Project2.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "XbAypvmqQEuG"
      },
      "source": [
        "# Setup"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "hCQ5-8BQQfGF"
      },
      "source": [
        "This project requires Python 3.7+."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "77_ZUgcWPrbS"
      },
      "outputs": [],
      "source": [
        "import sys\n",
        "\n",
        "assert sys.version_info >= (3, 7)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "LC8Qwni0QjMN"
      },
      "source": [
        "It also requires Scikit-Learn 1.0.1+."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Tyo_gZjeQbHT"
      },
      "outputs": [],
      "source": [
        "from packaging import version\n",
        "import sklearn\n",
        "\n",
        "assert version.parse(sklearn.__version__) >= version.parse(\"1.0.1\")"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "meb_w3kDU6IJ"
      },
      "source": [
        "# Get the Data"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "i69K1OLiZdN5"
      },
      "source": [
        "Now that we have the setup complete, we need to import the data. The following function checks if the data is stored locally. If not, it will download it from a repository. Finally, the function returns the data as a Pandas DataFrame for the sake of convenience."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "soMHp8M0VRsg"
      },
      "outputs": [],
      "source": [
        "from pathlib import Path\n",
        "import pandas as pd\n",
        "from scipy.io import arff\n",
        "import tarfile\n",
        "import urllib.request\n",
        "\n",
        "def load_dataset():\n",
        "    file_path = Path(\"dataset_31_credit-g.arff\")\n",
        "    if not file_path.is_file():\n",
        "        Path(\"datasets\").mkdir(parents=True, exist_ok=True)\n",
        "        url = \"https://www.openml.org/data/download/31/dataset_31_credit-g.arff\"\n",
        "        urllib.request.urlretrieve(url, file_path)\n",
        "    arff_file = arff.loadarff(file_path)\n",
        "\n",
        "    return pd.DataFrame(arff_file[0])\n",
        "\n",
        "default_dataset = load_dataset ()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "2tlLBC6MZwyK"
      },
      "source": [
        "Now that we have the data, we can take a quick look at it's structure."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "CiLTADPRXFKZ"
      },
      "outputs": [],
      "source": [
        "default_dataset.head ()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "CP_lldrkZ4Pk"
      },
      "source": [
        "It looks like the strings aren't formatted correctly. This is likely because they are byte strings, which are not handled by Scipy by default. We can fix this by iterating over the data and decoding any byte strings we find."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "2OrkPGbtY8ta"
      },
      "outputs": [],
      "source": [
        "default_dataset = default_dataset.map(lambda x: x.decode() if isinstance(x, bytes) else x)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "mxlVw7F3Zbrb"
      },
      "outputs": [],
      "source": [
        "default_dataset.head ()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "DajvkHitaO5u"
      },
      "source": [
        "Great! Looks like our data is in order now. Let's take a look at all the columns and their datatypes."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "v_F5gVjMXF0z"
      },
      "outputs": [],
      "source": [
        "default_dataset.info ()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "ILiCbGXtXN-E"
      },
      "outputs": [],
      "source": [
        "default_dataset.describe ()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "0_bH2j_BXTTa"
      },
      "outputs": [],
      "source": [
        "object_columns = default_dataset.select_dtypes(include=['object'])\n",
        "\n",
        "for column in object_columns:\n",
        "    print (default_dataset[column].value_counts())\n",
        "    print ()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "lcwovehYacur"
      },
      "source": [
        "It looks like there are many non-numerical columns that we'll need to handle later. Thankfully, they all seem to be categorical columns, so we can handle them similarly to how we did in Project 1. It also looks like non of the columns are missing values, which means we won't have to assume/impute anything this time."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "jjpMSRf9-Ug6"
      },
      "source": [
        "It also seems like the class column contains the labels we need to train the model."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "KO6kBtuTwr4W"
      },
      "source": [
        "Just like in Project 1, we need to split our data into train/test sets. We also need to keep the output consistent (even if the code is run multiple times) and stratified (so the samples represent the population). For this, we've split the labels and data, and then we feed them into train_test_split () with a constant seed and a train set size of 20%. This produces separate predictor/target value pairs for the train/test sets."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "J4hKZcg36jQA"
      },
      "outputs": [],
      "source": [
        "from sklearn.model_selection import train_test_split\n",
        "\n",
        "X_dataset = default_dataset.drop('class', axis=1) # Predictors\n",
        "y_dataset = default_dataset['class'] # Labels\n",
        "X_train, X_test, y_train, y_test = train_test_split(X_dataset, y_dataset, test_size=0.2, stratify=y_dataset, random_state=42)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "HRR7A8Ly88HV"
      },
      "source": [
        "We can confirm that the stratified sampling worked by looking at the value counts for each target value set. We can also set normalize=True to see the proportion of values rather than a count."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "l1OVek-S7iRU"
      },
      "outputs": [],
      "source": [
        "print(\"Original class distribution:\")\n",
        "print(y_dataset.value_counts(normalize=True))\n",
        "print(y_dataset.value_counts())\n",
        "print ()\n",
        "\n",
        "print(\"Training set class distribution:\")\n",
        "print(y_train.value_counts(normalize=True))\n",
        "print(y_train.value_counts())\n",
        "print ()\n",
        "\n",
        "print(\"Test set class distribution:\")\n",
        "print(y_test.value_counts(normalize=True))\n",
        "print(y_test.value_counts())"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "1_BF5YT79x2Y"
      },
      "source": [
        "As we can see, the data has been stratified quite nicely!"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ISqW97Mlwo4i"
      },
      "source": [
        "# Discover and Visualize the Data to Gain Insights"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "IOin8a63yYvK"
      },
      "source": [
        "In order to create an effective model, we must first understand the data so we can look for patterns ourselves. Let's start by generating some graphs for the numerical columns."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "_pj_xDPayKMi"
      },
      "outputs": [],
      "source": [
        "import matplotlib.pyplot as plt\n",
        "\n",
        "# extra code – the next 5 lines define the default font sizes\n",
        "plt.rc('font', size=14)\n",
        "plt.rc('axes', labelsize=14, titlesize=14)\n",
        "plt.rc('legend', fontsize=14)\n",
        "plt.rc('xtick', labelsize=10)\n",
        "plt.rc('ytick', labelsize=10)\n",
        "\n",
        "default_dataset.hist(bins=50, figsize=(10, 6))\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "WkNcYJWZyiSm"
      },
      "source": [
        "This makes it clear that some of the numerical columns are skewed. We can also generate a scatter matrix to see how each column is correlated with each other."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "collapsed": true,
        "id": "ILtXduxtzTv7"
      },
      "outputs": [],
      "source": [
        "from pandas.plotting import scatter_matrix\n",
        "\n",
        "# Finds all numeric columns\n",
        "numeric_column_names = []\n",
        "numeric_columns = default_dataset.select_dtypes(include=['float64'])\n",
        "for column in numeric_columns:\n",
        "    numeric_column_names.append(column)\n",
        "\n",
        "plt.rc('font', size=10)\n",
        "plt.rc('axes', labelsize=8, titlesize=10)\n",
        "scatter_matrix(default_dataset[numeric_column_names], figsize=(18, 12))\n",
        "plt.subplots_adjust (wspace = 0.1, hspace = 0.1)\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "vrpbXvzn--WK"
      },
      "source": [
        "Let's try to see the correlation between the categorical columns too. We can do this by using an ordinal encoder to convert one into a numerical column."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "sVrDd2il_Iiy"
      },
      "outputs": [],
      "source": [
        "from sklearn.preprocessing import OrdinalEncoder\n",
        "\n",
        "ordinal_encoder = OrdinalEncoder()\n",
        "corr_dataset = default_dataset.copy()\n",
        "\n",
        "for column in object_columns:\n",
        "    corr_dataset[[column]] = ordinal_encoder.fit_transform(default_dataset[[column]])\n",
        "\n",
        "class_column = default_dataset[['class']]\n",
        "# ordinal_encoder = OrdinalEncoder()\n",
        "# corr_dataset = default_dataset.drop('class', axis=1)\n",
        "# class_column_encoded = ordinal_encoder.fit_transform(class_column)\n",
        "# corr_dataset = default_dataset.copy()\n",
        "# corr_dataset[['class']] = ordinal_encoder.fit_transform(default_dataset[['class']])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "9ODt1jwxGj5A"
      },
      "outputs": [],
      "source": [
        "corr_dataset.info ()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "CBkPlQd0A6x7"
      },
      "outputs": [],
      "source": [
        "# class_column_encoded[:8]\n",
        "corr_dataset[['class']][:8]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Ottbt0BGE3SM"
      },
      "outputs": [],
      "source": [
        "corr_dataset.dtypes[\"class\"]"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "MRSJmytszYei"
      },
      "source": [
        "As you can see, all the columns are now numerical. Now we can use corr_dataset to check the correlation between columns. Since class is our label, let's start with that."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "mvWkwBLzCjb4"
      },
      "outputs": [],
      "source": [
        "corr_matrix = corr_dataset.corr(numeric_only=True)\n",
        "corr_matrix[\"class\"].sort_values(ascending=False)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "js3z4XtOzqwo"
      },
      "source": [
        "It looks like none of the columns have a strong correlation with our label. Let's plot a matrix of correlations to get a better idea of our data's relationships."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "collapsed": true,
        "id": "ko5JnYyuFrxf"
      },
      "outputs": [],
      "source": [
        "from pandas.plotting import scatter_matrix\n",
        "\n",
        "# Finds all numeric columns\n",
        "corr_numeric_column_names = []\n",
        "corr_numeric_columns = corr_dataset.select_dtypes(include=['float64'])\n",
        "for column in corr_numeric_columns:\n",
        "    corr_numeric_column_names.append(column)\n",
        "\n",
        "plt.rc('font', size=6)\n",
        "plt.rc('axes', labelsize=6, titlesize=8)\n",
        "scatter_matrix(corr_dataset, figsize=(18, 12))\n",
        "plt.subplots_adjust (wspace = 0.1, hspace = 0.1)\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "bSG9hzIVdwkj"
      },
      "source": [
        "It doesn't seem like there are any relationships that stick out here. Instead we can use the following function to find correlations above a certain threshold."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "HaeGjR6fXTav"
      },
      "outputs": [],
      "source": [
        "# Prints correlation values for columns with correlation values above a threshold\n",
        "# threshold is the absolute value of the minimum correlation you're looking for\n",
        "def checkForCorrelation (threshold, corr_matrix, columns, only_show_above_threshold = False):\n",
        "  assert threshold >= 0 and threshold <= 1\n",
        "\n",
        "  # Check if correlation values in each column are above the threshold\n",
        "  for column in columns:\n",
        "    corr_values = corr_matrix[column].sort_values(ascending=False)\n",
        "    if abs (corr_values.iloc[1]) > threshold or abs (corr_values.iloc[-1]) > threshold:\n",
        "      print (\"=====\" + column + \"=====\")\n",
        "      if only_show_above_threshold:\n",
        "        print (corr_values[abs(corr_values) > threshold])\n",
        "      else:\n",
        "        print(corr_values)\n",
        "      print ()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "RYHAeOxcae9t"
      },
      "outputs": [],
      "source": [
        "checkForCorrelation (0.5, corr_matrix, corr_numeric_column_names, True)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Th7EFBtSd-VD"
      },
      "source": [
        "It seems like only a couple of columns show any significant correlation. We will need to do some feature engineering to help the model understand the data better. Some ratio features such as monthly_credit_burden, installment_per_credit, dependents_per_credit might be more useful to the model."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "17ecqb4jnQKn"
      },
      "source": [
        "**Feature engineering suggestions from AI:**\n",
        "\n",
        "\n",
        "*   Bin age column into groups\n",
        "*   Bin credit_amount\n",
        "*   Create categories for duration\n",
        "* Extract separate features for marital status and gender from person_status column\n",
        "*   Use ordinal encoding for categorical columns whose categories are related in terms of order (like property magnitude)\n",
        "*   Use one hot encoding for the other categorical columns\n",
        "*   Create a new feature to represent the ratio of credit_amount to income\n",
        "*   Create a combination of savings_status and credit_amount\n",
        "*   Average duration of previous credits\n",
        "*   Create time-based features if possible, like time since last credit or the month/season the application was made\n",
        "*   Develop risk ratios for good/bad applicants\n",
        "\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "7wKTSZEO7jBU"
      },
      "source": [
        "# Prepare the Data"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "pwkcNQjyB7oT"
      },
      "source": [
        "First, let's make a copy of the training set and labels."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "zwrOdruX8BQy"
      },
      "outputs": [],
      "source": [
        "X_train_copy = X_train.copy()\n",
        "y_train_copy = y_train.copy()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "_edRXs51CKJ4"
      },
      "source": [
        "Now we can start cleaning the the data by constructing a pipeline of transformers. This will also give us an chance to implement some of our feature engineering ideas too. First, we should define some helper functions that will compute ratio features that  may be more helpful to the model. For this, we can reuse some of our code from Project 1."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "jKu8kyenu2Ko"
      },
      "source": [
        "There are a couple major changes we need to make first though. The first change addresses a data incompatibility in the column_ratio () function. Now it uses the DataFrame syntax. The other change is that the ratio_pipeline () function now computes the log of the new column. This is to ensure that the new columns are not skewed."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "SO6Jw4gqCobp"
      },
      "outputs": [],
      "source": [
        "from sklearn.preprocessing import StandardScaler\n",
        "\n",
        "from sklearn.pipeline import make_pipeline\n",
        "def column_ratio(X, columns):\n",
        "    # print ((X[columns[0]] / X[columns[1]]).to_frame().head (5))\n",
        "    return (X[columns[0]] / X[columns[1]]).to_frame()\n",
        "\n",
        "def ratio_name(function_transformer, feature_names_in):\n",
        "    return [\"ratio\"]  # feature names out\n",
        "\n",
        "def ratio_pipeline(columns):\n",
        "    print (columns)\n",
        "    return make_pipeline(\n",
        "        FunctionTransformer(lambda X: column_ratio (X, columns), feature_names_out=ratio_name),\n",
        "        FunctionTransformer(np.log, inverse_func=np.exp, feature_names_out=\"one-to-one\"),\n",
        "        StandardScaler())"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ZsyNUxyRvhl7"
      },
      "source": [
        "Next, we need to convert our categorical columns into numerical columns. We will use ordinal encoding for columns whose categories are related to each other. This will allow the model to better see the relationship between the categories. We will also use ordinal encoding for binary categories, since they only have two values."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "uLkUsWwuK08A"
      },
      "outputs": [],
      "source": [
        "from sklearn.preprocessing import OrdinalEncoder\n",
        "\n",
        "# Defines the intended order of categories for each column\n",
        "# This will ensure their relationships are preserved\n",
        "ordered_categories = {\n",
        "    \"checking_status\": [\"no checking\", \"<0\", \"0<=X<200\", \">=200\"],\n",
        "    \"savings_status\": [\"no known savings\", \"<100\", \"100<=X<500\", \"500<=X<1000\", \">=1000\"],\n",
        "    \"employment\": [\"unemployed\", \"<1\", \"1<=X<4\", \"4<=X<7\", \">=7\"],\n",
        "    \"own_telephone\": [\"yes\", \"none\"],\n",
        "    # \"foreign_worker\": [\"yes\", \"no\"]\n",
        "}\n",
        "\n",
        "# Performs ordinal encoding with the provided category orders\n",
        "class OrderedOrdinalEncoder(OrdinalEncoder):\n",
        "    def fit(self, data, y=None):\n",
        "        self.categories_ = [ordered_categories[column] for column in data.columns]\n",
        "        return super().fit(data, y)\n",
        "\n",
        "ordered_ordinal_encoder = OrderedOrdinalEncoder()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "vfFErEWwCu5v"
      },
      "source": [
        "Now we should convert the rest of the columns. These ones don't have the same type of relationship between categories, so we'll use one-hot encoding instead."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "BviqGy4GwilL"
      },
      "outputs": [],
      "source": [
        "from sklearn.preprocessing import OneHotEncoder\n",
        "\n",
        "cat_onehot_pipeline = make_pipeline(OneHotEncoder(handle_unknown=\"ignore\"))"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "yfZFMsHZw1Wv"
      },
      "source": [
        "Now that the categorical pipelines have been defined, let's work on the numerical column pipelines. Most columns require computing their log (to fix their skew) and standardizing them (to make them easier for the model to understand)."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "JrqbQiwFCuqZ"
      },
      "outputs": [],
      "source": [
        "import numpy as np\n",
        "from sklearn.impute import SimpleImputer\n",
        "from sklearn.preprocessing import FunctionTransformer, OneHotEncoder, OrdinalEncoder\n",
        "\n",
        "log_pipeline = make_pipeline(\n",
        "    FunctionTransformer(np.log, inverse_func=np.exp, feature_names_out=\"one-to-one\"),\n",
        "    StandardScaler())\n",
        "default_num_pipeline = make_pipeline(StandardScaler())"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Y9db6wrHEurB"
      },
      "source": [
        "Now we can wrap all our pipelines in  a ColumnTransformer to keep our code clean and provide an easy way to apply the transformers."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Elqpcb_mgQZL"
      },
      "outputs": [],
      "source": [
        "from sklearn.cluster import KMeans\n",
        "from sklearn.base import BaseEstimator, TransformerMixin\n",
        "from sklearn.metrics.pairwise import rbf_kernel\n",
        "from sklearn.pipeline import Pipeline\n",
        "\n",
        "class ClusterSimilarity(BaseEstimator, TransformerMixin):\n",
        "    def __init__(self, n_clusters=10, gamma=1.0, random_state=None):\n",
        "        self.n_clusters = n_clusters\n",
        "        self.gamma = gamma\n",
        "        self.random_state = random_state\n",
        "\n",
        "    def fit(self, X, y=None, sample_weight=None):\n",
        "        self.kmeans_ = KMeans(self.n_clusters, n_init=10,\n",
        "                              random_state=self.random_state)\n",
        "        self.kmeans_.fit(X, sample_weight=sample_weight)\n",
        "        return self  # always return self!\n",
        "\n",
        "    def transform(self, X):\n",
        "        return rbf_kernel(X, self.kmeans_.cluster_centers_, gamma=self.gamma)\n",
        "\n",
        "    def get_feature_names_out(self, names=None):\n",
        "        return [f\"Cluster {i} similarity\" for i in range(self.n_clusters)]\n",
        "\n",
        "cluster_simil = ClusterSimilarity(n_clusters=10, gamma=1., random_state=42)\n",
        "ordinal_geo_pipeline = Pipeline([\n",
        "    ('ordered_ordinal_encoder', ordered_ordinal_encoder),  # Assign a name to each step\n",
        "    ('cluster_simil', cluster_simil),                    # Assign a name to each step\n",
        "])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "2g2C-Hi6u_5H"
      },
      "outputs": [],
      "source": [
        "# from sklearn.preprocessing import QuantileTransformer\n",
        "\n",
        "# # Apply quantile transformation to the same columns you binned\n",
        "# quantile_transformer = make_pipeline(\n",
        "#     QuantileTransformer(output_distribution='uniform', random_state=42),\n",
        "#     StandardScaler()  # Standardize the result after quantile transformation\n",
        "# )"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "zhtlJQuXi1n-"
      },
      "outputs": [],
      "source": [
        "# from sklearn.preprocessing import PolynomialFeatures\n",
        "\n",
        "# polynomial_pipeline = Pipeline([\n",
        "#     ('poly', PolynomialFeatures(degree=2, include_bias=False))\n",
        "# ])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "hYuEzjZ2Es8m"
      },
      "outputs": [],
      "source": [
        "from sklearn.compose import ColumnTransformer, make_column_selector\n",
        "\n",
        "preprocessing = ColumnTransformer([\n",
        "        (\"monthly_credit_burden\", ratio_pipeline([\"credit_amount\", \"duration\"]), [\"credit_amount\", \"duration\"]),\n",
        "        (\"installment_per_credit\", ratio_pipeline([\"installment_commitment\", \"credit_amount\"]), [\"installment_commitment\", \"credit_amount\"]),\n",
        "        (\"dependents_per_credit\", ratio_pipeline([\"num_dependents\", \"credit_amount\"]), [\"num_dependents\", \"credit_amount\"]),\n",
        "        # (\"age_to_existing_credits\", ratio_pipeline([\"age\", \"credit_amount\"]), [\"age\", \"credit_amount\"]),\n",
        "        # (\"binning\", quantile_transformer, [\"age\", \"credit_amount\", \"duration\"]),\n",
        "        # (\"polynomial\", polynomial_pipeline, [\"duration\", \"credit_amount\", \"age\", \"installment_commitment\"]),\n",
        "        (\"log\", log_pipeline, [\"duration\", \"credit_amount\", \"installment_commitment\",\n",
        "                                \"residence_since\", \"age\", \"existing_credits\", \"num_dependents\"]),\n",
        "        (\"cat_ordinal\", ordered_ordinal_encoder, list(ordered_categories.keys())),\n",
        "        (\"cat_onehot\", cat_onehot_pipeline, [\"job\", \"housing\", \"other_payment_plans\",\n",
        "                                \"property_magnitude\", \"other_parties\", \"personal_status\",\n",
        "                                \"credit_history\"]),\n",
        "        (\"geo\", ordinal_geo_pipeline, [\"savings_status\", \"checking_status\"]),\n",
        "        ('drop_out', 'drop', [\"num_dependents\", \"foreign_worker\", \"purpose\"])\n",
        "    ],\n",
        "    remainder=default_num_pipeline)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "1dG1gB5ExRcO"
      },
      "source": [
        "With that done, the last part is simply running the training set through the transformers."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "FFmeNSLXSGw5"
      },
      "outputs": [],
      "source": [
        "X_train_prep = preprocessing.fit_transform(X_train_copy)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-T32XwzbxZ6I"
      },
      "source": [
        "Once we have the new data, we can use a method to see what new columns it has generated."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "collapsed": true,
        "id": "UbTsUlulnZfI"
      },
      "outputs": [],
      "source": [
        "preprocessing.get_feature_names_out()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "KeixoUp0xhRo"
      },
      "source": [
        "It's clear that the one-hot encoding added a lot of new columns, so we should keep that in mind later when we're choosing a model."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "sYsKvXpuxnzJ"
      },
      "source": [
        "We can also convert the data from a Numpy array back to a DataFrame. This will make it easier to view the data again and generate graphs."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "cBEVM35_1oNQ"
      },
      "outputs": [],
      "source": [
        "X_train_prep_df = pd.DataFrame(X_train_prep, columns=preprocessing.get_feature_names_out())"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "J7hp5NeXyKK4"
      },
      "outputs": [],
      "source": [
        "X_train_prep_df.info ()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 367
        },
        "id": "e32lskUSyNha",
        "outputId": "38aaf658-648c-43f2-db54-37a1c37aa4f3"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "       monthly_credit_burden__ratio  installment_per_credit__ratio  \\\n",
              "count                  8.000000e+02                   8.000000e+02   \n",
              "mean                  -5.195844e-16                  -1.720846e-16   \n",
              "std                    1.000626e+00                   1.000626e+00   \n",
              "min                   -2.820775e+00                  -3.025196e+00   \n",
              "25%                   -7.385604e-01                  -6.662208e-01   \n",
              "50%                   -6.767474e-02                   1.103665e-01   \n",
              "75%                    6.718776e-01                   7.711438e-01   \n",
              "max                    4.827774e+00                   2.494270e+00   \n",
              "\n",
              "       dependents_per_credit__ratio  log__duration  log__credit_amount  \\\n",
              "count                  8.000000e+02   8.000000e+02        8.000000e+02   \n",
              "mean                  -5.551115e-17  -5.928591e-16       -1.436629e-15   \n",
              "std                    1.000626e+00   1.000626e+00        1.000626e+00   \n",
              "min                   -2.690702e+00  -2.530398e+00       -2.886549e+00   \n",
              "25%                   -6.693017e-01  -6.575752e-01       -7.331832e-01   \n",
              "50%                    4.723493e-02   3.362790e-02       -5.658197e-02   \n",
              "75%                    6.526396e-01   5.240443e-01        6.445850e-01   \n",
              "max                    2.631514e+00   2.396867e+00        2.655598e+00   \n",
              "\n",
              "       log__installment_commitment  log__residence_since      log__age  \\\n",
              "count                 8.000000e+02          8.000000e+02  8.000000e+02   \n",
              "mean                  3.830269e-17          6.661338e-18 -1.643130e-16   \n",
              "std                   1.000626e+00          1.000626e+00  1.000626e+00   \n",
              "min                  -2.107715e+00         -2.004683e+00 -1.752317e+00   \n",
              "25%                  -6.496691e-01         -5.372006e-01 -7.488095e-01   \n",
              "50%                   2.032331e-01          3.212215e-01 -7.779471e-02   \n",
              "75%                   8.083768e-01          9.302818e-01  6.480387e-01   \n",
              "max                   8.083768e-01          9.302818e-01  2.667450e+00   \n",
              "\n",
              "       log__existing_credits  log__num_dependents  ...  \\\n",
              "count           8.000000e+02         8.000000e+02  ...   \n",
              "mean            1.088019e-16         1.576517e-16  ...   \n",
              "std             1.000626e+00         1.000626e+00  ...   \n",
              "min            -7.541061e-01        -4.425095e-01  ...   \n",
              "25%            -7.541061e-01        -4.425095e-01  ...   \n",
              "50%            -7.541061e-01        -4.425095e-01  ...   \n",
              "75%             1.159358e+00        -4.425095e-01  ...   \n",
              "max             3.072821e+00         2.259839e+00  ...   \n",
              "\n",
              "       geo__Cluster 0 similarity  geo__Cluster 1 similarity  \\\n",
              "count                 800.000000                 800.000000   \n",
              "mean                    0.301861                   0.242608   \n",
              "std                     0.395431                   0.389909   \n",
              "min                     0.000328                   0.000002   \n",
              "25%                     0.006810                   0.000123   \n",
              "50%                     0.018315                   0.018316   \n",
              "75%                     0.367869                   0.367879   \n",
              "max                     0.999971                   1.000000   \n",
              "\n",
              "       geo__Cluster 2 similarity  geo__Cluster 3 similarity  \\\n",
              "count               8.000000e+02               8.000000e+02   \n",
              "mean                1.194435e-01               7.206779e-02   \n",
              "std                 2.997730e-01               2.048517e-01   \n",
              "min                 2.521746e-11               3.266012e-10   \n",
              "25%                 4.104263e-06               3.499898e-05   \n",
              "50%                 4.975913e-04               2.251829e-03   \n",
              "75%                 1.812930e-02               1.960769e-02   \n",
              "max                 9.898261e-01               8.306351e-01   \n",
              "\n",
              "       geo__Cluster 4 similarity  geo__Cluster 5 similarity  \\\n",
              "count               8.000000e+02               8.000000e+02   \n",
              "mean                6.712836e-02               6.342229e-02   \n",
              "std                 2.105861e-01               2.088672e-01   \n",
              "min                 1.331001e-10               2.948304e-11   \n",
              "25%                 1.151916e-05               4.798507e-06   \n",
              "50%                 3.344317e-04               5.510880e-04   \n",
              "75%                 1.702192e-02               1.801705e-02   \n",
              "max                 9.458139e-01               9.836978e-01   \n",
              "\n",
              "       geo__Cluster 6 similarity  geo__Cluster 7 similarity  \\\n",
              "count                 800.000000                 800.000000   \n",
              "mean                    0.236633                   0.210484   \n",
              "std                     0.347395                   0.233344   \n",
              "min                     0.000002                   0.000335   \n",
              "25%                     0.000123                   0.006738   \n",
              "50%                     0.018316                   0.135335   \n",
              "75%                     0.367879                   0.367879   \n",
              "max                     1.000000                   1.000000   \n",
              "\n",
              "       geo__Cluster 8 similarity  geo__Cluster 9 similarity  \n",
              "count               8.000000e+02               8.000000e+02  \n",
              "mean                1.363701e-01               1.480321e-01  \n",
              "std                 2.282274e-01               2.158751e-01  \n",
              "min                 2.606142e-08               2.493622e-08  \n",
              "25%                 7.768799e-05               7.433383e-05  \n",
              "50%                 9.613104e-03               9.338486e-03  \n",
              "75%                 3.648516e-01               3.653336e-01  \n",
              "max                 9.917696e-01               9.930796e-01  \n",
              "\n",
              "[8 rows x 50 columns]"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-d4766eff-4ef7-4926-881b-002aa357ea40\" class=\"colab-df-container\">\n",
              "    <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>monthly_credit_burden__ratio</th>\n",
              "      <th>installment_per_credit__ratio</th>\n",
              "      <th>dependents_per_credit__ratio</th>\n",
              "      <th>log__duration</th>\n",
              "      <th>log__credit_amount</th>\n",
              "      <th>log__installment_commitment</th>\n",
              "      <th>log__residence_since</th>\n",
              "      <th>log__age</th>\n",
              "      <th>log__existing_credits</th>\n",
              "      <th>log__num_dependents</th>\n",
              "      <th>...</th>\n",
              "      <th>geo__Cluster 0 similarity</th>\n",
              "      <th>geo__Cluster 1 similarity</th>\n",
              "      <th>geo__Cluster 2 similarity</th>\n",
              "      <th>geo__Cluster 3 similarity</th>\n",
              "      <th>geo__Cluster 4 similarity</th>\n",
              "      <th>geo__Cluster 5 similarity</th>\n",
              "      <th>geo__Cluster 6 similarity</th>\n",
              "      <th>geo__Cluster 7 similarity</th>\n",
              "      <th>geo__Cluster 8 similarity</th>\n",
              "      <th>geo__Cluster 9 similarity</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>count</th>\n",
              "      <td>8.000000e+02</td>\n",
              "      <td>8.000000e+02</td>\n",
              "      <td>8.000000e+02</td>\n",
              "      <td>8.000000e+02</td>\n",
              "      <td>8.000000e+02</td>\n",
              "      <td>8.000000e+02</td>\n",
              "      <td>8.000000e+02</td>\n",
              "      <td>8.000000e+02</td>\n",
              "      <td>8.000000e+02</td>\n",
              "      <td>8.000000e+02</td>\n",
              "      <td>...</td>\n",
              "      <td>800.000000</td>\n",
              "      <td>800.000000</td>\n",
              "      <td>8.000000e+02</td>\n",
              "      <td>8.000000e+02</td>\n",
              "      <td>8.000000e+02</td>\n",
              "      <td>8.000000e+02</td>\n",
              "      <td>800.000000</td>\n",
              "      <td>800.000000</td>\n",
              "      <td>8.000000e+02</td>\n",
              "      <td>8.000000e+02</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>mean</th>\n",
              "      <td>-5.195844e-16</td>\n",
              "      <td>-1.720846e-16</td>\n",
              "      <td>-5.551115e-17</td>\n",
              "      <td>-5.928591e-16</td>\n",
              "      <td>-1.436629e-15</td>\n",
              "      <td>3.830269e-17</td>\n",
              "      <td>6.661338e-18</td>\n",
              "      <td>-1.643130e-16</td>\n",
              "      <td>1.088019e-16</td>\n",
              "      <td>1.576517e-16</td>\n",
              "      <td>...</td>\n",
              "      <td>0.301861</td>\n",
              "      <td>0.242608</td>\n",
              "      <td>1.194435e-01</td>\n",
              "      <td>7.206779e-02</td>\n",
              "      <td>6.712836e-02</td>\n",
              "      <td>6.342229e-02</td>\n",
              "      <td>0.236633</td>\n",
              "      <td>0.210484</td>\n",
              "      <td>1.363701e-01</td>\n",
              "      <td>1.480321e-01</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>std</th>\n",
              "      <td>1.000626e+00</td>\n",
              "      <td>1.000626e+00</td>\n",
              "      <td>1.000626e+00</td>\n",
              "      <td>1.000626e+00</td>\n",
              "      <td>1.000626e+00</td>\n",
              "      <td>1.000626e+00</td>\n",
              "      <td>1.000626e+00</td>\n",
              "      <td>1.000626e+00</td>\n",
              "      <td>1.000626e+00</td>\n",
              "      <td>1.000626e+00</td>\n",
              "      <td>...</td>\n",
              "      <td>0.395431</td>\n",
              "      <td>0.389909</td>\n",
              "      <td>2.997730e-01</td>\n",
              "      <td>2.048517e-01</td>\n",
              "      <td>2.105861e-01</td>\n",
              "      <td>2.088672e-01</td>\n",
              "      <td>0.347395</td>\n",
              "      <td>0.233344</td>\n",
              "      <td>2.282274e-01</td>\n",
              "      <td>2.158751e-01</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>min</th>\n",
              "      <td>-2.820775e+00</td>\n",
              "      <td>-3.025196e+00</td>\n",
              "      <td>-2.690702e+00</td>\n",
              "      <td>-2.530398e+00</td>\n",
              "      <td>-2.886549e+00</td>\n",
              "      <td>-2.107715e+00</td>\n",
              "      <td>-2.004683e+00</td>\n",
              "      <td>-1.752317e+00</td>\n",
              "      <td>-7.541061e-01</td>\n",
              "      <td>-4.425095e-01</td>\n",
              "      <td>...</td>\n",
              "      <td>0.000328</td>\n",
              "      <td>0.000002</td>\n",
              "      <td>2.521746e-11</td>\n",
              "      <td>3.266012e-10</td>\n",
              "      <td>1.331001e-10</td>\n",
              "      <td>2.948304e-11</td>\n",
              "      <td>0.000002</td>\n",
              "      <td>0.000335</td>\n",
              "      <td>2.606142e-08</td>\n",
              "      <td>2.493622e-08</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>25%</th>\n",
              "      <td>-7.385604e-01</td>\n",
              "      <td>-6.662208e-01</td>\n",
              "      <td>-6.693017e-01</td>\n",
              "      <td>-6.575752e-01</td>\n",
              "      <td>-7.331832e-01</td>\n",
              "      <td>-6.496691e-01</td>\n",
              "      <td>-5.372006e-01</td>\n",
              "      <td>-7.488095e-01</td>\n",
              "      <td>-7.541061e-01</td>\n",
              "      <td>-4.425095e-01</td>\n",
              "      <td>...</td>\n",
              "      <td>0.006810</td>\n",
              "      <td>0.000123</td>\n",
              "      <td>4.104263e-06</td>\n",
              "      <td>3.499898e-05</td>\n",
              "      <td>1.151916e-05</td>\n",
              "      <td>4.798507e-06</td>\n",
              "      <td>0.000123</td>\n",
              "      <td>0.006738</td>\n",
              "      <td>7.768799e-05</td>\n",
              "      <td>7.433383e-05</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>50%</th>\n",
              "      <td>-6.767474e-02</td>\n",
              "      <td>1.103665e-01</td>\n",
              "      <td>4.723493e-02</td>\n",
              "      <td>3.362790e-02</td>\n",
              "      <td>-5.658197e-02</td>\n",
              "      <td>2.032331e-01</td>\n",
              "      <td>3.212215e-01</td>\n",
              "      <td>-7.779471e-02</td>\n",
              "      <td>-7.541061e-01</td>\n",
              "      <td>-4.425095e-01</td>\n",
              "      <td>...</td>\n",
              "      <td>0.018315</td>\n",
              "      <td>0.018316</td>\n",
              "      <td>4.975913e-04</td>\n",
              "      <td>2.251829e-03</td>\n",
              "      <td>3.344317e-04</td>\n",
              "      <td>5.510880e-04</td>\n",
              "      <td>0.018316</td>\n",
              "      <td>0.135335</td>\n",
              "      <td>9.613104e-03</td>\n",
              "      <td>9.338486e-03</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>75%</th>\n",
              "      <td>6.718776e-01</td>\n",
              "      <td>7.711438e-01</td>\n",
              "      <td>6.526396e-01</td>\n",
              "      <td>5.240443e-01</td>\n",
              "      <td>6.445850e-01</td>\n",
              "      <td>8.083768e-01</td>\n",
              "      <td>9.302818e-01</td>\n",
              "      <td>6.480387e-01</td>\n",
              "      <td>1.159358e+00</td>\n",
              "      <td>-4.425095e-01</td>\n",
              "      <td>...</td>\n",
              "      <td>0.367869</td>\n",
              "      <td>0.367879</td>\n",
              "      <td>1.812930e-02</td>\n",
              "      <td>1.960769e-02</td>\n",
              "      <td>1.702192e-02</td>\n",
              "      <td>1.801705e-02</td>\n",
              "      <td>0.367879</td>\n",
              "      <td>0.367879</td>\n",
              "      <td>3.648516e-01</td>\n",
              "      <td>3.653336e-01</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>max</th>\n",
              "      <td>4.827774e+00</td>\n",
              "      <td>2.494270e+00</td>\n",
              "      <td>2.631514e+00</td>\n",
              "      <td>2.396867e+00</td>\n",
              "      <td>2.655598e+00</td>\n",
              "      <td>8.083768e-01</td>\n",
              "      <td>9.302818e-01</td>\n",
              "      <td>2.667450e+00</td>\n",
              "      <td>3.072821e+00</td>\n",
              "      <td>2.259839e+00</td>\n",
              "      <td>...</td>\n",
              "      <td>0.999971</td>\n",
              "      <td>1.000000</td>\n",
              "      <td>9.898261e-01</td>\n",
              "      <td>8.306351e-01</td>\n",
              "      <td>9.458139e-01</td>\n",
              "      <td>9.836978e-01</td>\n",
              "      <td>1.000000</td>\n",
              "      <td>1.000000</td>\n",
              "      <td>9.917696e-01</td>\n",
              "      <td>9.930796e-01</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>8 rows × 50 columns</p>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-d4766eff-4ef7-4926-881b-002aa357ea40')\"\n",
              "            title=\"Convert this dataframe to an interactive table.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "    <script>\n",
              "      const buttonEl =\n",
              "        document.querySelector('#df-d4766eff-4ef7-4926-881b-002aa357ea40 button.colab-df-convert');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      async function convertToInteractive(key) {\n",
              "        const element = document.querySelector('#df-d4766eff-4ef7-4926-881b-002aa357ea40');\n",
              "        const dataTable =\n",
              "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                    [key], {});\n",
              "        if (!dataTable) return;\n",
              "\n",
              "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "          + ' to learn more about interactive tables.';\n",
              "        element.innerHTML = '';\n",
              "        dataTable['output_type'] = 'display_data';\n",
              "        await google.colab.output.renderOutput(dataTable, element);\n",
              "        const docLink = document.createElement('div');\n",
              "        docLink.innerHTML = docLinkHtml;\n",
              "        element.appendChild(docLink);\n",
              "      }\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "\n",
              "<div id=\"df-a00113b7-63e5-434d-82d1-a87926ad0156\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-a00113b7-63e5-434d-82d1-a87926ad0156')\"\n",
              "            title=\"Suggest charts\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-a00113b7-63e5-434d-82d1-a87926ad0156 button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "\n",
              "    </div>\n",
              "  </div>\n"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "dataframe"
            }
          },
          "metadata": {},
          "execution_count": 161
        }
      ],
      "source": [
        "X_train_prep_df.describe ()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 261
        },
        "id": "GIY_Fg7mNPjS",
        "outputId": "47bb9459-3fab-428b-c812-d244ff6fad87"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "    checking_status  duration                  credit_history  \\\n",
              "675     no checking      30.0  critical/other existing credit   \n",
              "703        0<=X<200      30.0              delayed previously   \n",
              "12         0<=X<200      12.0                   existing paid   \n",
              "845        0<=X<200      21.0                   existing paid   \n",
              "795     no checking       9.0                   existing paid   \n",
              "\n",
              "                 purpose  credit_amount    savings_status employment  \\\n",
              "675             radio/tv         4530.0              <100     4<=X<7   \n",
              "703             business         2503.0        100<=X<500        >=7   \n",
              "12              radio/tv         1567.0              <100     1<=X<4   \n",
              "845  furniture/equipment         3976.0  no known savings     4<=X<7   \n",
              "795  furniture/equipment         2301.0        100<=X<500         <1   \n",
              "\n",
              "     installment_commitment     personal_status other_parties  \\\n",
              "675                     4.0  female div/dep/mar          none   \n",
              "703                     4.0         male single          none   \n",
              "12                      1.0  female div/dep/mar          none   \n",
              "845                     2.0         male single          none   \n",
              "795                     2.0  female div/dep/mar          none   \n",
              "\n",
              "     residence_since property_magnitude   age other_payment_plans housing  \\\n",
              "675              4.0                car  26.0                none    rent   \n",
              "703              2.0     life insurance  41.0              stores     own   \n",
              "12               1.0                car  22.0                none     own   \n",
              "845              3.0                car  35.0                none     own   \n",
              "795              4.0     life insurance  22.0                none    rent   \n",
              "\n",
              "     existing_credits                        job  num_dependents  \\\n",
              "675               1.0  high qualif/self emp/mgmt             1.0   \n",
              "703               2.0                    skilled             1.0   \n",
              "12                1.0                    skilled             1.0   \n",
              "845               1.0                    skilled             1.0   \n",
              "795               1.0                    skilled             1.0   \n",
              "\n",
              "    own_telephone foreign_worker  \n",
              "675           yes            yes  \n",
              "703          none            yes  \n",
              "12            yes            yes  \n",
              "845           yes            yes  \n",
              "795          none            yes  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-1255ee7c-1abe-47be-8db4-3181fee5e99e\" class=\"colab-df-container\">\n",
              "    <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>checking_status</th>\n",
              "      <th>duration</th>\n",
              "      <th>credit_history</th>\n",
              "      <th>purpose</th>\n",
              "      <th>credit_amount</th>\n",
              "      <th>savings_status</th>\n",
              "      <th>employment</th>\n",
              "      <th>installment_commitment</th>\n",
              "      <th>personal_status</th>\n",
              "      <th>other_parties</th>\n",
              "      <th>residence_since</th>\n",
              "      <th>property_magnitude</th>\n",
              "      <th>age</th>\n",
              "      <th>other_payment_plans</th>\n",
              "      <th>housing</th>\n",
              "      <th>existing_credits</th>\n",
              "      <th>job</th>\n",
              "      <th>num_dependents</th>\n",
              "      <th>own_telephone</th>\n",
              "      <th>foreign_worker</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>675</th>\n",
              "      <td>no checking</td>\n",
              "      <td>30.0</td>\n",
              "      <td>critical/other existing credit</td>\n",
              "      <td>radio/tv</td>\n",
              "      <td>4530.0</td>\n",
              "      <td>&lt;100</td>\n",
              "      <td>4&lt;=X&lt;7</td>\n",
              "      <td>4.0</td>\n",
              "      <td>female div/dep/mar</td>\n",
              "      <td>none</td>\n",
              "      <td>4.0</td>\n",
              "      <td>car</td>\n",
              "      <td>26.0</td>\n",
              "      <td>none</td>\n",
              "      <td>rent</td>\n",
              "      <td>1.0</td>\n",
              "      <td>high qualif/self emp/mgmt</td>\n",
              "      <td>1.0</td>\n",
              "      <td>yes</td>\n",
              "      <td>yes</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>703</th>\n",
              "      <td>0&lt;=X&lt;200</td>\n",
              "      <td>30.0</td>\n",
              "      <td>delayed previously</td>\n",
              "      <td>business</td>\n",
              "      <td>2503.0</td>\n",
              "      <td>100&lt;=X&lt;500</td>\n",
              "      <td>&gt;=7</td>\n",
              "      <td>4.0</td>\n",
              "      <td>male single</td>\n",
              "      <td>none</td>\n",
              "      <td>2.0</td>\n",
              "      <td>life insurance</td>\n",
              "      <td>41.0</td>\n",
              "      <td>stores</td>\n",
              "      <td>own</td>\n",
              "      <td>2.0</td>\n",
              "      <td>skilled</td>\n",
              "      <td>1.0</td>\n",
              "      <td>none</td>\n",
              "      <td>yes</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>12</th>\n",
              "      <td>0&lt;=X&lt;200</td>\n",
              "      <td>12.0</td>\n",
              "      <td>existing paid</td>\n",
              "      <td>radio/tv</td>\n",
              "      <td>1567.0</td>\n",
              "      <td>&lt;100</td>\n",
              "      <td>1&lt;=X&lt;4</td>\n",
              "      <td>1.0</td>\n",
              "      <td>female div/dep/mar</td>\n",
              "      <td>none</td>\n",
              "      <td>1.0</td>\n",
              "      <td>car</td>\n",
              "      <td>22.0</td>\n",
              "      <td>none</td>\n",
              "      <td>own</td>\n",
              "      <td>1.0</td>\n",
              "      <td>skilled</td>\n",
              "      <td>1.0</td>\n",
              "      <td>yes</td>\n",
              "      <td>yes</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>845</th>\n",
              "      <td>0&lt;=X&lt;200</td>\n",
              "      <td>21.0</td>\n",
              "      <td>existing paid</td>\n",
              "      <td>furniture/equipment</td>\n",
              "      <td>3976.0</td>\n",
              "      <td>no known savings</td>\n",
              "      <td>4&lt;=X&lt;7</td>\n",
              "      <td>2.0</td>\n",
              "      <td>male single</td>\n",
              "      <td>none</td>\n",
              "      <td>3.0</td>\n",
              "      <td>car</td>\n",
              "      <td>35.0</td>\n",
              "      <td>none</td>\n",
              "      <td>own</td>\n",
              "      <td>1.0</td>\n",
              "      <td>skilled</td>\n",
              "      <td>1.0</td>\n",
              "      <td>yes</td>\n",
              "      <td>yes</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>795</th>\n",
              "      <td>no checking</td>\n",
              "      <td>9.0</td>\n",
              "      <td>existing paid</td>\n",
              "      <td>furniture/equipment</td>\n",
              "      <td>2301.0</td>\n",
              "      <td>100&lt;=X&lt;500</td>\n",
              "      <td>&lt;1</td>\n",
              "      <td>2.0</td>\n",
              "      <td>female div/dep/mar</td>\n",
              "      <td>none</td>\n",
              "      <td>4.0</td>\n",
              "      <td>life insurance</td>\n",
              "      <td>22.0</td>\n",
              "      <td>none</td>\n",
              "      <td>rent</td>\n",
              "      <td>1.0</td>\n",
              "      <td>skilled</td>\n",
              "      <td>1.0</td>\n",
              "      <td>none</td>\n",
              "      <td>yes</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-1255ee7c-1abe-47be-8db4-3181fee5e99e')\"\n",
              "            title=\"Convert this dataframe to an interactive table.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "    <script>\n",
              "      const buttonEl =\n",
              "        document.querySelector('#df-1255ee7c-1abe-47be-8db4-3181fee5e99e button.colab-df-convert');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      async function convertToInteractive(key) {\n",
              "        const element = document.querySelector('#df-1255ee7c-1abe-47be-8db4-3181fee5e99e');\n",
              "        const dataTable =\n",
              "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                    [key], {});\n",
              "        if (!dataTable) return;\n",
              "\n",
              "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "          + ' to learn more about interactive tables.';\n",
              "        element.innerHTML = '';\n",
              "        dataTable['output_type'] = 'display_data';\n",
              "        await google.colab.output.renderOutput(dataTable, element);\n",
              "        const docLink = document.createElement('div');\n",
              "        docLink.innerHTML = docLinkHtml;\n",
              "        element.appendChild(docLink);\n",
              "      }\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "\n",
              "<div id=\"df-aaa897bc-8c14-47a0-95ba-6b2afac65265\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-aaa897bc-8c14-47a0-95ba-6b2afac65265')\"\n",
              "            title=\"Suggest charts\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-aaa897bc-8c14-47a0-95ba-6b2afac65265 button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "\n",
              "    </div>\n",
              "  </div>\n"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "dataframe",
              "summary": "{\n  \"name\": \"X_train\",\n  \"rows\": 5,\n  \"fields\": [\n    {\n      \"column\": \"checking_status\",\n      \"properties\": {\n        \"dtype\": \"category\",\n        \"num_unique_values\": 2,\n        \"samples\": [\n          \"0<=X<200\",\n          \"no checking\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"duration\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 9.813256340277675,\n        \"min\": 9.0,\n        \"max\": 30.0,\n        \"num_unique_values\": 4,\n        \"samples\": [\n          12.0,\n          9.0\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"credit_history\",\n      \"properties\": {\n        \"dtype\": \"string\",\n        \"num_unique_values\": 3,\n        \"samples\": [\n          \"critical/other existing credit\",\n          \"delayed previously\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"purpose\",\n      \"properties\": {\n        \"dtype\": \"string\",\n        \"num_unique_values\": 3,\n        \"samples\": [\n          \"radio/tv\",\n          \"business\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"credit_amount\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 1232.8370938611477,\n        \"min\": 1567.0,\n        \"max\": 4530.0,\n        \"num_unique_values\": 5,\n        \"samples\": [\n          2503.0,\n          2301.0\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"savings_status\",\n      \"properties\": {\n        \"dtype\": \"string\",\n        \"num_unique_values\": 3,\n        \"samples\": [\n          \"<100\",\n          \"100<=X<500\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"employment\",\n      \"properties\": {\n        \"dtype\": \"string\",\n        \"num_unique_values\": 4,\n        \"samples\": [\n          \">=7\",\n          \"<1\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"installment_commitment\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 1.3416407864998738,\n        \"min\": 1.0,\n        \"max\": 4.0,\n        \"num_unique_values\": 3,\n        \"samples\": [\n          4.0,\n          1.0\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"personal_status\",\n      \"properties\": {\n        \"dtype\": \"category\",\n        \"num_unique_values\": 2,\n        \"samples\": [\n          \"male single\",\n          \"female div/dep/mar\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"other_parties\",\n      \"properties\": {\n        \"dtype\": \"category\",\n        \"num_unique_values\": 1,\n        \"samples\": [\n          \"none\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"residence_since\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 1.3038404810405297,\n        \"min\": 1.0,\n        \"max\": 4.0,\n        \"num_unique_values\": 4,\n        \"samples\": [\n          2.0\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"property_magnitude\",\n      \"properties\": {\n        \"dtype\": \"category\",\n        \"num_unique_values\": 2,\n        \"samples\": [\n          \"life insurance\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"age\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 8.46758525200662,\n        \"min\": 22.0,\n        \"max\": 41.0,\n        \"num_unique_values\": 4,\n        \"samples\": [\n          41.0\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"other_payment_plans\",\n      \"properties\": {\n        \"dtype\": \"category\",\n        \"num_unique_values\": 2,\n        \"samples\": [\n          \"stores\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"housing\",\n      \"properties\": {\n        \"dtype\": \"category\",\n        \"num_unique_values\": 2,\n        \"samples\": [\n          \"own\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"existing_credits\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 0.44721359549995804,\n        \"min\": 1.0,\n        \"max\": 2.0,\n        \"num_unique_values\": 2,\n        \"samples\": [\n          2.0\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"job\",\n      \"properties\": {\n        \"dtype\": \"category\",\n        \"num_unique_values\": 2,\n        \"samples\": [\n          \"skilled\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"num_dependents\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 0.0,\n        \"min\": 1.0,\n        \"max\": 1.0,\n        \"num_unique_values\": 1,\n        \"samples\": [\n          1.0\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"own_telephone\",\n      \"properties\": {\n        \"dtype\": \"category\",\n        \"num_unique_values\": 2,\n        \"samples\": [\n          \"none\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"foreign_worker\",\n      \"properties\": {\n        \"dtype\": \"category\",\n        \"num_unique_values\": 1,\n        \"samples\": [\n          \"yes\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    }\n  ]\n}"
            }
          },
          "metadata": {},
          "execution_count": 162
        }
      ],
      "source": [
        "X_train.head (5)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 273
        },
        "id": "54QGc5xc2JMV",
        "outputId": "28faad6c-a11b-4662-d335-f0a99678e26e"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "   monthly_credit_burden__ratio  installment_per_credit__ratio  \\\n",
              "0                      0.209324                      -0.269834   \n",
              "1                     -0.769248                       0.316195   \n",
              "2                     -0.030305                      -0.590627   \n",
              "3                      0.582502                      -0.825704   \n",
              "4                      1.077973                      -0.285412   \n",
              "\n",
              "   dependents_per_credit__ratio  log__duration  log__credit_amount  \\\n",
              "0                     -0.954237       0.904441            0.847377   \n",
              "1                     -0.219971       0.904441            0.082768   \n",
              "2                      0.359696      -0.657575           -0.520853   \n",
              "3                     -0.792779       0.296411            0.679247   \n",
              "4                     -0.115819      -1.147992           -0.025687   \n",
              "\n",
              "   log__installment_commitment  log__residence_since  log__age  \\\n",
              "0                     0.808377              0.930282 -0.875008   \n",
              "1                     0.808377             -0.537201  0.648039   \n",
              "2                    -2.107715             -2.004683 -1.433613   \n",
              "3                    -0.649669              0.321222  0.118960   \n",
              "4                    -0.649669              0.930282 -1.433613   \n",
              "\n",
              "   log__existing_credits  log__num_dependents  ...  geo__Cluster 0 similarity  \\\n",
              "0              -0.754106            -0.442509  ...                   0.018315   \n",
              "1               1.159358            -0.442509  ...                   0.006883   \n",
              "2              -0.754106            -0.442509  ...                   0.367869   \n",
              "3              -0.754106            -0.442509  ...                   0.006595   \n",
              "4              -0.754106            -0.442509  ...                   0.000343   \n",
              "\n",
              "   geo__Cluster 1 similarity  geo__Cluster 2 similarity  \\\n",
              "0                   1.000000               1.812930e-02   \n",
              "1                   0.000002               2.521746e-11   \n",
              "2                   0.000123               4.104263e-06   \n",
              "3                   0.000002               2.240852e-04   \n",
              "4                   0.018316               1.113903e-07   \n",
              "\n",
              "   geo__Cluster 3 similarity  geo__Cluster 4 similarity  \\\n",
              "0               3.499898e-05               1.151916e-05   \n",
              "1               2.156210e-07               9.458139e-01   \n",
              "2               2.310620e-02               3.257759e-02   \n",
              "3               8.306351e-01               3.764233e-07   \n",
              "4               3.266012e-10               3.344317e-04   \n",
              "\n",
              "   geo__Cluster 5 similarity  geo__Cluster 6 similarity  \\\n",
              "0               1.801705e-02                   0.000123   \n",
              "1               2.619896e-04                   0.018316   \n",
              "2               4.798507e-06                   1.000000   \n",
              "3               2.948304e-11                   0.018316   \n",
              "4               9.836978e-01                   0.000002   \n",
              "\n",
              "   geo__Cluster 7 similarity  geo__Cluster 8 similarity  \\\n",
              "0                   0.367879               3.648516e-01   \n",
              "1                   0.000335               7.768799e-05   \n",
              "2                   0.018316               7.768799e-05   \n",
              "3                   0.000335               2.606142e-08   \n",
              "4                   0.006738               3.648516e-01   \n",
              "\n",
              "   geo__Cluster 9 similarity  \n",
              "0               3.653336e-01  \n",
              "1               2.493622e-08  \n",
              "2               7.433383e-05  \n",
              "3               7.433383e-05  \n",
              "4               1.225558e-04  \n",
              "\n",
              "[5 rows x 50 columns]"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-f158fe82-a6ff-46df-ad35-bdce67db7485\" class=\"colab-df-container\">\n",
              "    <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>monthly_credit_burden__ratio</th>\n",
              "      <th>installment_per_credit__ratio</th>\n",
              "      <th>dependents_per_credit__ratio</th>\n",
              "      <th>log__duration</th>\n",
              "      <th>log__credit_amount</th>\n",
              "      <th>log__installment_commitment</th>\n",
              "      <th>log__residence_since</th>\n",
              "      <th>log__age</th>\n",
              "      <th>log__existing_credits</th>\n",
              "      <th>log__num_dependents</th>\n",
              "      <th>...</th>\n",
              "      <th>geo__Cluster 0 similarity</th>\n",
              "      <th>geo__Cluster 1 similarity</th>\n",
              "      <th>geo__Cluster 2 similarity</th>\n",
              "      <th>geo__Cluster 3 similarity</th>\n",
              "      <th>geo__Cluster 4 similarity</th>\n",
              "      <th>geo__Cluster 5 similarity</th>\n",
              "      <th>geo__Cluster 6 similarity</th>\n",
              "      <th>geo__Cluster 7 similarity</th>\n",
              "      <th>geo__Cluster 8 similarity</th>\n",
              "      <th>geo__Cluster 9 similarity</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0.209324</td>\n",
              "      <td>-0.269834</td>\n",
              "      <td>-0.954237</td>\n",
              "      <td>0.904441</td>\n",
              "      <td>0.847377</td>\n",
              "      <td>0.808377</td>\n",
              "      <td>0.930282</td>\n",
              "      <td>-0.875008</td>\n",
              "      <td>-0.754106</td>\n",
              "      <td>-0.442509</td>\n",
              "      <td>...</td>\n",
              "      <td>0.018315</td>\n",
              "      <td>1.000000</td>\n",
              "      <td>1.812930e-02</td>\n",
              "      <td>3.499898e-05</td>\n",
              "      <td>1.151916e-05</td>\n",
              "      <td>1.801705e-02</td>\n",
              "      <td>0.000123</td>\n",
              "      <td>0.367879</td>\n",
              "      <td>3.648516e-01</td>\n",
              "      <td>3.653336e-01</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>-0.769248</td>\n",
              "      <td>0.316195</td>\n",
              "      <td>-0.219971</td>\n",
              "      <td>0.904441</td>\n",
              "      <td>0.082768</td>\n",
              "      <td>0.808377</td>\n",
              "      <td>-0.537201</td>\n",
              "      <td>0.648039</td>\n",
              "      <td>1.159358</td>\n",
              "      <td>-0.442509</td>\n",
              "      <td>...</td>\n",
              "      <td>0.006883</td>\n",
              "      <td>0.000002</td>\n",
              "      <td>2.521746e-11</td>\n",
              "      <td>2.156210e-07</td>\n",
              "      <td>9.458139e-01</td>\n",
              "      <td>2.619896e-04</td>\n",
              "      <td>0.018316</td>\n",
              "      <td>0.000335</td>\n",
              "      <td>7.768799e-05</td>\n",
              "      <td>2.493622e-08</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>-0.030305</td>\n",
              "      <td>-0.590627</td>\n",
              "      <td>0.359696</td>\n",
              "      <td>-0.657575</td>\n",
              "      <td>-0.520853</td>\n",
              "      <td>-2.107715</td>\n",
              "      <td>-2.004683</td>\n",
              "      <td>-1.433613</td>\n",
              "      <td>-0.754106</td>\n",
              "      <td>-0.442509</td>\n",
              "      <td>...</td>\n",
              "      <td>0.367869</td>\n",
              "      <td>0.000123</td>\n",
              "      <td>4.104263e-06</td>\n",
              "      <td>2.310620e-02</td>\n",
              "      <td>3.257759e-02</td>\n",
              "      <td>4.798507e-06</td>\n",
              "      <td>1.000000</td>\n",
              "      <td>0.018316</td>\n",
              "      <td>7.768799e-05</td>\n",
              "      <td>7.433383e-05</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>0.582502</td>\n",
              "      <td>-0.825704</td>\n",
              "      <td>-0.792779</td>\n",
              "      <td>0.296411</td>\n",
              "      <td>0.679247</td>\n",
              "      <td>-0.649669</td>\n",
              "      <td>0.321222</td>\n",
              "      <td>0.118960</td>\n",
              "      <td>-0.754106</td>\n",
              "      <td>-0.442509</td>\n",
              "      <td>...</td>\n",
              "      <td>0.006595</td>\n",
              "      <td>0.000002</td>\n",
              "      <td>2.240852e-04</td>\n",
              "      <td>8.306351e-01</td>\n",
              "      <td>3.764233e-07</td>\n",
              "      <td>2.948304e-11</td>\n",
              "      <td>0.018316</td>\n",
              "      <td>0.000335</td>\n",
              "      <td>2.606142e-08</td>\n",
              "      <td>7.433383e-05</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>1.077973</td>\n",
              "      <td>-0.285412</td>\n",
              "      <td>-0.115819</td>\n",
              "      <td>-1.147992</td>\n",
              "      <td>-0.025687</td>\n",
              "      <td>-0.649669</td>\n",
              "      <td>0.930282</td>\n",
              "      <td>-1.433613</td>\n",
              "      <td>-0.754106</td>\n",
              "      <td>-0.442509</td>\n",
              "      <td>...</td>\n",
              "      <td>0.000343</td>\n",
              "      <td>0.018316</td>\n",
              "      <td>1.113903e-07</td>\n",
              "      <td>3.266012e-10</td>\n",
              "      <td>3.344317e-04</td>\n",
              "      <td>9.836978e-01</td>\n",
              "      <td>0.000002</td>\n",
              "      <td>0.006738</td>\n",
              "      <td>3.648516e-01</td>\n",
              "      <td>1.225558e-04</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>5 rows × 50 columns</p>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-f158fe82-a6ff-46df-ad35-bdce67db7485')\"\n",
              "            title=\"Convert this dataframe to an interactive table.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "    <script>\n",
              "      const buttonEl =\n",
              "        document.querySelector('#df-f158fe82-a6ff-46df-ad35-bdce67db7485 button.colab-df-convert');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      async function convertToInteractive(key) {\n",
              "        const element = document.querySelector('#df-f158fe82-a6ff-46df-ad35-bdce67db7485');\n",
              "        const dataTable =\n",
              "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                    [key], {});\n",
              "        if (!dataTable) return;\n",
              "\n",
              "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "          + ' to learn more about interactive tables.';\n",
              "        element.innerHTML = '';\n",
              "        dataTable['output_type'] = 'display_data';\n",
              "        await google.colab.output.renderOutput(dataTable, element);\n",
              "        const docLink = document.createElement('div');\n",
              "        docLink.innerHTML = docLinkHtml;\n",
              "        element.appendChild(docLink);\n",
              "      }\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "\n",
              "<div id=\"df-4e648d80-51db-4f09-a186-aa62367c75d6\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-4e648d80-51db-4f09-a186-aa62367c75d6')\"\n",
              "            title=\"Suggest charts\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-4e648d80-51db-4f09-a186-aa62367c75d6 button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "\n",
              "    </div>\n",
              "  </div>\n"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "dataframe"
            }
          },
          "metadata": {},
          "execution_count": 163
        }
      ],
      "source": [
        "X_train_prep_df.head (5)"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Let's see what the data looks like now."
      ],
      "metadata": {
        "id": "gLC1m1XDu589"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 842
        },
        "collapsed": true,
        "id": "FbZGPMOkme37",
        "outputId": "3f30837a-fcb9-49df-e37a-5b64c9aa5151"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1400x1000 with 56 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAABM0AAAM5CAYAAAAUjsZ5AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/bCgiHAAAACXBIWXMAAA9hAAAPYQGoP6dpAAEAAElEQVR4nOzdd1wU1/o/8A9dLKCIIioRCyJ9UUBBwY7tatQQa1CvLWoSTTGxXBON2BNLor+oN8YWlahJxFw1FhRjL+RK1KhcNGqixBJBQFBYdp/fH3x3srvsLtsbz/v18iVbZuY5M8+cmT1z5owDEREYY4wxxhhjjDHGGGMCR0sHwBhjjDHGGGOMMcaYteFGM8YYY4wxxhhjjDHGlHCjGWOMMcYYY4wxxhhjSrjRjDHGGGOMMcYYY4wxJdxoxhhjjDHGGGOMMcaYEm40Y4wxxhhjjDHGGGNMCTeaMcYYY4wxxhhjjDGmhBvNGGOMMcYYY4wxxhhTwo1mjDHGGGOMMcYYY4wp4UYzxhhjjDHGGGOMMcaUcKOZlu7cuYN169YpvOfv74+srCyV3+/SpQvS0tJMH5gOvL29cefOHQBA3759kZ2dDQDYvHkzbty4UeX0pirT1atX4e/vb/T5GktWVha++eYbhfdEIhGKiooqfVfd+9qaN28eXrx4odV35bfHmDFjsGrVKr2Xq6u0tDScO3fObMszp2fPnsHBwUF4Lb9NV61ahQcPHpglDuX98ocffsA777yj0zzk93lLMkYc5lz31mL69OmYN28eAGDdunX45JNPAKiuk0xF1bFP/vhhyxwcHPD06VOLLT8pKQmbN282+nyVjyMfffQRtm/fbvTlVHeWzh9j2LdvH7p06QIAyMzMxNChQwEAT58+xZIlSywYmeHMVU+aOw/Gjx+PjIwMvafX9pwfqKhL3n77bWG6gQMH6r1cXZnzOGcs8tvGEufJplxntljfadpX5M+vrFVubi7i4+MtHQb7P9xopiVVPxwsrby8XO9pDxw4gMDAQAC6HUANZUjMplJVTKoOQllZWahTp47K76p6X1sff/yx1o1mlmTOkwFj5IxEItF7WvltasyGm6rKpbxfDhgwACtXrjTKsm2RNTSaGZqLhkw/adIkvP/++wCMe2JcVUyqjn3yxw9mXlKpFFKpVON3lI8j8+fPx8iRI00dGrMChtQxUVFR2LlzJwBuNLNmGzZsQNeuXfWe3pzn/Iawxe0nv23srdHMkvSt1wzdVyytcePGOHnypKXDYP/HbhrNHBwcsHDhQrRv3x7+/v5IS0vD4sWLERUVhYCAABw/flz47tdff43w8HCEh4ejX79+uH//PoCKA0mPHj0wfPhwhIWFISoqCr/99huAih8s2dnZEIlEGDBggDCv77//HrGxsWjevDkWLFhQKa7c3Fz4+PigpKREeG/EiBFYu3atxvIsXboUYWFhiIiIQIcOHVBSUoLjx48jJCQE48aNg0gkwp49e5CTk4N+/fohOjoa4eHhWLNmjTCPH374AUFBQQgPD8cHH3ygMH9ZL7kNGzYgMzMT77zzDkQiEQ4cOKAxrqNHjyI6OhqtWrXCe++9ByICULkXmvwV9DFjxmDs2LFISEhAaGgogIqrVwEBAWjXrl2lCv7QoUPo1KkT2rVrh5iYGOEqwfHjxxEaGoopU6YgIiICISEhyMzM1BivOv7+/pgxYwZiYmIwevRoPHjwAF27dkW7du0QEhKCN998E1KpFI8ePcJHH32EjIwMiEQiTJo0CYDiFZfMzEzExcUhPDwcDg4OOHjwoLCMjz76SGV+LFiwAEFBQRCJRBCJRLh7964w7/j4eIhEIjx69Ag7duxA+/btERkZiYiICPznP/+psmzz5s3DkCFD0L9/f7Ru3Rr/+Mc/cPXqVfTq1QutW7fG8OHDhR9dRUVFmDBhAmJiYhAeHo6JEyeirKwMQMU2nT59OuLj49GyZUshvgMHDuCHH37AJ598ApFIhA0bNqiNRbbt4+Li0Lp1a4wePRrPnz/XatlTp05FbGwsEhMT1c6/rKwM77//PkJDQxEREYHevXsDqNiXu3btildeeQVhYWG4cOECLl68iG7duiEqKgqRkZHYvXu3MJ/169cjICAAkZGRlRqmZNt6/vz5yM3NxdChQyESidT2MlVH1f6rbvuq2i+Vr/J+8sknCAkJQVhYGEaOHImCggK1+7ymesLBwQFz5sxBZGQkWrdurdAjRd06u3PnDurWrYu5c+eiXbt2aNWqlULdoanuqSqWRYsWISYmBs2bN8emTZsAQOW6/89//oPw8HCIRCKEhoZi7969ate9LN7p06cjPDwcISEhSE9PFz7XVOcobzN1zp49i06dOiEiIgLh4eFCPMp1jVgsxsyZMxETEwORSIQhQ4YgPz8fAPDnn3+iV69eCA4ORo8ePXDv3j1h/rIr/urqJF0o71/l5eXo1asXoqKiEBISghEjRqC4uBiA6mOffC/rmzdvokePHsK2sLYe1tqSr8djYmJw+vRp4bP169ejdevWaNu2LVJSUhR6ourixo0biIuLQ0hICAYOHIjCwkLhM+XewvJXwefNm4dXXnkFvXr1QmhoKP78809Mnz4d0dHREIlESEhIEHr+qTqOyM/72bNnGDt2LEJDQxEaGoqPP/5YWKa6Op9VzRz5c//+fSQlJSEsLAzh4eH48MMPAag+x/r666/Rvn17tG3bFgkJCfjll18AAGKxGFOmTEFAQIBCXQdU1HcikQhARR4VFRVBJBIhKipKbUya6g7ZOdvkyZMRHh6OsLAwXL58GWPGjEFYWBjat28vnH9LJBLhWB4aGoq33npLOB+oat8YOnQo+vfvj+DgYHTr1g15eXlGqSf1YY48UL7L4PXXX0f37t3RunVrDB48WFhvqo6Rqs4trly5gk6dOqFt27YIDg5W+VtGmS7bFgA+/fRTxMTEoG3btujduzfu3r0LwHjbT9VvJkD1eZJsudqeJ48ZMwYTJ05Ejx490Lx5c4wdOxYXLlxAly5d0KJFC7z77ruVto2q82RrW2eGMFeeK/8OULdO1J0Pyu8rms6vNJ2XadrH1P0G0RSrKlKpFG+++SaCgoIQERGBdu3a4cWLF8K5q4y6c2QAuH79Onr16iW0b8gudj548ABDhgxBTEwMwsLCMGfOHL22BwNAdgIArVq1ioiI0tPTqVatWrRp0yYiItq1axdFRUUREdGVK1fIx8eH7t27R0RECxYsoN69exMR0aZNm8jDw4N+++03IiKaMWMGTZw4kYiIMjIyKCIiQmGZzZo1o7feeouIiB4/fkweHh7CfDt37kx79uwhIqIRI0bQ+vXriYjowYMH1KBBAyoqKlJbls2bN1N0dDQ9ffqUiIjy8vKovLycMjIyyMHBgY4fP05EROXl5dSuXTu6fv06EREVFxdTWFgYXbhwgR4+fEheXl7066+/EhHR+vXrCQDdvn1biP3SpUuVYtWkc+fO1K1bNyorK6Pi4mJq164dbd++XeU8XnnlFWH9jx49msLDw6mwsJCIiPbt20fBwcFUUFBAUqmURo4cSc2aNSMiolu3blGHDh2ooKCAiIhycnKoUaNG9OLFC8rIyCAnJyc6d+4cERGtXbuWEhMTq4xblWbNmtG4ceNIKpUSEdHz58+FbVJeXk79+vWj1NRUIqrIi5dffllhegCUn59PpaWl5OfnRwcPHhTel21fdfmRl5dHnp6eVFJSQkQV2+358+cK85X566+/hBhv375NPj4+9OLFi0rrfPTo0bRy5UoiIpo7dy41b96c8vLySCqVUkJCArVv354KCwtJLBZTREQE7du3j4iIJkyYQFu2bCEiIqlUSuPGjaNly5YJ8x84cCCJxWIqKSkhf39/OnPmTKXlaTJ69GgKDg6mwsJCKi8vp3/84x+0cOFCrZbdq1cvKisr0zj/efPm0YABA4R18ujRIyKq2Gbu7u5048YNIiLKz88nkUhEubm5wvbw8/Oje/fuCXWC7LNZs2aRfNUov03k9xtdKe+/RNpvX1mZZHl44MABatOmjRDXhAkTaPTo0Sr3+Zs3b6qtJ2TlmzNnDhFV7H/16tWj27dva1xnt2/fJgD07bffEhHRjz/+SK1btyYi0lj3aKqzZLF8+umnRER0/fp1ql27NonFYpXrPjw8XMhHiUSisN8ok8W7YcMGIiI6e/YsNWjQgAoLC6usc5S3mSpPnjyhhg0b0okTJ4R4njx5IsQtX9csXLiQ5s+fL0w7f/58mjJlChERJSUlCdvi3r175O3tTXPnziWiiv162rRpRKS6TtKF8v4llUrpr7/+Ev6eNGkSLV68mIjUH/tk2yImJobWrVtHRET/+9//yMvLi+7cuaN3bOakrh4/efIk+fj4UFFREV25coUaNWpEf/75JxERffTRR6TvqVNUVJSQg5cvXyZXV1eF46R8nfree+8pbHtfX1968OCB8LmsriMiSk1NpV69elUql4z8vD/44AMaMWIESSQSevbsGYlEIvrmm2+ISHOdzyozd/506dKFFi1aJLyW5YDyOdapU6eoT58+wrHkxIkTFBwcTEREa9asoW7dulFpaSmVlpZSly5dqHPnzkSkuK/fvn2bPD09q4ypqrrDycmJMjMziYhozpw5VL9+faH+nzJlCk2fPp2IiL744gvq3LkzvXjxgsRiMfXp04eWLFkilE/TvtGsWTMhhqFDhwrryNB6UlvmzgPlc7+YmBgqLi6m8vJyiouLox07dhCR+mOk8rlFYWGhkCslJSUkEono7NmzRKT+uKPLtt2+fTuNHz+eysvLiYho69at1LdvX2H+hm4/db+ZVJ0nTZo0SViutufJo0ePpg4dOtDz58+ptLSUWrZsSQMHDqSysjJ69uwZNWzYkK5evapy28jnrTWtM31YIs/lz1M0rRNtcl3T+ZWm8zJN+5i63yCaYlXlv//9L7Vp04YkEgkRET19+pQkEkmleljdObJYLKaAgAAhLqKKc3YiosTEROEcViwWU69evWjXrl1VrH2mirMZ2+dMTjYWQ1RUFIqLizFs2DAAQExMDHJycgAAGRkZ6N27N5o0aQIAmDJlCubPny/cviXrFST7e/Xq1RqXOWLECAAVY/a0aNECt2/fFuYtM23aNEyYMAETJ07El19+ieHDh6N27dpq57lv3z5MmjQJnp6eAIB69eoJn7Vo0QKdO3cGAGRnZ+PXX38VyglU9N65du0a/vzzT4SHhyM4OBgAMG7cOLz11lsay6KNUaNGwcXFBS4uLnjttdeQnp4urANNXn31VeEWt6NHj2LIkCHw8PAAALz++us4deoUAODgwYO4efMmEhIShGkdHR3x+++/AwBatWqF9u3bA6jYPp9++qneZRkzZoxwBUQqlWLGjBk4deoUiAiPHj1CaGiowrpVJTs7G46OjujVq5fwXsOGDYVeGKryIzY2FgEBAXjttdeQmJiIfv36oWnTpirnf/v2bYwcORL37t2Ds7Mz8vLycPv2bbRp00ZjXImJiULetG3bFm5ubsL6j4yMFPaHtLQ0nD17FitWrAAAPH/+HE5OTsJ8hg4dCmdnZzg7O0MkEuHWrVuIjY3VuGxlQ4YMEZY9btw4fP7555g9e3aVy37ttdfg4uKicd779u3D0qVL4ebmBgBo0KCB8FlcXJxwC9mZM2fw22+/oU+fPgrTZ2dn4+rVq+jTpw98fX0BAJMnT8bixYt1KqO25PdfQP/tm56ejqFDhwpXoCZPnoy+ffuq3OdLS0vV1hPR0dEAKsZ9kMWXkJCAEydOwNvbW+06a9GiBWrUqIHBgwcDqNgXb926BQA4d+6c2rpHU50li0V2G1mbNm3g7OyMBw8eqNw/unfvjmnTpiEpKQmJiYlCDwl1nJ2dMWbMGABAhw4d0LhxY1y6dAlXr17VWOcobzNVzp49i8DAQGHsCUdHR3h5eQmfy9c1aWlpKCgowHfffQeg4kqlbEzHo0ePCnVakyZNFHo1G5v8/kVEWLlyJfbv34/y8nIUFBQgLi6uynkUFRXhv//9r3CVOSAgAJ06dcLJkyfRrFkzk8VubMr1eKdOneDj44OsrCz897//Re/evdGoUSMAwIQJEzB//nydl1FYWIisrCwhB8PCwtCpUyetp+/bty98fHyE10eOHMHq1atRVFQEqVSKvLw8reaTnp6O5cuXw9HREbVq1cKoUaNw5MgR4fzJGHV+dWOO/Hn27BlOnTqFQ4cOCe/JH+/kz7H27t2LX375RThXAoC8vDw8f/4cR48exahRo+Dq6goAGDt2LL766ivdC/1/qqo7WrVqhXbt2gGoODdv1aqVcHyLiYkReu+mp6djzJgxwrF8woQJ+H//7/9hxowZVcbQu3dv1K9fH0DFsejKlSt6l8cQ5sgDVQYNGoSaNWsCqFinsmOxtsfI58+fY8qUKcjKyoKjoyP++OMPZGVloUOHDhqXq+22TUtLw8WLF4XvKg+XYej2U/ebSdV50quvvipMp+15MgC8/PLLqFGjBoCKurtXr17Cb6Hg4GDk5OQgJCSkylitZZ0Zwpx5Ln+eommdaJPrms6vNJ2XAer3MXW/QarafspatGiB8vJyjB07Fl27dkW/fv3g6Kj6ZkBV58gFBQV48eIFhg8fLnzP29sbxcXFOHr0KB4+fCi8/+zZM7sYk9YS7KrRTFahyX54y79Wdz+0crdR2TRVTafL92NiYlCzZk1kZGTg3//+t8JtQbqSb2wjInh5eam8TeyHH35QeK1v99iqyObr7OysUCkoj8ulqZFQPjYiQs+ePbFjx45K37t//77O20cT+ZhWrFiBR48e4fz586hRowbeffddvccWky+PqnidnJxw7tw5nDlzBsePH0eHDh2QmpqqcrDHYcOGYcmSJUhKSgIAeHl5aRWX8nLVrTciwnfffYfWrVtrNR9jjC8mWz9VLVtTzmhDeV8JCQnBmTNnKn3v6tWrKuMzBeUy6bt9lamKWX49q6snNM1P0zq7c+cO3NzchGU4OTmpPSlQ3r+rikXbnFuxYgV+/fVXZGRkYPTo0Rg5cmSlW0GrIiunpjrH0DwEKufi6tWrNd52LB+fqcjHtGPHDhw7dgw//fQTPDw88Pnnn+PYsWN6zdeUMZuTunIYs3zy81J1DJXfRvJ///7773jzzTdx8eJFtGzZEpcvX1Zo9NU3BsA0dX51ZI78kadcx4wePRqLFi2qcjpD46mq7tD2XERTXFXtG9acs+bIA3Xl1/YYOXv2bHh7e+PSpUtwdnbG4MGDjX6eOWvWLEycOFGn+I2tqrpOUxz65rEyW1tn2jJVnivXa+rWiT7ng8rnpprOy3Rd31VtP2Wenp64evUqfvrpJ2RkZGDWrFk4ceIEnJ0rN9PoEgv93xBK586dU5iO6cduxjTTVteuXXHw4EHk5uYCqHgiWffu3RV6uKji4eEh3Auvj2nTpmHUqFEICgpS20ggM2DAAKxbt05Y3tOnT1X+IA0MDISHh4fCPc03b95EXl4eYmNjcfnyZWGwz40bNwr3YCvTpWzbtm2DWCzG8+fPsWPHDvTo0QNAxdWT8+fPA6joPSPrOaZKjx49sHv3bhQVFYGI8O9//1v4rFevXkhPT8fly5eF9y5cuKBVbIbIz89Ho0aNUKNGDTx48EBhvCtN6ycwMBBSqRRHjhwR3nv48KHGXi9FRUV4+PAh4uPj8eGHH6JTp064dOkSAKBOnToKy8rPzxd6Pm7btk24x95YBg4ciKVLlwqVbn5+Pm7evFnldLrkzLfffotnz55BIpFg06ZNQs7ou2x5AwYMwGeffYbS0lIAwOPHj1V+Ly4uDrdv31ZosM7KykJZWRm6deuGgwcPCoPMa3rgh6H1gDJN21fTsnr06IFdu3YJYyKtX78evXv3VrnPu7m5qa0nZGSf3blzBydPnkR8fLzGdaaJprpHU51VFeX1cePGDWH8wcmTJ1c54G55eTm+/vprABV1Sm5uLkQikVHqnLi4OOTk5AgDtmrq+TNw4ECsXLlSGHOlpKQEv/76K4CK7bpx40YAFeNvKF/8kDFFHnp7e8PDwwNFRUUKT3TUtKw6deqgbdu2wva8efMmTp06pXcDjqUo1+NnzpzBgwcPIBKJ0LVrVxw6dAiPHj0CAL175Xh4eCAyMhJbt24FAPz6668Kx8lWrVoJeffkyRON44sWFBTAxcUFvr6+ICKFcQGByscReT169MBXX30FIkJxcTG+/vprrRpwmXrmyJ/atWsjISEBy5cvF95Td7wbMGAAtm3bJvSWlUqlwvivPXr0EM7jysrKFOpieR4eHnj+/HmV9b2mukMXPXr0wNatW1FWVoby8nJs2LBByEtd9g3lMhiznqyKOfJAF+qOkcrrJT8/H02bNoWzszOys7MVzmeNYeDAgVi3bp1wTBSLxcI5rybabj91v5lUnSeZs64zJP9Mvc4MYak817ROtDkf1HR+pem8TBN1v0F03X6PHz9GcXExEhMTsWjRIvj7++PatWtVLl8mMDAQNWvWRGpqqvDeX3/9hdq1a6Nr164KD3XJzc1VGM+Naa/aNZqFhobik08+Qe/evREeHo6TJ0/iyy+/rHI62eDRoaGhet0yk5SUhGfPnuHNN9+s8rvJycl45ZVXEBcXh4iICPTt21fYIeU5Oztj3759+P7774X4xo0bh+fPn6NBgwbYuHEjBg0ahIiICOTk5AhdeZVNnDgRixYt0upBAEFBQejYsSPCwsIQHx8v3Gb1wQcfICMjA2FhYZg1a5bCbQHK+vbti6SkJLRt2xZRUVF46aWXhM9atWqFHTt24PXXX0dERASCgoIUBoA1lWnTpuH8+fMICQlBcnKy0LADVHT7LS0tRXh4eKWBNV1dXfH9999j7ty5CA8PBwBs2bJFY++UgoICDB48WBjMVywWY/To0QCA9957Dz179hQGcP7ss8+QlJSEyMhIXLp0SWFdGcPKlSvh7u4OkUiE8PBwdO/eHXfu3KlyuuTkZOzatQuRkZEaHwQAANHR0ejVqxeCgoJQt25d4fHl+i5b3owZM4QBR0UikbAeldWrVw/79+/HokWLEBERgeDgYMycORNSqRShoaGYN28e4uPjERkZKXSzVmXq1KmYMGGCXg8CUEXT9tW0X/bp0wf//Oc/ERsbi7CwMBQWFmLVqlUq93lN9YSMRCJBZGQkEhMT8fnnn8Pf31/jOtNEU92jTSzqKK/72bNnIyQkBJGRkfj666+rfHS47EpeREQE/vnPf2LHjh2oU6eOUeqcevXqYc+ePZg5cybCw8PRtm1bhYFx5c2YMQPR0dFo3749wsPD0aFDByGXPvvsM5w7dw7BwcEYNWoUunXrpnIemuokfYwaNQolJSUIDAxEnz59FHq9VnXs2759O3bu3ImIiAgkJSVhw4YNRq+nTE25Hn/77bfx7bffonbt2sLAuR07dkTbtm3x4sUL4TYgXW3duhX//ve/ERoaijlz5ig0Lk6cOBGPHz9GUFAQRo0apfHWqLCwMAwbNgwhISGIjo6utL6VjyPyPvzwQ7i4uAgDTw8YMABDhgzRqzysgrny5+uvv0ZmZiZCQkIgEokqNZbKxMfHY9myZUIdHBISIjxwacKECQgICEBwcDA6deqk9gKfl5cXRo0ahfDwcI0PAtBUd+hi4sSJaNu2rXAs9/f3F84VdNk35Bm7nqyKufJAW+qOkcrnFnPmzMGmTZsQHh6OmTNnqj3u6GvkyJEYM2YMunbtioiICIhEIq16Mmu7/dT9ZlJ1nmSqoTfUxaXtebIyU68zQ1gqzzWtE23OBzWdX2k6L9NE3W8QXbffH3/8gZ49eyI8PFx4GIry0CiaODs7Y+/evdi0aZPwQAzZrabbt2/HzZs3ERoairCwMAwePBhPnjzRet7sbw4k67vHTCozMxMjRozAjRs31N6nzJi9GTNmDEQikXDyy6yPg4MD8vPzFZ7QY2/u3LkDkUgkPPGWMV0UFRUJY9189tlnOHjwIH788UcLR8VsBecPAzgPWPXAec7slV2NaWatxo8fj8OHD2PDhg3cYMYYY4zZkJkzZ+L06dMQi8Vo3Lgx1q9fb+mQmA3h/GEA5wGrHjjPmb3inmYW8ujRI5X31vfs2ROffPKJBSKqsGHDBpVd/levXq13t3tTk38imbxBgwYJT6KR991336Fly5ZmiKx6OHDgAGbPnl3p/VmzZglPZDOEuu07evRovPPOOwbPX1/z58/H999/X+n9u3fvVnpyYEhICLZv326u0KqtAQMGCOP4yNSrVw8ZGRlGmf+kSZNUjpVx9uxZuLu7G2UZulJ3LFGVh4B11+W2hNc7M4St5Y+p69bqytbywFKs9TcT0w7neQV1vxv4d6lt4EYzxhhjjDHGGGOMMcaU8L2CjDHGGGOMMcYYY4wp4UYzxhhjjDHGGNNBamoqGjRoAADYvXs34uLi0L17d9y7dw8AcOPGDSQkJCAuLg5Hjx61ZKiMMcYMwLdnqiCVSpGbm4s6derAwcHB0uEwEyMiFBUVoXHjxlbzoAbOweqD849ZEucfsyTOP2Zp+uagRCLBq6++ijt37uDChQuIj4/HTz/9hIsXL2Lr1q1Yv349Bg8ejKVLl8LHxwd9+vTB6dOntZo352D1wXUgsyRrzD9rxU/PVCE3Nxd+fn6WDoOZ2R9//IGmTZtaOgwAnIPVEecfsyTOP2ZJnH/M0nTNwdTUVLz66qtYvnw5cnJyEBQUBFdXV3Ts2BHTp08HUJFLAQEBAAAvLy/89ddf8Pb2rjSv0tJSlJaWCq/v37+P4OBgA0vEbAnXgcySrCn/rBU3mqlQp04dABUJ5OHhofG7YrEYhw8fRmJiIlxcXMwRnk3FA1h/TM+fP4efn5+w3a2BLjloCGvcNsqsPUZD4yssLLT7/LP2bShTHeO05/yzle0pzxZjBvSP257zTx1r38bWHh9g+TpQIpFg165dSEtLw/Lly5Gfn6+QKxKJBEBFjx0ZT09P5OXlqWw0W7x4MT7++ONK72/YsAE1a9bUpTjMxpSUlGD8+PE2UQdaa93AcenG2n8DWytuNFNB1hXVw8NDq0azmjVrwsPDwyp2CGuLB7CdmKypC7IuOWgIa9w2yqw9RmPFZ8/5Z+3bUKY6x2mP+Wcr21OeLcYMGB63PeafOta+ja09PsDydeC2bdswZMgQ4XamunXrorCwUPjcyckJABRudyooKICXl5fK+c2aNQvvvvuu8FrWkDdw4EC1OSgWi3HkyBH07NnT6raTtcZmjXEVFhZi/PjxNlEHWmvdwHHpxtp/A1srbjRjjDHGGGOMMS1cu3YNly5dwrZt25CTk4PVq1fj+vXrKCsrQ2ZmJsLDwwEAvr6+uHXrFho2bKi2lxkAuLm5wc3NrdL7Li4uVf7Y1uY7lmKtsVlTXNYSB2NMM7trNDt+/DhSUlIglUoxdepUNGjQAB988AEcHR2xdu1ahIWFWTpEk/CfuR8AkJOSaOFIGLMc2X5wZ0k/C0fCbB3nErN2nKNME84P01m6dKnwd1RUFNauXYudO3eiS5cuqFGjBrZs2QIAWLhwIcaMGQOJRKLy9ktWPfC+yCxJln8A56Ah7KrR7Pnz51i+fDl+/PFHuLq6AgA6d+6M/fv3o6ioCJMmTcKBAwcsHCVjjDHG7BmfpDJLkM87Zh6ZmZkAgKFDh2Lo0KEKnwUHB+PkyZOWCIsxxpgR2VWj2dmzZ+Hu7o7+/fujZs2a+OKLL+Dk5IR69eqhXr16yMvLUzmd8lNrZOMSiMViiMVijcuUfV7V90zNzYkU4rB0PPKsPSZriosxZjuUf6By4whjjDHGGGP2xa4azR4+fIibN2/i3LlzSE9Px9y5cxUGMHR2dkZZWZnQC01G3VNrDh8+rPVTa44cOWJY8AZaFqMYh6XjUcVaYyopKbF0GIwZTWpqKqZOnYrHjx9j9+7dWLlyJdzd3bFlyxZ+nDRjjDHGmA17+PAhBg0aBBcXFzg5OWH79u24detWpeGIHjx4gFGjRqG4uBiTJ0/Ga6+9ZunQGbNZdtVoVrduXXTs2BGurq7o3r075s6dq/AI1fLy8koNZoD6p9YkJiZq9fRMa3gSS+i8QwCAS//qZhXxyLOWdSRPPqbnz59bOhzGjEIikWD37t3w8/NDeXk5VqxYgZ9++gkXL15ESkoK1q9fb+kQGWOMMcaYnry9vXHq1Ck4Ojpi8+bN+Oqrr3DkyJFKwxEtXboUH3zwAbp06YL4+HgkJSWhRo0alg6fMZvkWPVXbEd0dDSuX78OIkJWVhaCg4NRXl6Op0+f4o8//lD7qGc3Nzfhsbryj9eVPV2lqn+6fNdU/0olDiiVOFhNPNa4jjTFZEypqalo0KABAGD37t2Ii4tD9+7dce/ePaMuhzFlqampePXVV+Ho6IicnBwEBQXB1dUVHTt2xOXLly0dHmOMMcYYM4CTkxMcHSt+whcVFaFly5bCcEQvvfSSMBzRhQsX0K1bNzg7OyMqKgpXr161ZNiM2TS76mnm7e2NQYMGoXPnznBwcMDGjRtx//599O3bFw4ODvjiiy8sHSKzc9zTh1mKRCLBrl27kJaWhuXLlyM/P1+hp6xEIlE7rSHjOmrDGsc1VEU+TuVxIlWRfUd5elMz5vq09m3CmK74FnXt8VP9GLNNWVlZeP311/H06VMcPnwYO3fuFD6TDUckFouFxjVPT0+Dx/a21nM5jksz+XNV+W3K43rrxq4azQDgjTfewBtvvCG8btmyJc6cOWPBiFh1Iuvps3z58ko9faZPn652OlM3WqhjLRW6JrrEqE1Dh7EZug6NFeu2bdswZMgQ4QSpbt26Qh4BFVcm1THGuI7asMZxDVU5cuSIME6kpicuy74jY+6nMxtjffKYjsye8IUrxlh1IBKJcP78eezatQsLFy5UON+TDUfk4uICqVQKR0dHFBQUqL3jStdzQGs9l+O4VJM/V5U/T+VxvXVjd41mjFmKIT19zNVooY6lK3RtaBOjNg0dpqLvOjTWAevatWu4dOkStm3bhpycHKxevRrXr19HWVkZMjMzER4ernZaQ8Z11IY1jmuoinyckQuPAQCuzuul8B3Z+JGqKH/XVIy5PuVPtBmzdfpcuDL3RStTXqxS7v1a1XJUXWiyt4tp2s6LMVsh/1A7T09P1K5dWxiOqKioSGgci46OxvHjx5GQkICff/4Zy5YtUzk/bc8BrfVcjuPSTP689eq8Xjyut5640cxGcZd662NITx9TN1qoYy0Vuia6xCg7MJir8QIwfB0aq9Fi6dKlwt9RUVFYu3Ytdu7ciS5duqBGjRrYsmWL2mnd3Nzg5uZW6X1jj/lnijEETUE2TqTsb3my99VNZ07GWJ+2sD0Y04a+F64sddHKFBerlHu/Atr1llX1HXu5mFYV7mnBbE1WVhamT58OJycn1KhRAxs3bkROTk6l4YhmzJiBUaNGYc6cOZg0aRLc3d1Vzk/Xc0BrPZfjuFSTP2+Vj8PFxQXl5eWWCMkmcaMZY0ZiSE8fczVaqGPpCl0b2sSorqHDHPRdh6aINTMzEwAwdOhQDB061OjzZ4wxa6PvhStzX7Qy5cUqVT1hNV1EUnWhyd4uplWFe9syWxMTE4MTJ04ovOfr61tpOCJfX1+baPxmzBZwoxljRmJITx+mO1lvS4B7XDLGWHWn74UrS120MsX8VfWE1bQMTRea7OVimjbzYIwxxjThRjPGTIB7+jDGGGPmwxeuGGOMMWYK3GjGGLMp8j3MDJmee6cxxph94gtXf+NjHmOMMWYYbjRjjNktvoWTMcYYY4wxxpi+uNGMMcZYtSZrXHVzIpVPn2OMMcYYY4xVT9xoxhhjjMmpeKJc5QG1GWOMVeCe3IwxxqoLbjRjjDHGGGOsmpJvAFPucctjojHGGKvuHC0dAGOMMcZsS0FBAWJiYlC7dm1cvXoVALB7927ExcWhe/fuuHfvHgDgxo0bSEhIQFxcHI4ePWrJkBljjDHGGNMZ9zRjjJmVpW7pMPSpm4yxv9WsWRP79+/H+++/DwAoLy/HihUr8NNPP+HixYtISUnB+vXrMXv2bHz11Vfw8fFBnz590L17dwtHzhhjjDFWgW81Z9rgRrNqgLvWM8YYMyYXFxc0aNBAeJ2Tk4OgoCC4urqiY8eOmD59OgAgNzcXAQEBAAAvLy/89ddf8Pb2VphXaWkpSktLhdeFhYUAALFYDLFYrHeMsmkNmYe+3JyoUhza0CVm2TIsUT5l+q5ra4idMcYYY0wTbjRjjDFWrfCFBOPLz8+Hh4eH8FoikQAApFKp8J6npyfy8vIqNZotXrwYH3/8caV5Hj58GDVr1jQ4tiNHjhg8D13Jjwl14MABnafXJmbZMvSZv6nouq5LSkpMFAljjDHGmHFwoxljzOIMbcTgWy8Zs6y6desKPcQAwMnJCQDg6Pj30KkFBQXw8vKqNO2sWbPw7rvvCq8LCwvh5+eHxMREhYY4XYnFYhw5cgQ9e/aEi4uL3vPRR8UTWCtcnddL6+l0iVm2DF3mbyr6rmv5nGGMMcYYs0Z212iWmpqKqVOn4vHjx9i9ezdWrlwJd3d3bNmyBU2bNrV0eBbFvSsYY9UF13fmFRAQgOvXr6OsrAyZmZkIDw8HAPj6+uLWrVto2LChyl5mAODm5gY3N7dK77u4uBilsctY89FFqcRBYfm60iZm2TLMXTZNdF3X1hQ7Y4wxxpgqdtVoJpFIsHv3bvj5+akdlLi6CJ13SOGknTFbIt9zLCcl0YKRMMbU6du3L7KyspCdnY3XX38db7/9Nrp06YIaNWpgy5YtAICFCxdizJgxkEgkKm/BZIwxxhhjzJrZVaNZamoqXn31VSxfvlztoMSqGDIIsbEGGtb1Vg7lAYBlr9vNP4iUKMDNkdROa+6Bdy05GLM68jFZU1yMMePipyKZjqqxtIYOHarwOjg4GCdPnjRXSIwxxphdu3DhAqZNmwYXFxc0adIEW7duRVpaWqW7q27cuIGJEyeivLwcKSkp/PRqxgxgN41mEokEu3btQlpaGpYvX652UGJVjDEIsaEDDes6aLDyAMDy0wNASpQU6lhq0GBLDMZclSNHjvBAxDaCe08ybfD4dsweyOo7buhljDEmz8/PD8eOHYO7uztmzZqFvXv3qry7avbs2fjqq6/g4+ODPn36cKMZYwawm0azbdu2YciQIcKgw+oGJVbFkEGIjTXQsK49zeS/L8/NkZASJcWHmY4olapuYDD3oMGWHIxZHfmYnj9/bulwqi1u4GCMMcasn6ZxInkMScbMx9fXV/jb1dUV2dnZKu+uys3NRUBAAADAy8sLf/31l8pxRRljVbObRrNr167h0qVL2LZtG3JycrB69WqVgxKrYoxBiA0daFjToMGqTkaq6nFTKnVQ+x1LNVxZYjDmqri4uKC8vNzSYTDGGGOMmYyxGrb4Yhdj1uHu3bs4fPgwlixZgsePHwvvy+6ukkr/vuvI09NT7cN4tB2myBqH2wEMj0s2xJEh81DFWtaXcvl4iCL92E2j2dKlS4W/o6KisHbtWuzcubPSoMSMMcYYY4wxxpgtKiwsRHJyMjZv3gyJRKLy7irZ3VcAUFBQAC8vL5Xz0nWYImscbgfQPy5dh0jSlaXXl7ry8RBFurGbRjN5mZmZACoGJFYelNja8BU7xhhjjDFmTXgcUcasU3l5OYYNG4a5c+ciMDAQYrFY5d1Vvr6+uHXrFho2bKi2lxmg/TBF1jjcDmB4XLoOkWSuuIxFuXw8RJF+7LLRzF5wgxqrblTlfOi8Q5UedMGYMXAdyxhjTFf89EJmSampqTh//jxSUlKQkpKCyZMn4+233650d9XChQsxZswYSCQSlT3JZHQdpsgah9sB9I9L0xBJxmDp9aWufDxEkW640awaUjWuBQ/iykyFc4sxxhhj9oKfXsgsKTk5GcnJyZXeV767Kjg4GCdPnjRXWIzZNW40Y4wxxqrAvdIYY4wB/PRCxhirbrjRjDHGGGOMMcZ0YO6nF8qzlifzqWKtsVkqLtnTC1Ut19rWEWNMNW40Y4wxxtTQpYcZ34rMGGPVgyWfXijP0k/m08RaYzN3XLJxeVU9mZGfXsiYbeBGs2qMbzdizH5oOzBxdWZNdZ58LNzIxhgzhDXVbdWBpZ5eKM9ansynirXGZqm4ZE8vVPVkRvnGVsaY9eJGM8aMhBstmCVpOzAxY4wxxvRn6acX6vodS7HW2Mwdl+zpheqeRMkYs37caMaYkXCjBbMkbQcmZowxe8QXrixPucebvfai5acXMsZY9cKNZowZiSGNFvoMAmsM5hgUVXkAVNlrrad3JIX/tRH4r33/tyz13zFWmQ1dh8Ze91UNTKyKqfPPnIPv6ppfCtPqkWuqaFNO+Th1XS/GXJ88CDGzF3zhyrbwGJBMhnOBMWbtuNHMhphjzAo+cBlOn0YLQwaBNQZTDoqqPACq7LWuUqKkVX9JB6oGZDWEvuvQmIPAajMwsSrmyj9zDL6rb37JMzTXtMkt+Tj1zUVjrM/qNAgxjyNn3/S9cGXui1amuIigy8UC+eWqms5YFw+Ul1VpORqeKKjtfPnCAWOMMXPgRjPGjEjfRgt9BoE1Bl0HRZUNZiqjalDTqqbRlZsjISVKig8zHVEqdTBoXvK0iV0bhg4sa6xBYLUdmFgVU+efqQbfNTS3lBkr1+RzS90AwKpi1zYnjbk+eRBiZm90vXBlqYtWxryIoMvFAvlGek3TGeNClaYLApqeKKgtvnDAGGPMHLjRjDEjMaTRwpBBYI1B2+XIBjOVCfjwMADVvTb+7tVhnIauUqlDpeUbwtjrVd9tZaw4tB2YWBVz5Z+x52fMfFCYr4G5Jl9GdQMAq5q/ruvGGOuTByFm9kSfC1fmvmhliosIulxAUNWoL8+YF6o0XQjQ9ETBqvCFA8YYY+bEjWaMGYkhjRbMOtnS7craDkzMTM8ct9IzZiy2VM9pou+FK0tdtDJk/srbTJdGflWN+qoY40KVpvJpeqKgLvPnCweMMcZMjRvNGDMSbrRgjFVnd+7cQXR0NEJCQgAAu3fvxvHjx/nphcws+MIVY4wxxkyBG80YY9WSqt4V9tLjgjFL6dy5M7799lsAFT1/+OmFlXE9Yxp84YoxxhhjpmBXjWYXLlzAtGnT4OLigiZNmmDr1q1IS0uzyqvcfPsOY4ypp+lJh1x/Wq/Tp08jPj4e8fHxSE5OtujTC5WfsCf/pEBTPzFP07I0PTVQiNWx6icLGvL0QWPPR9+nGfKTC5kMP92WMcaYtbKrRjM/Pz8cO3YM7u7umDVrFvbu3ctXuRljjDEz8PX1xc2bN1GzZk1MmDAB33//vcJA6pZ6eqHsCXvyTwo05Il92tC0LG2eGih7cqGpnz5ozPkAuj/NkJ9cqJ499Ei0hzIwZm0KCgrQs2dPXLt2DefOnUNoaCh2795dqZPIjRs3MHHiRJSXlyMlJQXdu3e3dOhMA64vrZtdNZr5+voKf7u6uiI7O9vkV7n1vboqfxVa3Tw1fUftfP/v6rTsf33JYpB/spI+TziSn5c1XVGWj8ma4rIHttYLyNbiZbbNnvNNfkD1wYMHY/Pmzahdu7bwubmfXqj8hD1jHM+0pWlZmp4aKItZ9uRCUz190Njz0fdphvzkQsaYtbCVRouaNWti//79eP/99wGoHwph9uzZ+Oqrr+Dj44M+ffpwoxljBrCrRjOZu3fv4vDhw1iyZAkeP34svG/Kq9y6Xl2VvwqtTHa1V9N3qiK7Sq0vVTEYehVa13VkDkeOHOEr3YxZAU0nq7ZyIlvdFRUVoU6dOgCAkydPol+/fli3bp3Fn14om4/8kwBN/cQ8TcvS5qmBsicXmvrpg8acj2weusyHn1zI7PlCAmOm4OLiggYNGgivc3JyVHYSyc3NRUBAAADAy8sLf/31F7y9vQ1efui8Q8he+A+D58OYLbG7RrPCwkIkJydj8+bNkEgkClcxTXGVW9+rq/JXoZXJrvZq+o46bo6ElCipcJXamAzpaabPOjIl+ZieP39u6XCYleGTeMZ0d+rUKcyZMwc1a9ZE8+bNkZKSgho1avDTCxkzEX2OVZY+vmmzfL5Qwpj28vPzVQ6FIJX+3YHC09MTeXl5KhvNtL3jSn7MTWu6S8fQO5pMNd6prnEZa5xSdfOVzZvvttKPXTWalZeXY9iwYZg7dy4CAwMhFotx/fp1s1zl1vXqqvxVaGUBHx7+v7/0b/SSXaU2JkMbvIzVa8CYXFxcUF5ebukwGGPM5vXp0wd9+vRReG/o0KH89ELGGGPMROrWrauyk4ijo6PwXkFBAby8vFROr+sdVylRUpOPC6oPfe9oMvV4p9rGZczxRVXNV3nefLeVbuyq0Sw1NRXnz59HSkoKUlJSMHnyZLz99tt8lZsxPfCVXsYY+1t1rhOrc9nNTV1PLEv3EGOMWaeAgACVnUR8fX1x69YtNGzYUG0vM0D7O67kx9z8+aPepi2UDgy9o8lU453qGpexxilVN1/ZvPluK/3YVaNZcnIykpOTK73PV7mNg0+aqyc+UWemVF3yi29JYpbAOcWY/ZA/juSkJFowEmZpffv2RVZWFrKzs/H666+r7CSycOFCjBkzBhKJRGVPMhld77gqlWoeb9NS9L2jydTjnWoblzHHF1U1X+V5891WurGrRjPGGGO2TVPjUnVpYGOMMcYYU0fVLXzKnUSCg4Nx8uRJc4XEmF3jRjPGmMG4MYMxxhgzPj6+MsYYY5blWPVXGGOMMcYYY4wxxhirXrinGWOMMcYYY8yqyXrduTmRwhPhGGOMMVPiRjPGGKuC/O0xPKA2Y4wxpj19bzHlW1MZY4xZA749k+nMf+Z+PpFhjDHGzIiPvYwxxhhj5sc9zRhjjDHGtMQ9TxljjDHGqg9uNGOsGpL96MtJSVT7GWPGEjrvEJbFVPxfKnGwdDhWT7YPqmqQ8Z+5n8fzYczO8HGXMcYYs158eyZjjDHGGGOMMcYYY0q40YzpjcdXYYwxxhhjjDHGmL3i2zMZqwa0adzkBlBmDDzek2mo2j9D5x1C9sJ/WCAaxhhjjDHGqgduNGOMMcYYY8wE+EKCaWkaA5IxxhgzBm40MzPuzcMYY8xYVB1T+Mcj43MNVh2EzjsEgB8uwxhjzLR4TDNmVDzOGWOMMWa7ZMdxfY7lFY0YjDHGGGP2g3uamRB3GWeWxI2XpqHNfl0d9n1dxslzczJ1NIwxZnv4OG08mtalPR+LGWOMmV61aDSbMWMGzpw5A39/f2zcuBEuLi6WDqna8J+5H25OhGUxlo7Ecsydf7qchIfOO4RlMXyLgz7+bhAiYR2WShxUfseSJ+xc/9kuY/2gtvSYSsbMQWvYp8yJG1UMZ646UNsLCfLHDD7u2j8+BjNL4xxkzHB2f3vmL7/8gvv37+PkyZNo06YNvv32W4Pmp80tC8rfMeRWB1ugqnyqyhs671CV3zF1jOZm7PxjTBecf8zSbC0H9T1ea5pGfp7WfB5g7fHpw9byj9kXzj9maZyDjBmH3fc0O3PmDBITEwEAvXv3xqZNmzB8+HCD56tNLxOmmvJJuT33HDB2/tnbD5rqwlI5zvlXPVnTdjLVMdhYLHn80bcHoCHbV9dl6tp7ytqYow60x3MXe2Mvx2DGdMU5yJhx2H2jWX5+Pnx9fQEAnp6eyMvLq/Sd0tJSlJaWCq8LCgoAAHl5eRCLxQrfdS4vVnwtJZSUSOEsdoREavlGM2uLB9A+pidPnpguhv/bbrJliMVilJSU4MmTJ3jx4gUAgIiMvlxt8g/QPgeV889Q1pgvyqw9Rl3iU5XjRUVFAOw7/6x9G8pUhziVc9CU+QeY7hiszfFCvp53cXFRyF/Z9MrzU/Ud5WXL02Y+2lA+NumybWXTtl989O9Yq4hT31hVrXfn8mKFnFT1HVls52d1V3jfVvNPnq7r0drrGWuPD9A/Rns6Bj958kShfrMmynVvVXSp100Zl6ni0DRfW6oD5Y9Ppt5WutA135RpOu6bMy5T559s3ub6DWxv7L7RrG7duigsLARQUQl4eXlV+s7ixYvx8ccfV3q/efPmWi1jhGEhGp21xQNoF5P3cpOHoXEZRUVF8PT0NOrytMk/wPAcNIQ15osya49R2/iqc/5Z+zaUsfc41eWgKfIPMN0x2NDjhfL0quanzTK0mY8+8Rh7Wn3Lp8s0spz0/kT36W0t/+Tpsx6tvZ6x9vgA/WK0p2OwrxnOmc3NHL8DtGGqOMydf4AJj8Ea6nlbZg05aMoYzH0MticOZOdNi1lZWVixYgW2bt2KRYsWoXnz5pW6pSq3sEulUuTl5aF+/fpwcNB8BauwsBB+fn74448/4OHhYZIy6MLa4gGsP6Y6deqgqKgIjRs3hqOjcYf50yb/AMNy0BDWuG2UWXuMhsZHRHaff9a+DWWqY5ymzD/A9MdgTWxle8qzxZgB/eO25/xTx9q3sbXHB9hOHWjKY7A1bydrjc0a47KlOtAa1x/AcenKXL+B7Y3d9zQTiUTw8fFBfHw8XnrpJUyfPr3Sd9zc3ODm5qbwXt26dXVajoeHh1XtENYWD2DdMZmqdV2b/AOMk4OGsMZto8zaYzQkvuqSf9a+DWWqW5ymvLpormOwJrayPeXZYsyAfnHbe/6pY+3b2NrjA6y/DjTHMdiat5O1xmZtcdlaHWht60+G49KNqX8D2xu7bzQDgE8+sdM+pMwmcP4xS+L8Y5bGOcgsifOPWRLnH7M0zkHGDMf98BhjjDHGGGOMMcYYU8KNZgZyc3PD3LlzK3VrtRRriwfgmKyZLawHa4/R2uOzBrayjjhO+2KL68kWYwZsN25LsPZ1Ze3xAbYRo6lZ8zqw1tisNS5bYa3rj+PSjbXGZe3s/kEAjDHGGGOMMcYYY4zpinua2RgHBwc8ffrU0mEwC7CFbb9u3boqx05IS0vDuXPnhNeZmZkYOnSoqUMzqqioKBw/fhwAMH78eGRkZACoXDZbZgv5Zire3t64c+eOpcMAYJxYVq1ahQcPHhgnICObN28eXrx4YdJlqNtf9TVmzBisWrXK8MD+z507d4w+8Ly6/VebOtrYNm/ejBs3bmj13ePHj+PgwYMmjkg9zkf9iUQiFBUVWToMnXLozp07WLdunYkjMhznpXHNmzcPb7/9tqXD0Ojp06dYsmSJWZfJeWZauvxG2Lx5MwYOHGjQ8owxD/Y3bjRjjBnNpEmT8P7772v8jvJBIyoqCjt37jR1aBqVl5frPe2GDRvQtWtXAPbVaMbshzU3mn388cdGPUmval+W31+rI23qaGOzpUYzzkf9ZWVloU6dOpYOwy4bzTgvqx9LNJpxnpkW/0awbdxopocLFy4gNjYWCQkJGD58OMRiscLnx48fh5+fH7p06YLu3bubLI7MzEw0adIEtWrVgre3N3766Sfhs7Vr18LT0xO1a9dGbGwsHBwc9FrGihUrEB0dDZFIhOjoaJw9e1b47MyZMxCJRAgLC8PYsWMRERGBtWvXIjY2Fh06dMBLL72E6OhohIWFYc6cOWZbL8ru3LmDBg0aoEuXLujSpQseP34sfCaRSDB27FjEx8db/VUneZmZmYiLi0N4eDhiYmJw+vRp4bP169ejdevWaNu2LVJSUrTa9qpy+sGDBxgyZAhiYmKEbQgAJ06cQMuWLZGXlwcAePPNNzFhwgQAilfvzp07h3bt2kEkEiE0NBRr167FgQMH8MMPP+CTTz6BSCTChg0bcPz4cYhEIgB/97iYO3cu2rVrh3r16iEkJATJyckQi8XYu3cvgoKCEBERgRkzZmjVC2fTpk0QiUSIiIhAVFQU7ty5IyxnxowZaNu2LdasWaO2vMDfuR4aGop//vOfwonAhQsX4OHhgeDgYHTu3LlS2XJyctCxY0dERERUmqctMVa+aao7u3TpgunTpyM+Ph4tW7bEpEmThM+UryROnz4d8+bNA1CRc0OGDEH//v3RunVr/OMf/8DVq1fRq1cvtG7dGsOHD4dUKtVYvh9++AFBQUFo2LAhmjZtimfPngmx5eTkoF+/foiOjkZ4eDjWrFkjTOfg4IA5c+YgMjISrVu3xvbt24XPLl68iG7duiEqKgqRkZHYvXs3gMo53qpVKxw4cKBSLOHh4fjggw8U4pTlU+3atVGrVi20a9dOiNPBwQGLFi1CTEwMmjdvjk2bNgEA5s+fj9zcXAwdOhQikQhZWVn4z3/+g/DwcCGn9+7dq3H96OLs2bPo1KkTIiIiEB4ejr1792L69OnCcSQhIQHZ2dkAIGzj+Ph4iEQiPHr0SO18Dx06hLZt2yI8PBydO3fGtWvXAFQcb0NCQjBu3DiIRCLs2bOn0v5aWFiIKVOmICEhAQ0bNsS3334LoCKvXn/9dXTv3h2tW7fG4MGDUVZWBgA4evQoYmNjERkZiZCQEHz11VdGWT+q6iOZuXPnolGjRnB3d0e3bt2EbasulwBg//79iI6ORkREBEQiEc6fP6+wPCLCjBkzMGDAAJSUlCjU0Zs3b0aPHj0wfPhwhIWFISoqCr/99ptCPK1atUJ0dDTmzJkDf39/lWWS7dchISGoW7cuIiIihLzasGEDMjMz8c4770AkEuHAgQO4cuUKOnXqhLZt2yI4OBgLFiwAUNHosm7dOmzfvh0ikQjz58+v1Avv2bNnQh3z/PlzDB06FMHBwYiIiEBiYmKl2KwxH1UdP4YPH45Ro0aZNR/VHZ+GDRsGDw8PeHl5ITg4WFjWH3/8gYYNGwoxARX70GeffQZAsXejv78/PvroI8TGxqJ58+bCNgaAGzduCPkyePBgJCYmYvPmzQAqfkQHBwcL55bK+Sxbb7GxsYiOjkaDBg0QHh4uxK8qh8rLy9GrVy9ERUUhJCQEI0aMQHFxMYCKbZ6dnQ2RSIQBAwYIsWdlZQnLk+8Zs2DBAgQFBUEkEkEkEuHu3bt6rXtj5qX8cbVLly6IjIw0KC+BiuNxWloaAMPqyefPnyMmJga1a9fG1atXtV4/X3/9Ndq3b4+2bdsiISEBv/zyCwDFOis4OBhxcXG4du0aBg0ahKCgICQmJuLZs2cAKs4NXnnlFXTr1g1t2rRB//79MXXqVMTHx+P7778XzgskEgnef/99hIaGIjQ0FG+99RbKysqQm5sLHx8flJSUCHGNGDECa9euBVCR7wsXLkT79u3h7++PtLQ0LF68GFFRUQgICBByBqioLzp16oR27dohJiZG6F11/PhxhIaGYsqUKYiIiEBISAgyMzOF7V5UVASRSISoqCit1508U9V/n3/+OWrVqoVatWqhYcOGwvaxVJ7JFBQU6JRvn3zyCUJCQhAWFoaRI0eioKAAANCkSRPk5uYCAIYMGYK4uDgAQGlpKerXr4/S0tIqj58AMGPGDMTHxyM5ORk//PBDpd8IgPpcV6Zpn+jWrRsGDBiA4OBgJCQkKJxXPHv2rFKMsvXk6uqKVq1aVSr/vHnzMHToUPTv3x/BwcHo1q2b8JsPAD799FPExMSgbdu26N27t971oM0hprPc3FwqKSkhIqKZM2fS7t27FT7PyMig9957zyTLBkD5+flUWlpKjRo1oq5duxIR0fjx48nT05OKioroypUrVK9ePZo2bRoREUVGRpK+m/rRo0fC32fPnqXAwEAiIiotLaWmTZvSsWPHiIjo2LFjBIC+/fZbKikpocTERBoxYgTt3r2bxGIx9erVi+bOnWuy9aLJ7du36ZVXXlH5WVpaGv3rX/8ioop1eObMGXOGphP5be/n50cHDx4kIqKTJ0+Sj4+PsO0bNWpEf/75JxERffTRR1pte1U5nZiYSMePHyciErbhrl27iIho0aJF1L9/f9q5cydFRETQ8+fPiYho7ty5Qt4NGDCAduzYISwjLy+PiIhGjx5NK1euFN7PyMigiIgIIqrYVrI8ysrKoq5du1Lr1q1pwYIFtHbtWvLy8qLr168TEdHGjRsJAN2+fVttuTIyMsjf359yc3OJiKi4uJiKi4uF5WzZskX4rrryynL9yJEjRER06NAhAkAZGRmUm5tL8fHxtGfPHpo5cyZ17txZoWxTp06lRYsWCa+fPHlS5bawFqbIN011Z+fOnWngwIEkFouppKSE/P39hf1ROWfee+89mjt3LhFV5Fzz5s0pLy+PpFIpJSQkUPv27amwsJDEYjFFRETQvn371Mb08OFD8vLyou+++45GjhxJ69evJwD02WefUXl5ObVr107IueLiYgoLC6MLFy4I62jOnDlERHTr1i2qV68e3b59m/Lz80kkEgl59/jxY/Lz86N79+4p5DgR0Y8//kitW7dWiOXXX38lIhJiuX37thDLTz/9RCUlJVRcXEwNGjSgxYsXC7F8+umnRER0/fp1ql27NonFYiIiatasGV26dEkoc3h4uLBuJRIJ5efnq10/unjy5Ak1bNiQTpw4Icz7yZMnCseR1NRU6tWrl/BalmeayNbL5cuXiYho27ZtFBQURFKplDIyMsjBwUHYd9Xtr7L89fPzo/fff5+IKvIqJiaGiouLqby8nOLi4oQ6Ky8vj8rLy4VyvfTSS/THH38I08nno7aqqo8+/fRTGjlyJP3444/k7e1NO3bs0JhL2dnZ1KBBAyE/y8rK6OnTp8J6ffDgAQ0dOpSmTJkilEW+jt60aRN5eHjQb7/9RkREM2bMoIkTJxIR0b59+ygkJIQKCwtJKpXSmDFjqFmzZirLJduvw8PDKTk5mXbv3q2QV507d6Y9e/YI3y8sLKQXL14QEVFJSQmJRCI6e/ZspfiIKo4Jnp6ewuuioiKhjvn+++8pMTFR+Ey5frXmfFR3/DBnPqo6PmVlZVFSUhKVl5fTggUL6N///rfCsnr27CnU20VFRVSvXj3666+/Kq27Zs2a0VtvvUVEFTnr4eFB9+7dIyKiqKgo2rhxIxERXbt2jdzc3GjTpk1EROTh4SHkellZGRUVFVWKW5ZvU6dOpc6dOwvxyLa/cg5JpVIhRqlUSpMmTRLqTfnzDxnl+rJdu3aUkZFBeXl55OnpKRzDiouLhXMfXRg7L2Xr4+HDh1SjRg3hOKBvXhIp7rOG1JOffvopPXr0iEaPHk1XrlzRav2cOnWK+vTpI9QRJ06coODgYCL6u866e/cuERG99tpr1KJFC3rw4AEREfXr14/WrFlDRBV50KBBA+H85NVXX6VWrVoREVHXrl2pd+/eRET0xRdfUOfOnenFixckFoupT58+tGTJEiIiGjFiBK1fv56IiB48eEANGjQQchIArVq1ioiI0tPTqVatWkIe79q1i6Kiooio4tygQ4cOVFBQQEREOTk51KhRI3rx4gVlZGSQk5MTnTt3joiI1q5dK9RpynWfrkxZ/9WrV084D+rfvz81adLEonkmq//Kysq0zrcDBw5QmzZthPJOmDCBJk2aREREycnJtGXLFpJIJNS6dWsKDg6mgoICSk9Pp549exKR5uMnEVFWVhaNHDmSiIgWLFhAO3bsqFRXV5XrL7/8slbfc3V1pWvXrhER0dKlS6uMsaysjFJTU8nT05NOnz5dqfxz586lZs2aCfXm0KFDhWPF9u3bafz48cL22Lp1K/Xt21fjurYXzmZpmbMzvr6+wt+urq5wdKzcYe+7777D+fPnkZSUhGnTphk9huzsbJSVlWHMmDEAKq4Q7NmzB1lZWfjvf/+LJk2aYPDgwcJnr7/+ul7LuXTpEhYuXIgnT57A2dkZ2dnZeP78OXJycuDs7Cx0q+3atStatmyJ+vXrQyqV4ujRo/D29saZM2ewYMECPHv2DE2aNMGxY8dMul7UOX36NOLj4xEfH4+FCxcKV6rPnDmDfv36AQB69+6N06dPIzY21mxx6SM7OxuOjo7o1asXAKBTp07w8fERtn3v3r3RqFEjAMCECRMwf/78KuepnNNlZWU4evQoHj58KLz/7Nkz4YrUzJkz0adPH0ycOBEXLlxAjRo1Ks2za9euSElJQU5ODrp164ZOnTppVb4aNWpg8ODBWLduHYYMGYI333wTvXv3xkcffYTw8HC0adMGADB69GiF3kiq7N+/H8nJyUL5atasKXzm4uKC1157DQBQXFystrw3btyAs7MzevToAQBITExEixYthPUm2/9dXV0r9bJKSEjA+++/j2fPnqFz587CPGyJMfOtqrpz6NChcHZ2hrOzM0QiEW7duqXV/piYmIh69eoBANq2bQs3NzfhNqHIyEjk5OSonfbcuXMIDw/Hw4cPkZiYiJEjR+LNN99EZmYmsrOz8euvv2LYsGHC94uKinDt2jVER0cDqBiTAwBatGiBhIQEnDhxAt7e3vjtt9/Qp08fhWVlZ2ejRYsWQo4DQGxsLG7duqUQS3BwMABg3LhxeOutt4Rpf/31V0ydOlWYn1gsxv3794XXI0eOBAC0adMGzs7OePDgAZo2bVqpzN27d8e0adOQlJSExMREoaenoc6ePYvAwEDEx8cDABwdHeHl5YUdO3Zg9erVKCoqglQqVbhiqY3z588jLCwMYWFhACrK+cYbbwhlb9GiBTp37gwAavdX2VOiHB0dFfbTQYMGCfVCTEyMsC2ePHmCcePG4X//+x+cnZ3x5MkTXL16VeX61Jam+qhGjRpwd3dHYmIiYmNjkZ+fj9OnT8PT01NtLl2/fh29e/cW6kQXFxd4enoK3+nXrx9efvllfPjhh2pjkvUEkv29evVqABVX9l999VVhPxo3bpzasWdk5enevbtwq32rVq3U5tXz588xZcoUZGVlwdHREX/88QeysrLQoUMHtXGqEhERgevXr2PKlCno3Lkz+vbtq/C5NecjoPr4Yc58VHV82rlzJ2JiYvDqq6/il19+QUFBAV68eCEs65///Cc2bdqEpKQk7N69G926dUP9+vVVzn/EiBEAKsZlbNGiBW7fvo06deogKysLo0aNAgAEBQUpnBt0794dycnJ6N+/P/r06YPWrVtXmq8s3xISErB161akpqbCw8ND7fGViLBy5Urs378f5eXlKCgoEHqO6MLDwwMBAQF47bXXkJiYiH79+um1/o2dl7L1kZ6eDh8fH2F/1jcvVdE3L52cnNCgQQOd1s/evXvxyy+/oH379sJ7eXl5eP78OYCKeuqll14CUNELUCwWw8fHBwAQHR2tcLzv16+fcH7SqlUrofdXq1athF6M6enpGDNmjHCMmDBhAv7f//t/mDFjBqZNm4YJEyZg4sSJ+PLLLzF8+HDUrl1bmL9sPN6oqCgUFxcL5woxMTFCHAcPHsTNmzeRkJAgTOfo6Ijff/9diEVW1tjYWHz66ac6rS91TFn/hYeHC+dAkZGROHr0qEXzTMbFxUXrfEtPT8fQoUOFnsyTJ0/Gq6++CgDo0aMH0tPTERISgoiICPj4+OD48eM4e/aswp1S6o6fQMXvS1nv5969ewu9/+VVlevafi8uLg5BQUEAgIkTJ2LOnDmQSCRqY3RxccHFixfh7+8PDw+PSuWXxSyr22NjY3HlyhUAFbeYXrx4Ee3atQMAYTnVATeaGeDu3bs4fPhwpVuuoqKihMaFl19+WeiSa2xSqVRIdk9PT4XELSsrEz6T/a+rsrIyDB48GBkZGYiOjkZhYSE8PT1RWlqq8vuyHyL0fw9kbdasGU6dOgUXFxcAFY0Qrq6uAEy7XpT5+vri5s2bqFmzJiZMmIDvv/8er7zyCgAgPz9fYR3qevCwFupuidP1tlxZTsvfYqmqQayoqAi//fYbatWqhcePH6s8sX377bfx8ssvIz09HbNnz0ZoaCi++OKLKmNwc3ODg4MD8vPz0bx5c0gkEnh6eqKoqAhOTk46lUeTmjVrCj9YZDmrqryXL1+uNK3yen306BEOHz6MgIAAhfdfeeUVxMXF4ciRI1izZg1WrVqlcCuerTI039TVnfLr3snJSejG7+zsrFC/vXjxQuHEVXk6dfPRJD8/H76+vkIZnj59CiKCl5eXwq06VXFwcAARISQkBGfOnKn0+Z07d4Qcl8Wn7qRDfn0qx3L37l0MGzYMK1asEL6jbblXrFiBX3/9FRkZGRg9ejRGjhxZ6VZQY/n999/x5ptv4uLFi2jZsiUuX76s8OPBGORzQRXZerx79y7y8vKEk31A/TqbNGkS+vbti++++w4ODg5o27atSQdIdnNzw9OnT9G4cWMhJ/Ly8jTm0vXr1zXOs1u3bjhy5AimTZum9jxA25zRZt+eNm0ajh49inbt2mnMq9mzZ8Pb2xuXLl2Cs7MzBg8erHbdqtr3ZVq0aIFr167h2LFjSE9PxwcffICsrCyhAV0Va8pHGfnjhznzUdXxqVOnTti5cydGjx6NpUuXYu7cubhx44awrEGDBmHq1Kn4888/sXnzZo31hj659d133+Hnn3/G8ePH0bdvXyxYsEDhooW8qKgotGjRAv369dN4fN2xYweOHTuGn376CR4eHvj8889x7NgxtXGryzknJyecO3cOZ86cwfHjx9GhQwekpqYKjRKGMEZePnr0CHl5eejfv7/Kz3XNS3nmzEsiwujRo7Fo0SKtYtE2zwoLC4Vzvho1aqiNU349xMTEoGbNmsjIyMC///1vpKenq4xFdm4q/1oWBxGhZ8+e2LFjR6Vl3b9/X6/zFX0Zu/6TncvJfucBtpNnmuLq0aMHZs2aheDgYPTo0QM+Pj5IT0/H2bNnhdtzNcUL/H1OCfz9+1L2G1imqlzX9Xuq6HuMVzcdEWHWrFmYOHGizrHYOh7TTIMHDx4I42DJ/3vw4AEKCwuRnJyMzZs3K1QWQEWF4erqCldXV/Tv31/t/cmGCAwMhIODg3AifeLECZSWlkIkEqFr1664f/++cI+x7H5xXb148QJlZWXCFR35FvTAwECIxWJhHLWffvoJN2/eBFDRmFenTh1ER0cL6yY3NxdPnz41+XpRxc3NDbVq1YKDgwMGDx6ssNy6deuisLAQQMW98F5eXmaJyRCBgYGQSqU4cuQIgIqrGQ8ePBC2/aFDh4SxCJTv+9c2p+vVq4euXbsqDEKam5uLe/fuAajodTBy5Ejs2rULycnJePLkSaU4s7Oz0bx5c0yYMAGzZ88WBr/08PAQ7pvXpG7dusKTuAoKCtCqVStcvnxZaJDetm2bwtgqqvTv3x/btm3Dn3/+CQAoKSlRGJ9Cpnbt2mrL26ZNG5SXlwu9LNLT04WrX0DFVZZVq1Zh8+bNqFu3rkLZcnJy4OPjg1GjRmHZsmU2OQCoPvmmb92pTqtWrXDhwgUAFVccjdnwGBsbi8uXL+PFixcoLCzExo0bIRaLUbduXQQGBsLDw0PhCuHNmzcVGtdln925cwcnT55EfHw84uLicPv2bYUT7KysrCrzVRaLbND0jRs3CtPIxyJbhx9//LFWT6tT3udu3LiBkJAQvPnmm5g8ebLR8jIuLg45OTk4efIkgIpjwe3bt+Hi4gJfX18QkcKYcABQp06dKuuDDh064MqVK8IYJd988w2aNGmCJk2aKHzvwYMHePPNN/Hnn38iIiICXbp0QUREBG7duoXi4mIkJycLvfCqkp+fj2bNmsHBwQEnTpwwyvGqqvpI/ngEAF5eXhpzqVevXjh06JCQL2KxWGFdzp49G4MHD0aPHj1U1tGadOvWDd999x2ePXsGIsLGjRs1fr+wsBCvvPIKdu3ahWnTpinklXL+5efno2nTpkLvdVndouq7jRo1AhEJYzNt3bpV+OzevXtwcHDAgAED8Omnn4KI8McffwifWzofgYpen2VlZZXy8e2338aDBw8Ujh+q7lqQX2fGzkdVxyfZMaxZs2YoLCzEixcvFJZVo0YNvPrqq5g3bx5u3bqF3r1767RMDw8PREREYNu2bQAqzhNOnToFoGLQ8Fu3biEqKgrTp09HUlKSUO8rKywsRFJSErZt24axY8cqHF9V5Zu3tzc8PDxQVFQkjJ+m6ruAYi+kCxcuCOccRUVFePjwIeLj4/Hhhx+iU6dOuHTpkk7lB/TPy5ycHI3H1S+//BKOjo5CvFXlpabzGm3J8vLhw4cQiUS4dOkS5syZgy5duuDgwYMK9Zm2BgwYgG3btgk9saRSqTDOl64OHDgg3EFw5coVoTfOixcvhEaBHj16YOvWrSgrK0N5eTk2bNigMD7itGnTMGrUKAQFBam8QFyVXr16IT09XeECrLq8lufh4YHnz59Xed6gjjHqP1XncosXL8bly5dx7tw5JCcn49VXXzVbnsnXf/J5tmbNGmFf0FaPHj2wa9cuIUfXr18vbPfGjRvD09MT69atQ48ePdC1a1fs27cPd+7cQdu2bbWav6rfl8r1jba5XtX3zp49K5wHyB6uUFUngx49euDOnTvCGIDy5ddk4MCBWLdunXAeLBaL9aoHbRH3NNOgUaNGCgM5ypSXl2PAgAGYO3cuAgMDK31eWFgoXNU9depUlbeQ6cPV1RWrV6/GO++8g4MHDyIvLw/vv/8+ateuLQzoN3bsWDRr1gxEVGXLvyoeHh5YsGABYmJi4O3trXC1z83NDd988w3eeOMNSKVStGvXDoGBgahduzaGDRuGL7/8EmlpaQgNDYWDgwNq1aqF5cuXC11pTbVeVCkqKhJuMTl58qRw0AQqDirp6elISEjAoUOH8M9//tMsMRnC1dUV33//PaZOnYr33nsPNWrUwLfffits+zlz5qBjx46oU6cOevfurXC7ji45vX37drz77rsK23D9+vVIS0tDXl4ePvzwQzg6OmLy5MkYNWoU9u3bpzDPNWvW4NixY3B1dYWTkxOWL18OAEhOTsaYMWOQlpaGN954A61atVJZzri4OCxduhRAxSCqPXv2RP/+/TFw4EC4ubmhZ8+eqF27tsIg0coSEhIwd+5c9OrVCw4ODnB1dRUGAVemrrxNmzbFzp07MWXKFEgkEmHQbdl6u3btGqZNm4bAwMBKZXv8+DG2bdsGV1dXSKVSm3hKlzJ98k3fulOdiRMnIikpCUFBQWjRooXOt3Fp0qBBA2zcuBHvvPMOnj59igkTJqBmzZqIioqCs7Mz9u3bh7fffhsrV66ERCKBt7e3whVjiUSCyMhIFBcX4/PPPxcGS9+/fz+mT5+O9957D2KxGC+99FKVFzBksQwaNAiurq4K3eNlsUybNg3Tpk2Dl5cXFi5cqPLqtbKpU6cK5dq8eTPmz5+P7OxsuLq6ombNmgpXTg1Rr1497NmzB++99x6Kiorg6OiIlJQUDBs2DCEhIahfv36lx5+/99576NmzJ2rWrInDhw+jYcOGKtfL9u3bMWrUKJSXl6NevXrYvXt3pSujjRo1wokTJ3DmzBlMmTIFT548EXqVzZ8/H4sWLUJKSopWZVmyZAmmTJmClJQUiEQihdsi9FVVfRQXF4cVK1YIt+527NgR9erVU5tLrVq1wqZNm/Daa69BLBbDyckJ69atQ0xMjDDPt99+G7Vq1UK3bt1w6NAhrWP9xz/+gfPnz0MkEqFu3bro3Lmz2rq2vLwcw4YNg7u7OwYPHlwpryZOnIj33nsPK1euxKJFizBnzhwkJydjy5YtaNmyJbp16ybMa9CgQfj6668hEokwePBgfPTRR1i9ejX+8Y9/oH79+khKShK+e+XKFcyaNQtEhPLyciQnJyM8PFz43NL5CFTUn7t371bIRwcHB6xatQre3t4Kxw9NTJGP3377baXjU+vWrdGqVSvMnDkTz549Q+vWrSst65///CdiYmIwY8YMvXp/b926FWPHjsUnn3wiPGiibt26woOZ8vLy4OzsjAYNGqi8pUmWb+Hh4UhKSqp0fFXOoWnTpmHv3r0IDAxEgwYNEB8fL1xUDg8PR0hICEJDQ9GiRQv88MMPWLBgAUaPHo3169cLDywAKn74JiUlobi4GA4ODggICMDo0aN1Lr++eTly5EiVeSk7ri5cuBBisVjrvFR3XqMLWV7WqlULIpEINWvWxAcffICBAwdizJgxet3pEh8fj2XLlmHQoEEoLy9HWVkZ+vXrp9dg+PHx8RgxYgTu378PHx8fYb3dunULjRs3BlBRP926dUtoDOnSpYvCg8GSkpIwefJkvPnmmzovH6hohN2xYwdef/11lJSUoKysDJGRkVUeu728vDBq1CiEh4ejdu3aOjccGqP+U3cut3//fvTq1Qv169fH3r17zZZn8vWffJ6JRCKdH+bWp08fXL16FbGxsXB0dER4eLjC3TA9evTAvn37hNtJGzVqhMjISI0XN+TJjuejRo3CoUOH0LFjR7Rq1UrhN8L48eO1yvWq9om4uDjMmDEDN2/eRP369RUuLmkqf6tWrfDaa6/B3d29UvnVGTlyJJ48eSIMz1ReXo6xY8ciMjJSq/Vi08w6gpqd2Lp1K3l5eVHnzp2pc+fO9M033xARCQMAfvnllxQdHU2xsbE0ffp0k8Yyffp06tSpE40YMYJKS0uFGPLy8mj06NHUqVMnio+PFwa8NKbCwkLh7wsXLlCjRo3oyy+/rLRuLLFe5B04cIDatm1LnTp1ouTkZBKLxUJMYrFYWE+yQWttnfx2WbVqlVbbXl1OW5pyfsuXbc+ePdSmTRsLRme9682cdMk3a19fyvlWFWgxaK6xWfs6tEa2tM50zUFTku3bUqmU3nnnHWGQYGW2tH6tibWuN1PnYFFREUmlUiIi+u2338jHx4d+//13rae31vVmKda+Pvr06UO+vr7UoUMHYaB8c1B+IASRfrl98eJFCggIIIlEYoIobZM155yl8k0VcxzP5R8YoAtrWk+2woHo/wbzYXbljTfewOnTpyEWi9G4cWOsX79e4+CL+ti8eTNWrlwJIoKzszOWLVtmkwOd2xtzbHtLWbRoEXbu3AmJRAIPDw+sWbNG667SzDTsOd+qIht7T1NvR8Zs1aBBg3Dnzh28ePECISEhWLduHby9vS0dFrNxhw8fxvvvvw+goqfuv/71LwwfPtzCUTF7M2/ePDx9+hSrVq3Sex7jx4/H4cOHK92yyZi12Lx5M9LS0vQeiolpjxvNqpFHjx6prPTv3r2LZs2aVXp/9erVRhnglFmevW/7qKioSoNbhoSEYPv27RaKqHqzxnybNGmSyrG7zp49C3d3d5Mum2nH1rbR/Pnz8f3331d6X1We20t9lJWVJTy1W97o0aPxzjvvmD8gE+J8tA58fFfEecnMgfNMPwcOHMDs2bMrvT9r1izhaavMNnGjGWOMMcYYY4wxxhhjSvjpmYwxxhhjjDHGGGOMKeFGM8YYY4wxxhhjjDHGlDhbOgBrJJVKkZubizp16qh8hC6zL0SEoqIiNG7cWOtHCZsa52D1wfnHLInzj1kS5x+zNM5BZkmcf8ySrDH/rBU3mqmQm5sLPz8/S4fBzOyPP/5A06ZNLR0GAM7B6ojzj1kS5x+zJM4/Zmmcg8ySOP+YJVlT/lkrbjRToU6dOgAqEsjDw8PC0ehPLBbj8OHDSExMhIuLi6XDMYgpy1JYWAg/Pz9hu1sDc+agPeVJVayxrNUx/6xxOyizhRgBw+O0pfyzxm3CMWlPVVycf9ajupbPVnKwum4fe2Hr+QdU321kL2zlGGytuNFMBVlXVA8PD5tvNKtZsyY8PDxsfuc3R1msqQuyOXPQnvKkKtZc1uqUf9a8HWRsIUbAeHHaQv5Z4zbhmLSnKS7OP8ur7uWz9hys7tvH1tl6/gG8jWydrRyDrRXfvMoYY4wxxhhjjDHGmBJuNLNC/jP3w3/mfkuHwawQ5wazJFn+cQ4yxnThP3M/QucdsnQYTAu8naxb6LxDfAxmFsV1BKuOuNGMMcYYY4wxxhhjjDEl3GjGGGOMMa09fPgQcXFx6Ny5M7p164Y///wTp06dQlxcHDp16oQrV64AAB48eIDExER07NgR27Zts3DUjDHGGGOM6Y4fBMAYY4wxrXl7e+PUqVNwdHTE5s2b8dVXX+HIkSPYv38/ioqKMGnSJBw4cABLly7FBx98gC5duiA+Ph5JSUmoUaOGpcNnjDHGGGNMa9xoxhhjNuj48eNISUmBVCrF1KlT0aBBA3zwwQdwdHTE2rVrERYWhgcPHmDUqFEoLi7G5MmT8dprr1k6bGYHnJychL+LiorQsmVLHDt2DPXq1UO9evWQl5cHALhw4QKWL18OR0dHREVF4erVq4iKiqo0v9LSUpSWlgqvCwsLAVQ86UksFgvvy/6Wf8/SOCbtuDkR3BwJAFRuU8YYY4wxa2V3jWapqamYOnUqHj9+jN27d2PlypVwd3fHli1b0LRpU0uHxxhjBnv+/DmWL1+OH3/8Ea6urgCAzp07c08fZjZZWVl4/fXX8fTpUxw+fBg7d+4UPnN2dkZZWRnEYjEcHStGgfD09BQa05QtXrwYH3/8caX3Dx8+jJo1a1Z6/8iRI0YqhfFwTJoti/n7b/m4SkpKLBANY4wxxpj27KrRTCKRYPfu3fDz80N5eTlWrFiBn376CRcvXkRKSgrWr19v6RAZY8xgZ8+ehbu7O/r374+aNWviiy++gJOTk949fRjTlUgkwvnz57Fr1y4sXLhQ6B0GAOXl5XB1dYWLiwukUikcHR1RUFAALy8vlfOaNWsW3n33XeF1YWEh/Pz8kJiYCA8PD+F9sViMI0eOoGfPnnBxcTFd4XTAMWkndN4huDkSUqKkCnHJ5w1jjDHGmDWyq0az1NRUvPrqq1i+fDlycnIQFBQEV1dXdOzYEdOnT1c7nba3hpiLm1PlWxj0YY23aOjLlGWxh/XDqpeHDx/i5s2bOHfuHNLT0zF37lyFxgVde/poWwfK6ibZZ/qyhbrJFmIEDI9Tn+nKysqEHo6enp6oXbs2ysvL8fTpUxQVFQmNY9HR0Th+/DgSEhLw888/Y9myZSrn5+bmBjc3t0rvu7i4qGz0Ufe+JXFMmpVKHIS/5eOylvgYY4wxxtSxm0YziUSCXbt2IS0tDcuXL0d+fr7Cj0iJRKJ2Wl1vDTE12W0MBw4cMMr8rOkWDUOZoix8ewizNXXr1kXHjh3h6uqK7t27Y+7cuahTp47wua49fbStA+VvsTJG/WQLdZMtxAjoH6c+9V9WVhamT58OJycn1KhRAxs3bkROTg769u0LBwcHfPHFFwCAGTNmYNSoUZgzZw4mTZoEd3d3vWJkjDHGGGPMUuym0Wzbtm0YMmSI0Kuibt26Ct3+5QcuVqbtrSHmEjrvEADg6rxeBs3HGm/R0Jcpy8K3hzBbEx0djeXLl4OIkJWVheDgYNy+fVvvnj7a1oGyugkwrH6yhbrJFmIEDI9Tn/ovJiYGJ06cUHjP19cXZ86cqfSerTQ6MttT1Ri2N27cwMSJE1FeXo6UlBR0797d0iEzxpjRcB3ImPnYTaPZtWvXcOnSJWzbtg05OTlYvXo1rl+/jrKyMmRmZiI8PFzttLreGmJqstsYjLVsa7pFw1CmKIu9rBtWfXh7e2PQoEHo3LkzHBwcsHHjRty/f1/vnj7a1oHKt1gZyhbqJluIEdA/TlsoG2PKtBnDdvbs2fjqq6/g4+ODPn368A9Gxpjd4DqQMfOym0azpUuXCn9HRUVh7dq12LlzJ7p06YIaNWpgy5YtZo/Jf+Z+AMCdJf3MvmxmenyFh1nSG2+8gTfeeEN43bJlS+7pwxirFrQZwzY3NxcBAQEAAC8vL/z111/w9vauNC9tx3S0lXEO9SUrl5sj2WUZ1W0/eywrs3+WrAOrWx1hL1SVz17Lagp202gmLzMzEwAwdOhQDB061MLRMHvEV3gYY4wx89N2DFupVCq8J3sQiqofjLqOa2vvFyJSoqRGG1PXGilvPx7XltkaS9eB1a2OsDfy5eP6T3t22WhmL7inmvUy5hUexhhjjGlH2zFsZZ8D0PggFG3HdLSVcQ71JSvfh5mO+Pmj3pYOx+jUbT8e15bZGkvXgdWtjrAXqsrH9Z/2uNGMMR0Z+woPoH3XaDcnEt43FnvvjizPGstqTbEwxpi103YMW19fX9y6dQsNGzbUePzVdVxbWxnnUF+lUge7Lp/y9tOnrBcuXMC0adPg4uKCJk2aYOvWrUhLS+NhOphZWLoOrG51hL2RL589l9PYuNHMwrg3me0x9hUeQPuu0ctiKv43Rbdoe++OLM+ayspdoxljTHvajmG7cOFCjBkzBhKJROXxlTF9+fn54dixY3B3d8esWbOwd+9eHqaDmQ3XgYyZHzeamYGsYQzgxjF7YOwrPID2XaND5x0CAFyd18to5bH37sjyrLGs3DWaMcb0o2kM2+DgYJw8edISYTE75+vrK/zt6uqK7OxsHqbDxtjLbzOuAxkzD240Y0xHprjCo23X6FKJg/C+sdl7d2R51lRWa4mDMcYYY9q7e/cuDh8+jCVLluDx48fC+7oO06HNEB3yTy6Uf20vzD18hmy4E3Mtk5/eypht40YzC5G/wsFsF1/hYYwxxhirXgoLC5GcnIzNmzdDIpEYNEyHLk8vTImqaIiz16cXmmv4DNlwJ4B51yU/vZUx28SNZjZKn7HQePw0xhhjjDHG9FdeXo5hw4Zh7ty5CAwMhFgsNmiYDm2G6JB/cmGp1MGow3RYA3MPnyEb7gQw7pAn6vDTWxmzbdxoxhhjjDHGGGNaSE1Nxfnz55GSkoKUlBRMnjwZb7/9tt7DdOjy9MJSqQNKJfb79EJzDZ8hG+5EtkxzMcbTWxlj5seNZowxxhhjjDGmheTkZCQnJ1d6n4fpYIwx++RY9VcYY4wxxpgy+Vt8GGOMMcaY/eFGM6Y1/5n7+QEGjDHGGGOMMcYYqxa40czGcUMWY4wxc7pw4QJiY2ORkJCA4cOHQywWY/fu3YiLi0P37t1x7949AMCNGzeQkJCAuLg4HD161MJRM8YYY4xVL7K2Au4ZbxhuNGOMMcaY1vz8/HDs2DGcOHEC/v7+2Lt3L1asWIHjx49j/vz5SElJAQDMnj0bX331FQ4ePIiPPvrIwlEzxhhjjDGmO240syGylmJDe5Zx7zTG7ENqaioaNGgAANzTh5mNr68v3N3dAQCurq7Izs5GUFAQXF1d0bFjR1y+fBkAkJubi4CAAHh4eMDLywt//fWXJcNmjDHGGGNMZ/z0TDshawS7s6SfhSNhjJmDRCLB7t274efnh/LycqxYsQI//fQTLl68iJSUFKxfv17o6ePj44M+ffqge/fulg6b2ZG7d+/i8OHDWLJkCR4/fiy8L5FIAABSqVR4z9PTE3l5efD29q40n9LSUpSWlgqvCwsLAQBisRhisVh4X/Z3u/kH8fNHvY1bGD25ORIAKMRpabJYrCkmNydSua6sKUbGGGOMMVW40YwxxmxQamoqXn31VSxfvhw5OTkKPX2mT58O4O+ePgCEnj6GNFq4OZHwtyE/dq3xR70yW4gRMDxOfacrLCxEcnIyNm/eDIlEIuQMADg5OQEAHB3/7sxeUFAALy8vlfNavHgxPv7440rvHz58GDVr1qz0fkqUFAcOHNArbmNLiar4/8iRI5YNRAVrimlZzN9/y8dVUlJigWgYY4wxxrTHjWbVmPwtmtxDjTHbIZFIsGvXLqSlpWH58uXIz8+Hh4eHwueA9j19tG20kP/ha4xGC2v6Ua+OLcQI6B+nPo0W5eXlGDZsGObOnYvAwECIxWJcv34dZWVlyMzMRHh4OICK2zhv3bqFhg0bqs09AJg1axbeffdd4XVhYSH8/PyQmJiokNdisRhHjhzBh5mOVtPTrN38g0iJkqJnz55wcXGxdDgA/l5P1hRT6LxDcHOkSutKvrGVMcYYY8wacaMZY4zZmG3btmHIkCFCT566desa1NNH20YL+SfvXJ3XS+/4rfFHvTJbiBEwPE59Gi1SU1Nx/vx5pKSkICUlBZMnT8bbb7+NLl26oEaNGtiyZQsAYOHChRgzZgwkEonKRlkZNzc3uLm5VXrfxcVFZZlKpQ5Ws01KpQ4A1MdqSdYUU6nEQfhbPi5riY8xxhhjTB1uNLMzPLYZY/bv2rVruHTpErZt24acnBysXr3aoJ4+2jZaKP/wNZQ1/ahXxxZiBPSPU59pkpOTkZycXOn9oUOHKrwODg7GyZMndZ4/Y4wxxhhj1oIbzYzIWhusQucdUvixyxizbUuXLhX+joqKwtq1a7Fz5069e/owxhhjjDHGGKuMG80YY1bDWhuerVlmZiaAil4+3NOHMcYYY4bi8zHGGPubY9VfYYwxxhhjjDHGTCd03iGFB5Uxxpg14EYzxhhjjDHGGGOMMcaUcKMZY4wxxhhjjDHGGGNK7KrR7MKFC4iNjUVCQgKGDx8OsViM3bt3Iy4uDt27d8e9e/csHSJjjDHGGGOMMcYYswF21Wjm5+eHY8eO4cSJE/D398fevXuxYsUKHD9+HPPnz0dKSoqlQ2SMMcYYY4wxxhhjNsCuGs18fX3h7u4OAHB1dUV2djaCgoLg6uqKjh074vLlyyqnKy0tRWFhocI/ABCLxTr9c3MiuDlRpdfy/9R9pmo+2sxP3T/ZNADg5qj993Upn7n/6bNNdJk3Y4wxxqyftncW3LhxAwkJCYiLi8PRo0ctHDVjjBmO6z/GzM/Z0gGYwt27d3H48GEsWbIEjx8/Ft6XSCQqv7948WJ8/PHHld4/fPgwatasqfVyl8VU/H/gwAGF1/LUfSZ7X9Nnquanjvz8UqKkOn1fHeXyWcKRI0eMPs+SkhKdp7lw4QKmTZsGFxcXNGnSBFu3bkVaWhpWrlwJd3d3bNmyBU2bNsWNGzcwceJElJeXIyUlBd27dzd6/Iwxxlh1IruzwN3dHbNmzRLuLPjpp59w8eJFpKSkYP369Zg9eza++uor+Pj4oE+fPnwMZozZPK7/GDM/u2s0KywsRHJyMjZv3gyJRCL0GgMAJycnldPMmjUL7777rsI8/Pz8kJiYCA8PD62XHTrvEADg6rxeCq/lqftM9r6mz1TNT52r83pBLBbjyJEj+DDTEaVShyq/L0+5LOreU0eX72pDVpaePXvCxcXFKPOUkc8RbfEBizHGGLMMX19f4W9VdxZMnz4dAJCbm4uAgAAAgJeXF/766y94e3srzKu0tBSlpaXCa+W7DWTke7zbI/m7E+yxjOq2nz2Wldk3Y9Z/gO51YHWrI2ydmxNV/O9Y8b+qbcqqZleNZuXl5Rg2bBjmzp2LwMBAiMViXL9+HWVlZcjMzER4eLjK6dzc3ODm5lbpfRcXF50aaEolDsJ08q+V56nqM/nlqPtM1fzUUZif1KHKaQM+PAwAuLOkn8KyVMWlzTrR5bu60HWbaDtPXRn7gMUYY4wx3VR1Z4FU+ndPe09PT+Tl5VU6But6t4Eperxbk5QoqUXvKDA15e2nz90GBQUF6NmzJ65du4Zz584hNDQUu3fv5rsNmFkZo/4DdK8Dq1sdYeuU71STL58+9V91ZVeNZqmpqTh//jxSUlKQkpKCyZMn4+2330aXLl1Qo0YNbNmyxdIhMjtjrAOWtld5ZFcLjHllwJqurJiifPKsqawy1hQLY4zZCm3uLHB0/Hvo3oKCAnh5eVWaj7Z3G5iyx7s1kL874eePels6HKNTt/30udugZs2a2L9/P95//30AFRft+W4DZk7Gqv8A3evA6lZH2DrZ3V9ujoSUKKlC+fSp/6oru2o0S05ORnJycqX3hw4daoFomL0z5gFL26s8phxXzhqurJhr3DxrKKsMX+VhjDHdaHtnga+vL27duoWGDRuqvWil690Gpujxbk1KpQ52XT7l7adPWV1cXNCgQQPhdU5Ojslvj5O/NU7+tamY+iKmMkuVzxzLkl+GMW4PNmb9B+heB1a3OsLWqbqDTVY+eyqnqdlVoxkznP/M/ZYOwSYY+4Cl7VUeY48VB1jXlRVTlE+eNZVVhq/yMMaYbrS9s2DhwoUYM2YMJBKJygtTjBlDfn6+wrmaKW+Pkz3cy9QXFy318C9zl88cy5JnjNuDuf5jzPy40YwxPRj7gKXtVR5TjRWnalmWYMryybOGsspYSxyMMWYrtL2zIDg4GCdPnjRXWKyaqlu3rslvj1N+uJepLi7KmPoipjJLlQ8wTxmNeXsw13+MmR83mjGmBz5gMUu6cOECpk2bBhcXFzRp0gRbt25FWloaD0LMzIYHwmaMsQoBAQFmuz1O9nAvU19wM9dFzErLNXP5APOW0Ri3BzPGzI8bzZjB5G/plD19kzFmOn5+fjh27Bjc3d0xa9Ys7N27lwchZmbFA2Ezxqqzvn37IisrC9nZ2Xj99df59jjGGLNj3GjGGGM2xtfXV/jb1dUV2dnZBg1CzJiujDkQtrZPD5YfKNpanjprrkGrdWGNTwl2cyKV68qaYmRMF6rGweK7DRhjzD5xoxnTiB8MwJj1unv3Lg4fPowlS5bg8ePHwvu6DkKsbaOFsZ42ZY0/6pXZQoyA4XEaq3yGDIStyyDYQMVA0eYenFqdlKiK/63pibwy1hST/KDb8nHx04OZOfjP3A83J1LIQ8YYY0xb3GjGGGM2qLCwEMnJydi8eTMkEolBgxBr22hh7KdNWdOPenVsIUZA/ziN1WhhyEDY2j49WH6g6J8/6m2UuA3Vbv5BpERJreqJvNb4lODQeYfg5kiV1hU/PZgxxhhj1o4bzRhjzMaUl5dj2LBhmDt3LgIDAyEWiw0ahFjbRgtjPW3KGn/UK7OFGAHD4zRWo4UhA2HrMgg2UDFQtLVsk1Lp34NlW0tMMtYUk/Kg27K4rCU+xhhjjDF1uNFMR7LbFavzgPe8DhizrNTUVJw/fx4pKSlISUnB5MmTDRqEWNtGC2M/bcqaftSrYwsxAvrHaUjZeCBsxhhjjDFm77jRjDHGbExycjKSk5Mrvc+DEDNz4oGwGWOMMcaYveNGMz1xbyvVDwnQ5sEBvO4YY4wxxhhjjDFm7Ryr/gpjjDHGGGOMMcYYY9ULN5oxxhhjjDHGGGOMMaaEb89kJmXKWzH5Nk/GGKua/G3zXF8yxhhjjDGmPe5pxhhjjDHGmBZC5x2ydAiMMcYYMyPuacYsTlWPMeX3QucdwrKYiv9LJQ7mD5IxxhhjjDHGGGPVCvc0Y4wxxhizM9wjijHGLM9/5n6ujxmzcdxoxhhjjDHGGGOMMcaYEr49k5mF/EDUhnxH03Q8wDVjjDHGGGOMMcaMhXuaMcYYY4wxxhhjjDGmhHuaMYvRpfeZm5Opo2GMMcYYsw7ci54xxhizDtzTjFUL/jP36337J2OMMcYYY4wxxqofbjRjjDHGGGOMmQ1fzGSMMWYruNGM2Q0+AWOMMcYYY4wxxpixVItGsxkzZiA+Ph7JyckQi8WWDodVM5x/zJI4/5ilcQ4yS+L8Y5bE+ccsjXOQMcPZfaPZL7/8gvv37+PkyZNo06YNvv32W0uHxKoRzj/bYm+9FTn/mKVxDjJL4vxjlsT5xyyNc5Ax47D7RrMzZ84gMTERANC7d2+cPn3awhExU7Omhg/OP2ZJnH/M0jgHmSVx/jFL4vxjlsY5yJhxOFs6AFPLz8+Hr68vAMDT0xN5eXmVvlNaWorS0lLhdUFBAQAgLy+vUjdW5/JihddPnjyp9JnsPeXvavpM1XyqmkaTJ0+eQCwWo6SkBM5iR0ikDlpPa42cpYSSEqlOZdG0beQVFRUBAIjICJEq0ib/AO1zUFM59CXLkydPnsDFxcVo89WHKconr6qymmr57RcfBQCcn9W90me2mH+AYevImnJOHVuIEdAuTk3bzZT5Bxj3GCx/TDNVHaErZ3ExSkqkVpUnVrmeyouF47j8uuL8U8/Ux0Nl9lw+dfkH2M4xWPmc3tTrzZL5Z87yAaYvo6XyD7DtOtDcbOW8T1eyXLfEMdie2H2jWd26dVFYWAigohLw8vKq9J3Fixfj448/rvR+8+bNq5y/93Lt3qvqM32m0USfaazdCB2/r+u2KSoqgqenp45L0Uyb/AN0z0F73L7yLF0+Uy3fXvIPsPw2YvpRt91MkX+A6Y7B3p8YL0ZD6XpsMidrWk+A+nXF+aeeJepaey2fpn2Vj8GWXY4ll2uuZZk7/wD7qAOZ8Zj7GGxPHMjOmxazsrKwYsUKbN26FYsWLULz5s0xfPhwhe8ot7BLpVLk5eWhfv36cHCw3R5ahYWF8PPzwx9//AEPDw9Lh2MQU5aFiFBUVITGjRvD0dG4dyxrk3+AZXPQnvKkKtZY1uqYf9a4HZTZQoyA4XGaMv8A4x6DrXGbcEzaUxUX55/1qK7ls5VjcHXdPvbCEvkHcB2oi+pYPlPnnz2x+55mIpEIPj4+iI+Px0svvYTp06dX+o6bmxvc3NwU3qtbt66ZIjQ9Dw8Pu9n5TVUWU7Wua5N/gHXkoD3lSVWsrazVNf+sbTuoYgsxAobFacqri6Y4BlvjNuGYtKccF+efdamO5bOlY3B13D72xJz5B3AdqI/qVj7uYaYdu280A4BPPuE+pMxyOP+YJXH+MUvjHGSWxPnHLInzj1ka5yBjhuN+eIwxxhhjjDHGGGOMKeFGMzvm5uaGuXPnVupya4vsqSzWpjqt2+pUVmtmC9vBFmIEbCdOY7DGsnJM2rPWuLRl6/FXhctn3Ww9/qpw+ayfPZRBEy4f08TuHwTAGGOMMcYYY4wxxpiuTN7TbN68eXjx4oWpF1PJ8ePHIRKJdJ4uKysL33zzjd7LHT9+PDIyMqr8nr+/P7KysrSer6by5ObmIj4+Xqv56LpcY7pz547CwJIikQhFRUUAgHPnziEsLAyRkZE4dOgQACAzMxN9+vTRa1nz5s3D22+/LbweN24cgoODMWjQIL3jN9Snn36KpUuXGnWe2u5fx48fx8GDB422XHvbv9RRziN58vkrv7wuXbogLS2tyvkqb7ekpCRs3rzZwIg1qy7bzZq8+eab2Llzp9Hna6rcVEVTbvIxXjVL57Km9SNfhjFjxmDVqlUANOeUJmvWrMGYMWP0C1QLffv2RXZ2tsrPDK03582bh127duHcuXN6z0MflspffesAfcjXQ8a2d+9eBAUFQSQS4cqVKwbNa/369WjTpg1EIhGePHlipAiNp7rVsfaSo/aC8890bCH/0tLSzH58tFYmbzT7+OOPLbKz6cvQnW3Dhg3o2rWrESOqWuPGjXHy5EmzLtMYsrKyUKdOHQDAli1bMGLECFy6dAm9evUCAOzZswcDBw40eDkPHz7EN998gytXrmDPnj0Gz09faWlpRimPPG33L2M3munLFvcvdeTzV1dcL1YPRIQff/wRffv2NetyDclNXZkzl8vLyw2eB+ey5cqg7/Y7cOAAAgMDjRxNhY8//timfhQYmr/mZMp6aN26dfjoo4+QlZWFsLAwradTlYOrVq3Cpk2bkJWVhfr161f5fXOrbucL5mTOY6Wt4vwzHXPlnyH1mC0dH01Np0azs2fPolOnToiIiEB4eDj27t0LAJg+fTqio6MhEomQkJAgXBGcNGkSACA+Ph4ikQiPHj1SO+9Dhw6hbdu2CA8PR+fOnXHt2jUAFT/2Q0NDMWXKFERERCAkJASZmZkK03Xq1Ant2rVDTEyMwhXg8vJytdN9/fXXCA8PR3h4OPr164f79+/j0aNH+Oijj5CRkQGRSCTErwv5FupHjx5h8ODBCAsLQ2hoKNavX6/w3e3bt6Ndu3Zo1aqVVk82UVce5R5csitwDRs2RNOmTVGjRg3k5OQIn3///feIjY1F8+bNsWDBArXLmzdvHgICAtCuXTvMmTMH/v7+Kpf37NkzODg4CK9HjhyJqKgoYd0+ePBA5fwdHBzw9OlTLFmyBDt37sSaNWsgEonw9OlTAMAPP/yAl19+GYcOHULdunVRq1Yt1K1bF6NHjxbm8emnnyImJgZt27ZF7969cffuXYVlPH36FF27dsWLFy/Qrl07LFmypFIcOTk56NevH6KjoxEeHo41a9YoxLhw4UK0b98e/v7+SEtLw+LFixEVFYWAgAAcP35cYZ1Mnz4d4eHhCAkJQXp6ujCfhw8f4unTpwgMDMTmzZvRo0cPDB8+HMHBwYiLi8O1a9cwaNAg+Pv7o169eggLC0N4eDi++eYbBAYGwt3dHbVr14avr6/QW072o8ff3x81atRAdHS0MJ+goCAkJibi2bNnyMrKwmeffYY1a9bA3d0d/v7+CvtXgwYN4Ovri3r16iE4ONhu968LFy4gNjYWCQkJGD58OMRisUIOdOzYEREREQgLC8OcOXMqLfvatWsIDQ3Fjz/+KOTG06dPcfbsWTx48ABDhgxBeHg4/vrrLwAV9WLbtm1Rv3591KlTB4GBgZg4cSImTJgAAIiJiUGtWrUQGBiIgQMHorCwsNIy7bVebN26tbA/f/HFFwrf1aVeVL4SefXq1Ur11Ny5c4X5HThwQPjuxYsX0a1bN0RFRSEyMhK7d+9WmO79999HrVq14OjoiK+//hrvvPMORCIRQkNDcfXqVWH5oaGhGDVqFEJDQ9GuXTuF3kWZmZkIDAxEnTp1KvXike+dI18nhIWFISoqCr/99hsAzbn54MEDxMbGol27dqhbty7+85//APg7NzUpKirChAkTEBMTg/DwcEycOBEnTpxAp06d0KZNG9SqVQsvvfQSBg4ciMzMTHz88ceVjvEjR44EAAQGBqJmzZrw9/dXWMfy+RsWFobg4GCEh4cjIiICrVq1wpQpU9CyZUvUqFEDSUlJKnOyZ8+e8PHxQZ06dVCjRg2EhYWhY8eOmDRpErKzs9G4cWMEBgYiJCQEmzZtQrt27SASidC0aVM0adLEpLl88+ZN1KtXD46OjggICDDoGK9vLj9//hxDhw5FcHAwIiIi0L59e8TGxmLq1Km4e/cuxGIxCgsLkZiYiPnz51cqgybqjq9FRUUYOnQoAgMD0alTJ409ffz9/TFjxgxERkbC19cX8fHxaNasGSIjIyESiTBkyBDk5+cDqGjMCw4OhkgkQlhYGM6fPy/MQ7Zf3bhxA3FxcQgJCalUb6anp6NRo0aoWbMm3N3dkZiYiLKyMkyfPh0eHh5o2LAhPD094efnh0mTJgnbf+fOnXj//fcREhKCDRs2qCzHoUOH4OPjg1q1asHHxwe//PKLsN1sqS4GgFOnTiE+Ph4tW7YU5nHhwgVERUXB29sbdevWRUhIiEI+K/ecjIqKEs5/FixYIPT6EolEQp7I10P+/v746KOPVJ573rhxA7GxsQgJCcHgwYORmJiosffg1KlTcfLkScyePRtxcXHCelZ3rAwJCcG4ceMgEonw3nvvoUGDBsK8kpKScOvWLYwZMwZJSUnCvjZjxgy0bdsWa9asEY7vMTExlepgTeeP6qj6LaXN7ygfHx+0b98eycnJCucvMrZ8viB/fnbt2jWcOHGiUo4Cms/7rClH1W2P1NRU1KhRAzt27AAAfPHFF3B1dUVxcTEAoFu3bjhx4kSV5y+G0CX/ZMek5s2bIyIiQvgdL5FIMHbsWMTHxwvnNbaYfzNmzEB8fHylfSo1NRVubm5wd3dHkyZNhPdtKf8cHBzQpEkTODk5YcKECZXO+QIDA9GxY0e8/fbb6NKlC6ZPn15pnztw4AB++OEHfPLJJxCJRGqPj9UGaenJkyfUsGFDOnHiBBERSSQSevLkCRERPXr0SPheamoq9erVS3gNgPLz8zXO++HDh+Tl5UWXL18mIqJt27ZRUFAQSaVSysjIICcnJzp37hwREa1du5YSExOJiOjWrVvUoUMHKigoICKinJwcatSoEb148ULjdFeuXCEfHx+6d+8eEREtWLCAevfuTUREmzZtopdfflnb1VJJ586dac+ePURENGTIEJo5c6ZQxqZNm9LZs2eJiKhZs2aUnJxMUqmUHj9+TH5+fnT69Gm189VUntu3b5Onp6fCutyzZw+NHDmSNm7cSADos88+E5b71ltvERHR48ePycPDQ1gP8vbt20fBwcFUUFBAUqmURo4cSc2aNau0PCKioqIikk8l+XxYvHgxvf766yqnk8+N0aNH08qVK4XP/ve//1FcXBwREc2bN4/Gjh1LREQzZ86kjRs3EhHR9u3bafz48VReXk5ERFu3bqW+ffsSEdHcuXNp2rRpKpcrr7y8nNq1a0fXr18nIqLi4mIKCwujCxcuCDGuWrWKiIjS09OpVq1atGnTJiIi2rVrF0VFRQnLAEAbNmwgIqKzZ89SgwYNqLCwkIiI1q9fT7NmzSKiihzz8PCgu3fvEhHRa6+9Ri1atKDr169Tw4YNKTY2ltasWUMSiYTeeOMNGjp0KEmlUiosLKSmTZtS48aNhfkAEPYb2XwePHhARET9+vWjNWvW0MOHD8nd3Z1GjhxJRIr715dffkkODg507tw5WrBgAY0dO9Zu96+9e/dSSUkJEVXk0e7du4X5Tp06lRYtWiS8ltVtsjzKyMigoKAg+vnnn4XvAKDffvuNGjZsSD4+PnTp0iWSSCQUFxdHe/bsoUePHtGECRNoy5YtQr04btw4WrZsGQEgkUgk5Mvly5fJ1dVVyC1ZmeyxXkxMTKTg4GBhO3h5eRlUL0ZERAivr1y5olBPAaBvv/2WiIh+/PFHat26NRER5efnk0gkotzcXCIiYVn37t0Tptu9ezc9evSI4uLiyN3dnY4dO0ZERMuWLaOkpCRh+QAoPT2diIh27txJgYGBJJVKiYho1qxZtH79eiJSrJOIiFavXk2jR48mor/rhN9++42IiGbMmEETJ04kIs25OX78eDp48CAFBQXRmDFjhJyWr1ubNWtGly5dqrQdZLlJRCSVSum1116j2rVr04kTJygqKor+/e9/05MnT+jy5cvk4uIi5Kb8MV62rmTzkV/H8vn78OFDql27NjVv3pykUinNnj2bANDZs2cpIyODHB0dKSYmhogq53KDBg0oJCSETp8+TQMGDKD69etTq1ataObMmeTk5ETr1q2j4OBgWrt2LTVo0IB27Ngh5PLVq1eJyHS5nJSURFOnTqXRo0fT8ePHDT7G65PL33//vbC+iIh+/fVXKikpoYyMDGrYsCGtXbuW2rZtK2wj5TLIH3vlc1TT8XX69OlC2Z4+fUpt2rQRcllZs2bNaNy4cSQWi0kikdDChQtp0KBBlJKSQkRE8+fPpylTphARkYeHh7BPlpWVUVFRkTAPWQ5HRUWprDefPHlC7u7uNHv2bCKqOLa/9tprtGzZMnr06BF17tyZBg4cSNu2baMePXqQv78/nTlzhgDQ8OHDFc4/lD18+JA8PT2F8iclJVGTJk1sti4eOHAgicViKikpEdZDbm4uvfLKKzRz5kyaOXMmbdiwoVI+y7YBEVG7du0oIyOD8vLyyNPTUzi2FhcX0/Pnz4mocj2k7twzKipKOK+7du0aubm5KRwL1ZVDlsNVHSsdHBzo+PHjVF5eToMGDaLIyEiFecmXTblOI6o4Zh0/fpyIiMRiMfXq1Yt27dpV5fmjKup+S1X1O+rEiRPCOdyCBQtox44dCvO19fOF3NxcIYf8/PwoOjq6Uo4SVX3eZy05qm57DBw4kF566SX65z//SUREAwcOpNjYWNq/fz8VFxeTl5cXlZWVaazzDaFL/mVlZdHIkSMJAP3rX/9SyLm0tDT617/+RURE48ePp3379tlc/snKJ5teVr7OnTtTp06d6J133rHZ/JPN+4MPPqDRo0fTlStXFM759uzZQxEREbRs2TIaP348RUZGqjwuEFX+fV6dad3T7OzZswgMDBTGznJ0dISXlxcA4MiRI4iNjUVoaCjmz5+v8zge58+fR1hYmNDFeuTIkcjNzcX9+/cBAK1atUL79u0BALGxsbh16xYA4ODBg7h58yYSEhIgEomQlJQER0dH/P777xqny8jIQO/evYXW4ylTpuDYsWOQSCQ6xV2V9PR0vP766wCAhg0bYvDgwQq9j8aNGwcHBwd4e3tX+kwVdeWRd+7cOYSHh+PPP/9EYmIiRo8eDRcXF4XW+REjRgAAvL290aJFC9y+fbvSfI4ePYohQ4bAw8MDDg4OQjm0sWPHDkRFRSE0NBQbNmzQa1wX+VszExMTceTIEbz33nu4ffs2atasCaCiy2h6errQs2DZsmXCttdWdnY2fv31VwwbNgwikQhxcXEoKioSrpAAwNChQwFUXDUoLi7GsGHDAFT0FJLvwefs7Cz0HOnQoQMaN26MS5cuCbHKj6cWGxuLl156SZhvdHQ0bt26hcDAQPTq1Qs5OTlwdHTEmTNnEBAQgLi4OMTGxqK0tBR5eXkKZfDz81OYj4+PDwAgOjoaOTk5OH/+PBo2bAhvb28AivvXr7/+ikaNGqF9+/bo3bs3iouL7Xb/ysrKgru7OwDA1dUVjo5/V38JCQn48ssv8a9//QuHDx9W6El57NgxTJo0CQcPHkTbtm0VYrhw4QICAwNRo0YNABX1oouLC4CKenHTpk2YOHEixowZg6NHj+LkyZO4efMmAODKlStCvoSFhaFTp04K87bXevH06dMYO3YsgIqr/E2bNjWoXtSkRo0aGDx4MADFsp45cwa//fYb+vTpA5FIhB49egCAcHVV1vOpQYMGqF+/PmrWrCn07FTe7/39/dG9e3cAwJAhQ/DgwQP88ccfAP7uLasN2dVF5Vg15eb58+cxbdo0HDx4EC+99JJCTlclLS1NuHoYGRmJY8eOoXbt2oiIiEBWVhbGjh0LLy8vhIWFISAgAAsWLFB7jO/fv3+luOXz9/z582jXrh3y8vJw//599OzZE46OjmjatCkA4KWXXhLGElLO5cLCQjx58gRDhw5FVlYWSktL8eTJE3zxxRdwdnbG2rVrkZeXh8jISBARUlJS8MEHHyAyMhIhISEATJfLx44dwzvvvAMAqF+/vsHHeE3U5XJERASuX7+OKVOmYOfOnfDz8xPquZKSEsydOxfLli3DqFGjdFqepuPr0aNHhbJ5enoK5xTqjBkzBs7OznB0dERaWhpOnz6Nr776CiKRCKmpqcI5SPfu3ZGcnIzPPvsMt2/fRu3atRXmU1hYiKysLJX15tmzZ1FeXo4ffvgBIpEI7dq1w4ULF3Dz5k0cOXIE//3vf5GZmYmFCxfiypUrEIlEKs+hVJEdQ2XnAjNnzsTjx49tti4eOnQonJ2d4e7uLqwHX19fZGRk4PXXX4erqyvq1aunVc56eHggICAAr732GtavX4+8vDzheKhM1bmnbJvK8jMoKKjSsbAqVR0rW7Rogc6dOyM1NRWvvvpqlfWki4sLXnvtNQBAcXExjh49imnTpkEkEiEqKgo3b95Edna2VuePytT9lqrqd9SFCxeQmJgIAOjduzdOnz6t0zqw9hz19fUV6i1HR0d06tSpUo4CVZ/3qWKJHFW1Pe7evYvu3bvD09MTR48ehUQiwbVr1/Duu+8iPT0dJ0+eRExMjHAOqa7ON4Qu+XfmzBkh53r06KGQc/Kf9e7dG999953N5Z9yGeTL17VrV+zZsweJiYmoU6eOzeWfzFtvvSX8LX/O9/rrryMvLw83b95E7969UVBQoPK4wBQ5GzqD33//HW+++SYuXryIli1b4vLly0hISDBGbAL55HJychLuzSUi9OzZU+jmKu/+/ftqp1Mmf2uhKVW1nKo+17Y8AJCfnw9fX1/htfytOrrMR1Vszs7OChWT/L3up06dwueff46zZ8+iYcOG+OGHH/DRRx9VOX9laWlp2LJlC4CKijIrKwupqanYunUrfv31VyQlJYGIMGvWLEycOFHn+csQEby8vDQ27MnWl5OTU6XXVa07BwcHFBUVITs7G1FRUZXmKZuPum1SVlaGzz77DJcuXULLli3x1ltvVboFSJv5qFNUVCQ8etjT0xOFhYV2v3/dvXsXhw8fVrjF4pVXXkFcXByOHDmCNWvWYNWqVUJX+FatWuHGjRs4d+6c0NBZlcePH2PGjBmoU6cOzp07hxcvXih0ef/3v/+tc/zKbLVelEqlwhgOnp6eVY6VoSkOTXURUPF4bdn0Tk5OwneJCCEhIThz5kyled65c0fhcdwODg4Kr6varxwcHODg4ID//e9/qFu3rtCIXVWs6raLNrn5n//8p1JOV4WI8N1336F169YAgP3796t8UMnvv/+O//3vf1i4cCE++OADlcd4VetYF25ubsL6UM7lxo0bY+rUqSgpKUFwcDBOnz6NZ8+e4X//+x+ePHki1N1Xr15FrVq1sH//fsyePRvHjh3DlClT8MUXX1hdHaSKvrncokULXLt2DceOHUN6ejo++OADZGVl4cGDBygvL0dUVBT27duHbt266bQedDm+VjVfWeNXVlYWrl27hrp16+L48eNo1qyZwve+++47/Pzzzzh+/Dj69u2LBQsWCBeptFm2ck4DFfkrEokQFBSEWbNmoUWLFkhISND63EemvLwcHh4eACrqLZJ78Lyt1cWalnvv3j2hLjlx4oTwvrr8dHJywrlz53DmzBkcP34cHTp0QGpqqsqHU1mqvLVr14ZEIsGuXbuQlpaG5cuXa/x+zZo1hYY12XY+d+5cpR+6v/76a5Xnj9rQ5nfU06dP0bJlSwAV+ad88bQqtpKjd+/eRV5ennDbrS7LteYclUgkKCsrw4ABA7B582a4ubkJt+93794dCxcuhJOTk3ABDlBf5xubuvyT/x3p4eGhkHP5+fkK9WFVg9lbY/7Jl095nwoKChLO11966SXhYrem5Vpj/slfeJI/Pk6cOBFTpkyBSCRCeno6ysvL9WobqG60viwdFxeHnJwcYcB5qVSKvLw8FBQUwMXFBb6+viCiSvfz16lTBwUFBRrn3aFDB1y5ckUYJ+abb75BkyZNFO4jVqVXr15IT0/H5cuXhfcuXLhQZVm6du2KgwcPIjc3F0DFgKLdu3eHk5MTPDw8qoxXWz169MCXX34JoOJH9Pfff4+ePXsKn8vuR87Ly8OePXsUKkt9dejQAZcvX0ZpaSkKCwuxbds2iMVihd4J2sa+e/duFBUVgYgUfuQ3atQIRCRcTdu6davwWX5+PurUqYP69eujrKxMbQOPJn/++SeePXuGgIAAAMDt27chlUqxc+dOHDx4EDk5OXj27BkGDhyIdevWCRWdWCwWenZpKzAwEB4eHti0aZPw3s2bN3U+IQEqTqq//vprABV5mJubC5FIhAMHDqBPnz5VVnKyfezOnTsAKvaxkJAQiMViNGrUCEVFRZUOLM7OzlrtX48fPxbmK79/1a5dG1KpFABQUFAAT09PYTp73L8KCwuRnJyMzZs3C1fzgIpxSXx8fDBq1CgsW7ZMYdDLl156CUePHsWCBQsU8gQA2rdvj5ycHOHgKJVKIRaLUVJSAhcXF7z88stYsmQJPv/8cwAV+8fNmzdRp04dhISECPvOr7/+ilOnTinM217rxaCgIOzbtw9ARQNVbm6u3vViixYtcPfuXTx+/BgAhP2vKnFxcbh9+7bCFcKsrCyUlZXpWhzcuXNHGIPj22+/hY+PD5o2bVrpQSatWrVCZmYmJBIJSkpK8N1332k1/6pyMy0tDR988AH69++vkNNVGThwIJYuXSqcGAUHB+PGjRv45ZdfEBkZic2bNyMvLw8///wzysvLhYYCVWP2qMoN+fzt0KEDfv755//P3p3HRVW9fwD/zAzDsK8KsgUuyD4zICCCiIpspmnmVq5kmrZZ5teyLE0ry1IrM+2XaRpqueTyVXNBIXFLLXHfMJcUV0BAUJbh+f3Ba+6XYRMQGGbmeb9evmS2O+fc+8w59z733HNhb28PFxcX7NmzB1KptE6xfPPmTWRmZsLGxgZ5eXno0KEDNmzYACOj8nN+ZWVlwkjqkpIStG3bFu+//z5KSkqE/ZXmaIOys7OfqI9vaCxfv34dIpEIzzzzDL788ksQEc6ePYtPP/1UmIczMzMTY8eOFdr6uqitf+3VqxeWLVsGIkJeXh5Wr15dp2UqlUpMnToVrVu3FuZXKywsxOnTp1FaWopLly4hODgYkydPxsCBA6u0W1ZWVggMDKy23QwPD4dUKsVbb72F0tJSlJWV4Z9//sHJkychlUphbGxc4z6qsbFxrTERFhaGO3fuCPs8a9asgYWFhc63xZVFRUXhhRdewE8//YT79+9rxHOHDh2EOeYOHz4sHFDm5+fj9u3biIyMxAcffICuXbvWaz/MysoKCoUCSUlJAMpH+1buCx+nLn1lUlISBg8eXK/RuED5QWePHj005sPNzMzE9evXG7T/WN2x1OXLlx97HGVsbCzM35ebmytc6VOfdVCdlhSj6v0zb29voX2vrLb9vpYUo5W3x2uvvQZHR0fhypBevXrhww8/RK9evWBrawupVIq1a9cKo96bSn3iT93nWlpaIjMzUyPm1K8B5fHo5+enc/FXuQ4V62dqagpjY2MYGxvDyclJGPmmK/FXnYr7fDY2Nrhx4wYyMjKQm5tb4+9N/f1N1efomjr3Hra2ttiwYQPeffddyOVyBAUFYf/+/QgICMDQoUPh5+eHkJCQKiMx3n77bcTExNR6I4DWrVtj5cqVGDlyJORyORYtWoS1a9c+NsnQoUMHrFq1Ci+//DIUCgV8fHyEW6fXxt/fH1988QXi4+Mhl8uRlpYm/Aiio6NRVFQEuVze4ElW1eX+5ptvcPbsWQQEBKBHjx54//33hWGm6nqrJz587bXXNM6sNJSDgwOWLFmCn3/+GW+++SZOnjwJY2Pjeo/+6927NwYOHIigoCAEBwdrbFcjIyMsWLAAffr0QUhIiMbkifHx8fDy8hKG/zbkdsGbNm3CM888IzzevXs32rZti8zMTIwaNQpffPEFrK2tMWzYMIwePRo9evSAQqGAUqnEnj176vVdRkZG2LJlC3777TdhAv8xY8bg4cOH9S63tbU1Tp06BYVCgcTERKxatQqWlpZ1vguo+je2a9cu/PLLLwgKCsKAAQPg6uoKW1tbODk5wc3NTeN30b59e+H3VdOZntatW2PhwoXYuXMnTE1N8e677wq/L39/f+FzO3bsQGBgoPA5fft9hYaGYujQoZg+fXqVO7GtW7cOAQEBCAwMxJAhQ7B48WKN152cnLBnzx4sXLhQSIAB5Z3uhg0bcP/+fQwePBhBQUHIzc2Fu7s7hg4dipSUFGzZsgX//e9/kZ+fj+joaFy5cgVvv/02cnNz8dprr8Hb2xvTpk2r8hvV13Zxzpw5wh3Phg0bhiFDhjS4XXR2dsaUKVMQGhqKsLCwKgcSNbG1tcXWrVvx6aefQqFQwNfXF++++269kgpqfn5++OmnnxAQEIDZs2dj9erVEIlEVe6WO2DAADg7O8PHxwd9+vTR+K3VprbYLCsrw5tvvomff/4Zv/32m0ZsPs78+fOFYfhyuRzPPfccpk2bhnfffRe5ubl44403hASFm5sbpk2bVm0fL5PJhDZInfABNOM3OjoaHh4ekMlkUCgU2LRpEzw8POoUy5GRkVizZg2++uorvP766/jjjz/QvXt3HD9+HOfPn4efn59wp6z8/Hz4+flhxIgRsLa2RmFhYbO0QZs2bcKYMWOeqI9vaCyfPHlSuFFEYGAghg0bho8//hijRo2CTCaDVCrFqlWroFKpMGzYsDqfPa6tf/3ggw/w8OFDeHt7o3fv3nW6VESdkH7nnXfg6+uLTZs2QS6XIywsDOnp6cLE0v7+/lAqlfjrr78wadKkKstZsWIF/u///g/+/v4a7aatrS22bNmCv//+GxYWFjA3N0dMTAxkMhmGDh2Ko0eP4j//+U+1+6h79uzB7NmzhSklKmvdujU+//xzzJ07F3K5HD/88AOmTp2qs21xdUpLS4W+a+DAgVX61I8//hgLFy6EQqHA0qVLhUufc3NzhYmx5XI5SkpKNG7WVBcrVqzAokWL4O/vj3feeQchISH1OtFbl77yzJkzWLFiBeLj43Hx4kW88cYbdV7+ypUrkZGRAX9/fwQEBGDAgAHIyspq0P5jdcdSeXl5jz2OWrJkCd58803cuXMHO3bsQERERL3XQXVaSoyWlpYK+2fqKViqU9t+X0uK0crbY8uWLXjqqaeQkJAgnBy/evWqkCTr1asXCgoKoFAo6lWu+qpP/IWHhyM5ORlvv/02Ro8ejS1btgjH8erXgPJjh7i4OJ2Lv8p1qPibKiwsFP7OysoSrhjQlfirTsV9vjVr1uCll17ClStXsGPHDo0BE5WNGDECa9asQWBgIN8IoLkmTzMUvr6+tG/fPq2WQT35/OTJk8nHx4esrKyoqKjoiZZZcVLiphYXF0dHjhwRHq9YsYLs7OwoKiqKoqKi6JdffmmWctRHTTcbKCoqInd3dyopKWnQcouLi4UJIx88eEARERGNXv/JkydT165d6YUXXnjiOGlqT/L70oU40leVt5uuxFxCQgI5OTlRWFhYtZOuVp68XS0zM5MCAgKavHyGFtMtIW4qx/LjYqS5tdSY+PPPPykyMpK6d+9O8fHxwoT/uqQlxF9T0Wbc5OfnCzdP+eeff8jR0ZGuXbvWZN/XqVOnJlt2U+L4057GjFFdir/KMae+QVFJSQmNGjWKunbtKkxer4tqqt8PP/xAISEh1KVLF5o8ebKWS9nw+Ku8f6Jv26+5iYgqTMzAnoiXlxd8fHywfv16Yf4rbfj000/x66+/QqVSwcrKCt9++22VCczr69SpU+jTp49wiR/TdOXKFSiVSo254xrDnTt3kJCQAJVKhUePHgmX+zXXHD0tSUv5fbH60eftlpqaijfffPOJ57RhukGfY5kxbdm5cyf+85//ACif++n999/H888/r+VSMfY/HKNMmzj+WoZmTZqNHz9eYz4WtYMHDwp3TWlJ7ty5I9xZo6KrV69WmcAWABYsWFDtxH71ERwcXOXyCT8/P6xcufKJlsv0H/++Hq8l/r54uz1eS9xurCqO5cfjWG65OH5bFl3bHs1B19YJx6h+0bX6cvyxxsQjzRhjjDHGGGOMMcYYq6R+t5FhjDHGGGOMMcYYY8wAcNKMMcYYY4wxxuooNTUV0dHR6NGjBzZs2IB9+/YhPDwcXbt2xcmTJwEAt27dQmxsLCIiIpCUlKTlEjPGGGsovjyzGmVlZcjMzISlpaVBTrhuaIgI+fn5cHZ2hljcMvLIHIOGg+OPaRPHH9Mmjj+mbQ2JwYcPH2Lw4MFYv349jI2NAQBRUVHYuHEj8vPzMX78eGzbtg1vvfUWnn76aXTv3h2RkZFISUmBiYnJY5fPMWg4uA1k2tQS46+lMtJ2AVqizMxMuLm5absYrJn9+++/cHV11XYxAHAMGiKOP6ZNHH9Mmzj+mLbVJwbVE2337dsXZmZm+O677yCRSGBrawtbW1tkZ2cDAA4fPoy5c+dCLBYjODgYp06dQnBwcJXlFRUVoaioSHh848YN+Pr6Nk7FmE7gNpBpU0uKv5aKk2bVsLS0BFAeQFZWVsLzJSUl2LlzJ2JjYyGVSrVVvCZjqPXLy8uDm5ubsN1bgupi0FC3j77g+Gv59L2O1dVPV+IPMMzto084/lo2Q61fQ2Lw9u3byMjIwKFDh5CcnIzp06drxIqRkRGKi4tRUlIijN6wtrYWkmmVzZ49Gx999FGV55csWQIzM7M6l4vpnsLCQrz00kvcBrYAhli/ltgHt1ScNKuGeiiqlZVVlcbCzMwMVlZWevtjMuT6taQhyNXFoKFvH13H8dfy6Xsda6tfS48/wLC3jz7g+GvZDL1+9YlBGxsbREREwNjYGNHR0Zg+fbrGQWdpaSmMjY0hlUpRVlYGsViM3Nxc2NnZVbu8qVOnYtKkScJj9YFs//79NfrgXbt2ISYmRm+3jyHWLy8vDy+99BK3gS2AIdevJcVfS8VJM8YYY4wxxhirg5CQEMydOxdEhPT0dPj6+uLy5cu4f/8+8vPzheRYSEgIUlNT0a1bN/z111+YM2dOtcuTyWSQyWRVnpdKpVUObqt7Tp8YWv30ua6M6RNOmjWA/4wdOP9JH20XgzGmIzze3QqZhDAntHGWd/DgQUydOhVA+dwTTz/9NI4dOwaVSgWJRIIxY8ZgxIgRuHXrFkaOHImCggJMmDABw4cPf+Lv9p+xA0UqEa589vQTL4uxhuA+mGkTxx9r1aoVnn32WURFRUEkEmHp0qW4ceMGevfuDZFIhO+++w4A8M4772DkyJGYNm0axo8fD1NTUy2XnGlDY+8Dahu3gcwQcdKMMcZ0TJcuXZCamgoAGD16NPr3749jx47h999/h4WFhfC+zz//HFOmTBHu3DVw4MA63bmLscdJTU3FrFmzUFZWhjfeeAOtW7fGlClTIBaLsWjRIgQEBDRJ0pYxxlqCV199Fa+++qrwuH379jhw4IDGe5ycnLBr167mLhpjjLFGxkkzxhjTUcXFxTh8+DCWLl0KsViM3r17w8bGBgsWLIC7u3uD79yVl5cHoHz+g5KSEuFvAJCJSeOxPqlcV31TXf0aUteHDx9i7ty5+P3332FsbAwAiIqKwtatW5Gfn4/x48dj27ZtnLRljDHGGGM6j5NmjDGmo5KTkxEdHQ2xWIy1a9fC3t4ef/zxB15//XVs3rz5ie/ctXPnzip37poVXAYA2LZtWyPXpuXQ95EBFetXWFhY788fPHgQpqam6Nu3L8zMzPDdd99BIpHA1tYWtra2QpzVNWnLGGOMsbrj0d6MNS9OmjHGmI5au3YtEhMTAQD29vYAykf8vP322wDwxHfuio2NrXLnrg+OilFUJsKpGXFNWTWtMMS7d6lHFdbH7du3kZGRgUOHDiE5ORnTp0/XuMOWkZERiouL65y0rctIR/VjoHy0oz6OBuSRjowxxh6HR3sz1vw4acYYYzqopKQER44cwY8//gigPNFgZWWFM2fOwNbWFkDT3LmrqEyEIpVIL5NKaoZ0966G1NPGxgYREREwNjZGdHQ0pk+fDktLS+H10tJSGBsb1zlpW5+RjkD5aEce6ai7nnSkI2OMGTIe7c1Y89PZpBkPS2WMGbLk5GT07NlTGMnTs2dP4c5cCxcuBMB37mJNIyQkBHPnzgURIT09Hb6+vrh8+TLu37+P/Px8ITlW16RtXUY6ApqjHf/6ML5pK6kFPNKRMcbY4/Bo76bBo71ZbXQyacbDUhljhi4hIQEJCQnC46NHj1Z5D9+5izWFVq1a4dlnn0VUVBREIhGWLl2KGzduoHfv3hCJRPjuu+8A1D1pW5+RjkD5aEd9TCqp8UhHxhhjNeHR3k1L3/ebebR3w+hk0qyxh6Vyhr2cIWbYq3vMGGOsdq+++ipeffVV4XH79u1x4MABjfdw0pYxxhhrXDzau2nwaG9WG51MmjX2sFTOsGvS94OcyvXjLDtjjDGmO3iKDsaYoeLR3k2LR3uz6uhk0qyxh6Vyhr2cIWbYAc6yM8YYY7qCp+hgjBk6Hu3NWPPSyaRZYw9L5Qy7JkPKsKsfM8YYY6zl4yk6mgZP0cEYY4xVTyeTZo09LJWxhli9ejXeeOMN3L17F2vXrsX8+fNhamqK5cuXw9XVFefOncO4ceNQWlqKWbNmITo6WttFZowxxnQaT9HRtPR9ZApP0cEYY6y+dDJpBvCwVKZdKpUKa9euhZubG0pLSzFv3jz88ccfOHLkCGbNmoXvv/8e7733Hn788Uc4OjoiISGBk2aMMcbYE+IpOpoGT9HBGGOMVU9nk2aMadPq1asxaNAgzJ07FxcvXoSPjw+MjY0RERGByZMnAwAyMzPh6ekJALCzs8O9e/fQqlWrapdXl8tDDPXSCX0gkxBkYgLAl4YwxtiT4Ck6mhZP0cEYY4xp4qQZY/WkUqmwZs0abNy4EXPnzkVOTo7G2WiVSgUAKCsrE55TXxpSU9KsPpeH6PvoSX2s35zQ//3Nl4YwxljD8RQdjDHGGGtOnDRjrJ6SkpIwePBgYa4UGxsbjeH9EokEAITXAdR6aQhQt8tDDPXSCX3gP2MHZGLCrOAyvjSEMcaeEE/RwRhjjLHmwkkzxurpzJkzOHbsGJKSknDx4kUsWLAAZ8+eRXFxMY4ePQq5XA6gfIf90qVLcHBwqHWUGVC/y0MM7dIJfVCkEgl/86UhjDHGGGOMMaYbOGnGWD19/vnnwt/BwcFYtGgRfv31V3Tv3h0mJiZYvnw5AOCTTz7B6NGjoVKpqr30kjHGGGOMMcYYYy2X+PFvYYzV5OjRowCAIUOG4MCBA9izZw/c3NwAAL6+vkhLS8OBAwcQExOjzWIyxhhjjLFGtHr1arRu3RoAsHbtWoSHhyM6OhrXr18HAJw7dw7dunVDeHg4du/erc2iMsYYewI80owxxhhjjDHG6kilUmHt2rVwc3NDaWkp5s2bhz/++ANHjhzBrFmz8P333+O9997Djz/+CEdHRyQkJCA6OlrbxWaMMdYAPNKMMcYYY4wxxupo9erVGDRoEMRiMS5evAgfHx8YGxsjIiICJ06cAABkZmbC09MTVlZWsLOzw71797RcasYYYw3BI80YY4wxxhhjrA5UKhXWrFmDjRs3Yu7cucjJyRHudK5+HQDKysqE56ytrWu8KVRRURGKioqEx+q7apeUlKCkpET4u+L/+kaf6yeTEGRiAlC1fvpYX8b0ESfNGGNMx1y5cgUhISHw8/MDUD6XSmpqKubPnw9TU1MsX74crq6uOHfuHMaNG4fS0lLMmjWLLw1hjDHGnlBSUhIGDx4Msbj8gh0bGxsh0QUAEokEAITXASA3Nxd2dnbVLm/27NnV3jBq586dMDMz03hu165dT1z+lkwf6zcn9H9/V65fYWFhM5eGMdYQnDRjjDEdFBUVhXXr1gEAz6fCtGL16tV44403cPfuXaxdu5aTtowxg3DmzBkcO3YMSUlJuHjxIhYsWICzZ8+iuLgYR48ehVwuBwA4OTnh0qVLcHBwqHGUGQBMnToVkyZNEh7n5eXBzc0NsbGxwgi2kpIS7Nq1CzExMZBKpU1fyWamz/Xzn7EDMjFhVnBZlfpVTLYyxlouTpoxxpgO2r9/PyIjIxEZGYkRI0ZozKcyefJkAP+bTwWAMJ/Kk14aUtMlBvpAny8PAaqvX0PrypNgM8YM1eeffy78HRwcjEWLFuHXX39F9+7dYWJiguXLlwMAPvnkE4wePRoqlarakWRqMpkMMpmsyvNSqbRKAqm65/SJPtavSCUS/q5cvyepK5+4Yqz5cNKMMcZ0jJOTEzIyMmBmZoaxY8fit99+e6L5VOpzacis4PJlbtu2rVHq0hLp4+UhFVWsX0MvDVFPgj137twqk2A3RdJW/RgoT9zqY2KTk7aM6Z6jR48CAIYMGYIhQ4ZovObr64u0tDRtFIvpOT5xxVjz0umkGWfYGWOGqOJZ6QEDBuCnn36ChYWF8Hp951Opz6UhHxwVo6hMhFMz4hq9Xtqmz5eHANXXryGXhjT2JNj1SdoC5YlbTtrqrsZI2jLGmCFrzBNXjLHH09mkGWfYGWOGKj8/H5aWlgCAtLQ0PP3001i8eHGD51Opz6UhRWUiFKlEeplUUtPHy0Mqqli/htSzsSfBrkvSFtBM3P71YXy9y93ScdKWMcbY42jj7q3qxwCP9tZVPNr7yehs0owvDWl8hthYVPeYsZZu3759mDZtGszMzNC2bVvMmjULJiYmDZ5PhbH6aOxJsOuTtAXKE7f6mFRS46Rt3fDVBowxQ6TNu7cCPNpb1/Fo74bRyaQZXxrStAypsQC4wWC6JyEhAQkJCRrP8XwqrLk09iTYjNUXX23AGDNU2rh7K8CjvXUdj/Z+MjqZNONLQ5qGITYWgO41GP4zduD8J320XQzGGONJsJlW8NUGjY+vNmBMN2jz7q0Aj/bWdY012tvQ6GTSjC8NaVqG1FioHzPGGGOs5eOrDZoWX23AmO7gE1eMNQ+dTJrxpSGMMcYYY4aHrzZoGny1AWOMMVY9nUyaVcQZdsYYY4wxw8BXGzQtvtqAMcYY06TzSTPGGGOMMWYY+GoDxhhjjDUnTpoxxhhjjDGdw1cbMMYYY6ypiR//FsYYY4wxxhhjjDHGDAsnzRhjjDHGGGOMMcYYq4STZowxxhhjjDHGGGOMVcJJM8YYY4wxxhhjjDHGKuGkGWMNcPjwYXTp0gXdunXD888/j5KSEqxduxbh4eGIjo7G9evXAQDnzp1Dt27dEB4ejt27d2u51IwxxhhjjDHGGKsrTpox1gBubm7Ys2cP9u7dCw8PD2zatAnz5s1DamoqZs6ciVmzZgEA3nvvPfz444/Yvn07PvzwQy2XmjHGGGOMMcYYY3VlpO0CMKaLnJychL+NjY1x/vx5+Pj4wNjYGBEREZg8eTIAIDMzE56engAAOzs73Lt3D61ataqyvKKiIhQVFQmP8/LyAAAlJSUoKSkR/gYAmZiEv/VJ5XrqE5mEIBMTgKr108f6MsYYY4wxxpg+4KQZY0/g6tWr2LlzJz777DPcvXtXeF6lUgEAysrKhOesra2RnZ1dbdJs9uzZ+Oijj6o8v3PnTpiZmWk8Nyu4DNu2bWusKrQ4u3bt0nYRGt2c0P/9Xbl+hYWFzVwaxhhjjDXU4cOHMXHiREilUri4uGDFihXYuHEj5s+fD1NTUyxfvhyurq44d+4cxo0bh9LSUsyaNQvR0dHaLjpjjLEG4KQZYw2Ul5eHESNG4KeffoJKpRJGhwGARCIBAIjF/7sCOjc3F3Z2dtUua+rUqZg0aZLGst3c3BAbGwsrKysA5SOSdu3ahQ+OivHXh/FNUSWtUtcvJiYGUqlU28VpVP4zdkAmJswKLqtSv4pxwxhjjLGWTT1Fh6mpKaZOnSpM0fHHH3/gyJEjmDVrFr7//nthig5HR0ckJCQ0StLMf8YOFKlEuPLZ041QE8YYY3XBSTPGGqC0tBRDhw7F9OnT4eXlhZKSEpw9exbFxcU4evQo5HI5gPLLOC9dugQHB4caR5kBgEwmg0wmq/K8VCqtkkAqKhPpXVKpourqrOuKVCLh78r1a0hdqzvL7evrCxcXFwDA+++/j5iYGD7LzRhjjDWyxp6igzHGWMvGSTPGGmD16tX4888/MWvWLMyaNQsTJkzAm2++ie7du8PExATLly8HAHzyyScYPXo0VCpVtZdfMtYQ1Z3ltra2Rmpqqsb7muIsN2N8aRJjjDXeFB31nde24mN9wfPa1h33wYw1P51MmnFjwbRtxIgRGDFiRJXnhwwZovHY19cXaWlpzVUsZiAqn+UWi8V48OABoqKi4OLigm+//RZ2dnZ8lps1CW1emsQYYy1BY07RUd95bQHo7dy2PK/t43EfzFjz08mkGTcWjDH2v7Pc06ZNQ1RUFOzt7bFixQpMnz4dCxYs4LPc9aTPZ7qB6uvXkLpq4+7BFcvKdxDWTY0VfwCfPGXa1dhTdNR3XtuiMhFOzYhr+oo2I57Xtu748mDGmp9OJs14h71pGOIOe3WPGdMFFc9yS6VS2NvbAwAGDhyIJUuWAOCz3A2lj2e6K6pYvye5e6s27h4M8B2EdV1jxB+fPGXa1NhTdNR3Xtsilf7Obcvz2tZdc14erH4M8HGwrmrME1eGSCeTZmq8w940DGmHHXiyg0bGtKHyWe7i4mIQEWQyGdLS0tChQwcAfJa7vvT5TDdQff0aevfW5r57cMXy8x2EdVNjxh+PtGDaxFN0MG3T1uXBAB8H67rGOnFqaHQ2acY77I3PEHfYgYbvtDOmLdWd5Z4zZw7Mzc0hk8mwdOlSAHyWu6H08Ux3RRXr15B6avPuwQDfQVjXPWn8VdQYJ095lEU5QxxlUd1jxlo6bVweDPBxsK5rzBNXhkgnk2a8w960DGmHXf2YMV1S3Vnuyme4AT7LzZoG3z2YtQSNdfKUR1loMqRRFgCPtGC6R5uXBwN8HKzrGvPElSHRyaQZ77Azxhhj2sGXJjFta8yTpzzKopwhjrIAeKQF0z3cBzPW/HQyacaNBWOMMcaYbvB4dytkEsKc0MZZXmOePOVRFpoMaZSF+jFjjDFWG51MmjHGGGOMMcPEJ08ZY4wx1lw4acYYY4wxxhhjjDGmRzze3QoAjTra2xCJH/8WxhhjjDHGGGOMMcYMCyfNGGOMMcYYY4wxxhirhC/PZAahsSchZowxxhhjjDHGmH7jkWaMMcYYY4wxxhhjjFXCSTPGGGOMMcYYY4wxxirhpBljjDHGGGOMMcYYY5XwnGaMMcYYY4wxvcTz2jLGGHsSPNKMMcYYY4wxxhhjjLFKOGnGGGOMMcYYY4wxxlglnDRjjDHGGGOMMcYYY6wSTpoxxhhjjDHGGGOMMVaJQSTN3nnnHURGRmLEiBEoKSnRdnGYgeH4Y9rE8ce0jWOQaRPHH9Mmjj+mbRyDjD05vU+aHT9+HDdu3EBaWhq8vb2xbt06bReJGRCOP6ZNHH9M2zgGmTZx/DFt4vhj2sYxyFjj0Puk2YEDBxAbGwsAiI+Px/79+7VcImZIOP6YNnH8MQDweHcrPN7dCv8ZO5r9uzkGmTZx/DFt4vhj2sYxyFjjMNJ2AZpaTk4OnJycAADW1tbIzs6u8p6ioiIUFRUJj3NzcwEA2dnZGsNYS0pKUFhYCKMSMbKyspq45M1PXb+srCxIpVJtF6dRGZUWwKiMUFhYVqV++fn5AAAiavTvrUv8AXWLQY4/3aVv8acqE3EM6hCj0oLy/6uJwaaMP4D74PrQ1/gDam4DOf5aDkOMP4D74JaC44/bQG3T1xjU5j6gPtH7pJmNjQ3y8vIAlDcCdnZ2Vd4ze/ZsfPTRR1Web9u2bY3LbfVF45WRNZ8XanktPz8f1tbWjfp9dYk/oP4xyPGnm/Ql/gCg1dzGKSNrXjXFYFPEH8B9MNPE8ce0iftgpk3NHX8At4FMU3P3wfpERHqeWkxPT8e8efOwYsUKfPrpp2jbti2ef/55jfdUzrCXlZUhOzsb9vb2EIlEwvN5eXlwc3PDv//+Cysrq2arQ3Mx1PoREfLz8+Hs7AyxuHGvWK5L/AF1i0FD3T76guOv5dP3OlZXv6aMP4D74PowxPpx/LUchlo/7oNbBkOtH7eBLYch1q+p40+f6P1IM6VSCUdHR0RGRuKpp57C5MmTq7xHJpNBJpNpPGdjY1PjMq2srPTyx6RmiPVrqux6XeIPqF8MGuL20Sccfy2fvtexcv2a8uwi98H1Z2j14/hrWQyxftwHtxyGWD9uA1sWQ6sfjzCrG71PmgHAF1/wGFKmPRx/TJs4/pi2cQwybeL4Y9rE8ce0jWOQsSfH4/AYY4wxxhhjjDHGGKuEk2b1IJPJMH369CpDWPUF169l0/XyPw7Xr2XT9fLXhb7XUdfrp+vlfxyuX8um6+V/HK5fy6br5X8crl/Lpw91qA3Xj9VG728EwBhjjDHGGGOMMcZYfWl9pNmMGTPw6NGjZv/e1NRUKJXKen8uPT0dv/zyS+MXCFXXxejRo/HVV181yXfpssaIGZFIhPv37zdb/M2YMQNvvvmm8Lgx4q/yMp8Ux1/dPGnMzJgxQ4i/5sRtnu5qzDZPG99d0U8//YRz58499n0tMV614f79+/jss8+0XQxBQ+Jh8eLFwpw6Dd0+6ngIDg5GampqnT/H8dDy9OnTB59++imAJ98+9YmHym1KxceZmZmIjIwUXtu0aRN8fHygVCqxdu3aOpfxww8/xMqVK+tafK3QhxhqjDZFrb5tSm0a2l96eHggPT29wd9raP1lTTHMsa2pMWO7oa5cuVLrTSXq66uvvsKtW7fq9N6NGzfi0KFDjfbdWk+affTRR1pJmjVUUzYwTbEuSktLG3V5LUFjrieOv//h+KubJ11PH330USOWpulxzGmfNtupxv7uuibNKqvrdm2ug4DmiLPS0tIWdxDQkHgYP348/vOf/wBo/oO0xv6+mra7ocZDQ2zduhWvvPIKgJZz0O7s7Iy0tDTh8eLFi/Hhhx8iPT0dBQUFdS7jzJkzMWzYsAaVoT4xVFZWhrKysgZ9hz7EkDbblNoY+jFFc7SPNcUwx3bTain71tpMmoEa2YEDBygiIoLkcjkFBATQxo0b6e2336bg4GBSKBQUGRlJ586dIyKil19+mQCQv78/KRQKun37do3L3b59OwUGBlJAQAB169aNTp8+TUREKSkp5OfnRxMmTCC5XE6+vr505MgRjc9FRERQUFAQhYSE0J49e+r0uRUrVlBAQAAFBARQ79696fr163T79m1yc3MjKysrUigU9PLLL9d7/Vy8eJGio6MpICCAFAoFbdiwocZ1MWrUKBo3bhz17NmTPD096dlnn6WioiIiIiouLqZ33nmHQkJCSKFQ0KBBgyg7O5uIiEaNGkWJiYkUGRlJXl5eNZZl2bJl1KNHD+rbty/5+PhQZGQkXb58mYiITpw4QRERERQYGEg+Pj40a9YsIiJ6+PAhOTo60rVr14TlTJ06laZMmUJERO7u7vT+++9Tly5dyNXVlRYtWkRLly6lsLAwcnd3p9WrVwufO3z4MPXo0YM6depESqWS1qxZQ0REly9fJmtra/rwww8pKCiI2rdvT1u3bq1xPVVUMf4A0MCBA6l169Ykk8noqaeeEuIPAI0ePZoAUKtWrcjU1JR8fX014pOICAB98skn1LFjRzI2NiZXV1fq1q0bnTx5kl599VVyc3MjmUxGrVq1ooCAgBrjz8nJiWxtbUmpVJKXlxeFhYWRt7c3TZgwgdq1a0empqbUsWNH8vX1pSVLlgjxZ2trS25ubhQREUFt27YlU1NTIf46depEEydOJCKi06dPk5+fH23bto3jr5njr6Y2DwA5ODiQubk5ubu7U1JSkvB59TJu375dY/u4fft2AkCOjo5kaWlJzs7OtHTpUqHt8vf3J5lMRjKZjLy8vOjhw4caMVexzZs+fTp1796dLC0tycbGhszNzcnLy4uOHDlCpaWlNHnyZHJxcSGZTEb29vYUHx9P169fp8GDB5O5uTlJJBIyNjYmDw+POsUAx1zztXkVAaD333+flEoleXp6UlJSksZrOTk5dODAAXJ2diZTU1MyMTEhX19fSkxMFGIQAL355psUEhJCFhYWGt+dmJhI3t7eJJfLKSgoSIg5ddxV7KO//fZbCggIoPbt25NMJqP4+HhydXUlkUhETk5OpFAoaOvWrbR161ZycXEhExMTMjU1pf79+1NRURGlpKSQjY0NeXt7k7m5ORkbG9ORI0doxYoVFBoaSu7u7mRmZkaenp7Uu3dv2rRpEwUEBJBUKiWxWEyurq4N6qOjoqLotddeo+DgYGrfvj1NmjSJysrKhNdef/11CgsLox49egi/HT8/P/Lz86PXXntNiFd1THbp0oU8PT1p5MiRVFhYSEREeXl59NJLL1FISAgFBATQ2LFjhc9V/o64uDgSi8VCm3/kyBHy8vISykRE1KVLl1rb/unTp9OgQYOoT58+5OnpSU8//TSdPHmSQkNDycTEhGxsbIS2a+XKldSmTRuN+FC3SUOGDCEAJJPJyNramqKiomjZsmVCfSv+dsPDw0mpVFJAQAC5u7vT8OHD6fbt29S6dWsSi8Vkb29PdnZ25OvrK8SmOo6MjIzI39+fQkJC6JtvviGFQkEeHh5kY2ND9vb21K5dO/L19aXff/+dBg0aRCEhIeTi4kIODg6P3WdbunQpKRQKksvl1KlTJ7p8+TKVlJRQbGwsderUiXx9fen555+nBw8eEBGRQqEgW1tbatWqFRkbG1OfPn30Nh5iY2PJ09OThg4dSiqVioiIVq5cSaGhoaRUKkkul9PmzZuF5Zw9e5b8/PyE/RILCwt6/fXXiYjI19eXWrduTRYWFmRiYkK9evWioqIioT2zt7cnPz8/cnZ2rrJPrY4HdZsikUgoNDSUTp8+Tfv37xfaFC8vLzIxMSF3d3c6cuQI3bx5kwYNGkQdO3YkMzMzatOmTa37/e7u7uTp6UlE5W2wqakpBQQEkL29PUkkEnJ1daXg4GByc3Mjc3NzsrCwoFatWmm03ZWNGjWK5s+fL6znwYMHk52dHdna2pKlpSW1bduWJk2aRAcOHKCgoCAyNzcnOzs7atu2LfXo0YNGjBhBPXr0EGJIqVTStGnThFh0d3cna2trMjY2poEDB9Ibb7xBwcHB5O/vT23atBGOW8aOHSvsv1paWpKlpSW1b99eI4YsLS3J1NSU2rdvT46OjlRWViasQ/V+y/vvv9/kMRQWFka+vr707LPPUkxMTI1tSsX9genTp9PEiRNrPC6r2KYQEdnb2wt9/f79+0mhUJCfnx+NHj2aFAoFpaSkEBEJ9Q8JCaF27dqRq6urxr4dEWnss8nlcurQoQNNmDCB7OzsCAC1b99e6KurO5YkqtpfOjk50bFjxyglJYWcnJzI0dGRfH19ydnZmUJCQmjo0KHk7+9PHTp0oE6dOgn7d+PGjSNfX1/y8PAga2trGjNmDMnlcjIyMhLagEGDBlGHDh0oICCA/P39SSqV0qVLl+jrr78mmUxGUqmUTExMqFWrVnTp0qVqt3dN/WJUVBQ5OzuTg4MDyWQycnNzo5KSEpo8eTKZmZmRnZ0dOTo6Uvfu3am0tJT8/PzIxsaGTE1NydramoYNG0aFhYU0atQoGj58uLAcGxsbSkxMpKKiIsrLy6M2bdqQg4MDmZmZkZOTE8XExJBYLCZzc3Nq3bo1BQQEkJmZmUb7aG5uTl999ZVW28eWGtv+/v41/rbVUlJSyNfXl1588UVSKBS0Zs0aunDhAvXu3ZuCg4MpICCAFixYILz/hRdeoE6dOgmxfvPmTSL63/5tdWrbr//hhx/Ix8eHFAoF+fv706FDh+ijjz4iqVRKHTt2JIVCQceOHaPk5GQKCwsjpVIpHEMTEW3dupVsbW3J2dmZFAoF/fDDD0REwj5kYGAgRUZGUnp6eq3roaJGTZplZWWRg4MD7d27l4iIVCoVZWVl0Z07d4T3rF69muLi4v5XgEobvzq3b98mOzs7OnHiBBERJSUlkY+PD5WVlVFKSgpJJBI6dOgQEREtWrSIYmNjiYjo0qVLFBYWRrm5uURUfvDWpk0bevToUa2fO3nyJDk6OtKECROoa9eupFAohDIvW7aMnnnmGUpMTKSuXbsKSYu6Cg0NpcWLFxMR0YULF8jOzo6uXLlS7boYNWoUhYaGUkFBAZWWllJ4eDitWrWKiIg++eQTmjlzpvDemTNn0iuvvCJ8Ti6XU15eXq1lSUhIIJFIRH379qXi4mL6/PPPKSYmhlJSUsjZ2ZkiIyOpZ8+eVFhYSEqlkg4ePEhERO+99x5NnTqViIgePXpEbdq0Eerg7u5Ob775prC+TUxMhIPPw4cPU6tWrYiIKCcnh5RKJWVmZhIR0d27d8nNzY2uX79Oly9fJgC0bt06IiL6/fffqWPHjkK5a4qZivF3//59AkBGRkb0xx9/0KVLl8jW1pa++eYbio2NJQAUFhZGAOjixYvCMqqLz+nTp5OdnR1t2rSJLCwsaPny5dS2bVvy8vKi3bt3k0QioeTkZFKpVDXGn3oZDg4O9OjRI3rmmWdIJBLRoUOHKDs7mxYuXEixsbGUlZVFbdq0odatW9OECRPIwcGBTE1NqVevXlRaWip0mImJieTm5kZKpZJSUlLIx8eH/vrrr1q3NxHHX2PHX21tHgBycnIimUxGnTp1IltbW2HZAOiFF16grl270tixY4Vlq+NP3eYBoC+//JKSkpKoffv2ZG5uTsnJySSRSMjd3V2IuZ49e5JKpaqxzXv//ffJ0dFRaPOGDBlC/fr1o9jYWPruu+8oODiYHB0d6cqVK5SQkECxsbHUtm1bIenXu3dvIQaSkpIoMTGR3N3dKSwsrNoYaKkxt2zZMjI2NqYzZ87QlClTyMPDg5ycnKi4uJjy8vKEvsHFxYUsLCyoU6dOLS7magNAOMBKT08niURCpqamdPLkSQJA//zzDzk4ONCGDRsoMTGRIiIiqGfPntSjRw+NZagTj2fPniUAdPfuXfr777/J29tb2Fm8f/++8Hd1fbRMJqP9+/cLfe2uXbuIiMjT05OUSiURlbeRbdu2pYiICHr06BGdPXuWZDIZffzxx5SSkiIkaIYMGULffvsthYaGUkJCAh0+fJhMTU3J29ub7Ozs6OOPPyYHBwdatWoVLVu2jPr161dtArcuoqKiqGfPnlRcXEwFBQXUqVMnWrlypfBaXFwcFRcXExHRd999R1FRUfTo0SMqKSmhhIQE+uyzz4ioPCZ9fX0pLy+PSktLqU+fPvTJJ58QEdHYsWNp+fLlNGXKFIqIiKD27dvT7Nmzhe9QJ4GioqKoS5cuVXY4w8PDaceOHURE9Pfff1OHDh00Dgoqmz59OrVt25ays7OprKyMunXrRp06daLWrVtTSkoKKRQK2rx5M2VlZdG9e/eExKx6ZzcmJoaIiIKDg4VYPHPmDMlkMo2DAPVvNzMzk4yMjMjV1ZXMzc0pOjqa7OzsqKysjN555x0SiUTUp08f6tq1K/Xo0UNYprr9UreVp0+fJrFYTFu3bqWUlBQSi8UEgFJSUmjRokVkb29Pqampwj5bVFQUrVmzhj7++GOKj48nIhLigaj8IMDDw0P4DRYUFFBBQQGVlZXRvXv3iIiorKyMxo8fT7Nnz6YpU6aQmZkZAaDt27cL8fDzzz9TYmIiWVlZkbu7e6PFg/r7x4wZQ3PmzKk25qo7AGloPLzxxhsUERFBDg4OFBoaSnl5eVRSUkLt27cne3t7ioqKosjISGFZly9fJkdHR3r06BERESmVSrKysqK9e/fSmTNnyNjYWDiIGjJkiBAPK1euJBsbG6EtB0Djx4+vsn3UAAh9xYkTJ8je3p7mz59P3t7e5OLiQl9++SVJJBL6+uuvCQC99dZbFBsbS7GxsbRq1SoKCwujy5cvk5WVFRkbG1OrVq009vv3799PiYmJ5OzsTFKplIjKkxgikUhIanh4eFBgYCARES1cuJCsrKyqbbsrq5w0c3d3p/DwcOrZsycNGjSIZsyYISQ9Vq1aJWxf9W+uc+fO1L59e3r06BFNnjyZZDIZtWnThoqLi0mhUJCRkRGdP3+eRo0aRba2thQcHCzE0IQJEyguLk6IIfX+Q1RUFHXt2pXMzc2ppKSELl68SGKxWNgeREQhISG0Y8cOio2NpR9++IE6dOhAxcXFFBcXV22CsLo2pXPnzkIMKRQK2rJlCxER3bt3r9oYmjJlCllYWFCXLl2ouLhYaFPeeecdcnV1JUdHR7Kysqp2f0CdWKgthqpLLBQVFZGrq6vQH+3YsUNoU4iIYmNjKTU1Vdi3Cw0NpTVr1gj7dkSkcUz7wQcfEADhWBKA0J8eOHCAjIyMhD5Y3S6p+8t+/fpR165dKS4ujqRSKR09epQGDhxIIpGIUlNTiYho5MiRJJFI6J9//qFLly6Rs7MzjR49moiIlixZQhKJhG7dukUpKSkkEono2WefFeLI39+fVCoVubu7k0QiobNnz1JycjK1b9+e4uPjadmyZWRiYiK07++88w6NGzdOYz1OmTKFunbtSg4ODtSjRw+NfrFbt25kZWVFMpmMYmJi6P79+8LxibrtiI2Npbi4OPrss8/ou+++I0dHR/L29qbs7GyKj48nb29v+uSTT2jUqFFkY2NDixcvFtrHTp060Zw5c2js2LHk5eVFcXFxVFRURGPGjKF3331XOHGjbh/Dw8Np+fLlZG1trdEO1tY+qus3fPhwoY2dPn06OTk5kbOzM3Xr1o1sbGzqHdtE5f3l0qVLiYhq7S+bKrZ/++03AkAmJiZC0k8dV48ePSIXFxfy9vauMY+hjin1Z0pLS6lTp0509uxZIirvPwMCAujw4cNVfhezZ88Wkny1Jc1q26+v2OYWFxdTfn4+EZXvcx87dkz4THZ2NpWWlhJR+THZU089Rf/++6+wntXtMRHRvn37KCEhQdhGe/fuJV9f32rLVp1GvTzz4MGD8PLyEuYFEIvFsLOzw65du9ClSxf4+/tj5syZ9b5u+88//0RAQAACAgIAAMOGDUNmZiZu3LgBAOjQoQM6d+4MAOjSpQsuXboEANi+fTsyMjLQrVs3KJVKDBw4EGKxGNeuXav1cykpKQgNDUVeXh7S0tLQp08f7N69GyqVCgBw69YtYSh3QUEBDh48WKd65Ofn4++//8aYMWMAAJ6enujatavGkPDKnn32WZiZmUEikSA0NFQo48aNG5GUlASlUgmlUonVq1fj8uXLwucGDRoES0vLGpd7/Phx5OTkICoqCp07d8a6deswbtw4pKamQqVSoV+/fnB3d8edO3cQFhaGq1evCtvtlVdewfLly1FUVIS1a9ciNDQU7u7uwrKHDBkirF8TExMMHDgQQPm11dnZ2bh//z4OHDiAf/75BwkJCVAqlejVqxcA4Pz58wAAExMTDBgwoMq2qU3F+DMzMwMA9OvXDwcPHsSwYcNQWlqKzz77DIcPHwYA/P777wCAJUuW1Bqf7dq1Q0BAAJ555hkYGRmhZ8+eyMrKQlFREebMmYPWrVtDoVBALBbXGH+LFy8GEcHIyAjXrl1D3759YWRkhM6dOyMrKwvr169HamoqevbsiezsbLRr1w55eXlISEhA586dhWvS27Zti9u3b8PZ2Rkvvvgirl69ilGjRmH79u0ICgqqdf1w/DV+/NXW5gGAmZkZTExMcOXKFXTr1g179+4VPuvk5IS0tDRcvHgR/v7+GvGnbvOA8vZu2LBhuHfvHiQSiRAfUqkUL774Iu7evYt//vkHYrG4xjYvNzcXoaGhQpvXpUsXiEQiXLp0CcnJyfDy8kJ8fDzc3d0xduxYZGVl4erVq4iLi4OHhwdu374txMC2bdvg7OwMBwcHnD17Fh07dqw2BoCWFXNq4eHhKC4uxo0bN3Ds2DHcuXMHa9aswcOHD/HSSy/hxRdfRHFxMaRSKV566SUALSvmHkddZl9fX8TExKBTp07Ca4cPH4aXlxf27t2LnTt34v79+/j7779x+PBhoQ0EgOzsbACAt7c3AOD27dto164dSktL8eKLL2L58uUoKSmBWFy+C1FdH61eb6tXr4abm5tQXysrK2Fo/fbt23Hjxg1cu3YNnTt3xtChQ2Fubo6tW7ciIyMDUqkU06ZNQ0BAAPLz83HhwgUcP34c8fHxICLIZDI8fPgQcrkcWVlZmDlzJjZt2oSsrCzY2to2aP0BwMiRIyGVSmFmZobhw4cjOTlZeG348OGQSqUAgOTkZIwePRoymQxGRkYYO3as8NsHgMGDB8PS0hISiQRjxowRlrNx40bMmjULP/zwAx48eIDc3FyNz8XExGDo0KFITU3FqlWrqpRv4sSJ+PbbbwEACxcuxCuvvAKRSFRrnWJjY2FrawuRSISgoCC0bdsW3t7e6N69OwIDA3Hp0iXY2dnh8uXL6NWrF8zMzODj44Pc3FwcO3YMeXl5Gn2jj48PunbtqvEd6t/u0aNH4ejoiOHDh2PgwIHo2LEjHjx4IOyzyWQyKBQKpKWlacx1om6/8vLy0Lt3bwwYMAAikQienp4AytuPdu3aAQCUSiWysrIwceJExMXFobi4GNevX8f58+fxyiuvYM+ePcI+m9rWrVsxYsQIODk5AShvn83MzEBEmD9/PgIDAyGXy7F161akpqbixo0b8PT0hLm5ObKzs4V4+Omnn+Ds7IzAwEB4eHjg6NGjTxwPX3zxBZRKJQIDA5GWloaMjAzhcxVjrjoNiYdOnTohKysL+/btQ4cOHeDo6AhLS0sYGRnB09MTAQEBSE1Nxbx585CQkAB/f3/0798f2dnZuHz5MvLy8nDixAnI5XJERkbCx8cHkZGRsLCwAFA+R9jNmzcRGhqKjz/+GMXFxfVq044eParRpvTv3x83btyASCRCp06d0KFDB7zxxhto164dfH19kZGRgd27d+M///kPjhw5Ah8fHzx8+BBisbjKfv/du3fh7OyMjz76CCqVCgcPHsTBgwdhZGQEFxcXAOWXa548eRIqlQoXL15EYWFhjW13beLj4yGVSjFy5EhERETg33//xfDhw2FkZIRZs2bh6tWrCAkJgYODAwAI6+zcuXO4efMm4uLiUFRUhHXr1gEAAgICcPbsWbi6umLp0qXIyMiAv78/li5diqVLl2L37t1CDAH/a4tffvllSKVS3Lp1C//88w+ICM8//7xQzsmTJ+Prr7/G7t278c4776CwsBAhISHIyMiosZ6V25Tu3bsLMRQYGIiLFy8CAC5fvlwlhn7//XdcuXIFjx49QkJCAtatW6fRpgwZMgTx8fGYOnVqtfsDDXXu3DkYGRkJ2zA2NlZoUwoKCrB7925MnDgRwcHBKCgoQFZWFs6fPy/s2wHQOKZdsWIFJBKJcCypri9Q3i8+99xzGDx4MAAI7dKBAwfg5OQEf39/pKWlwc3NDaWlpUhMTERhYSE6duyIqKgoAED79u1hbGyMtm3bYvv27Xjw4AHWrVsHpVIprJu8vDwAgLu7O06cOAGgfN7Aq1ev4tixY7C1tYWnpycuXLiA5ORkDBs2DHv27EFZWRk6dOggHC9V3u84fvw4bty4gbS0NJiZmcHLy0voFxUKBe7fv4/AwECYmJigS5cusLa2xvDhw7Fnzx6MHj0aYrEYI0aMwMsvv4xdu3YhOTkZHTp0wNChQ2Fra4tx48bB2NhYaAeLi4vx7bffolOnTjh16hTOnDmDjIwMbNy4EdevX8fZs2cRGhqKtLQ0XL16VSinun2cOHEiVqxYAUCzHaypfaxYP29vb+E3BpTvuz3//PP4448/MHr06HrFtrp9TE9Px8iRIwHU3l82VWz36dMH7u7uiIqKQmFhoRDbSqUS3t7eKCwsxLBhw2rNY7Rr106IxfPnz+P06dMYOnQolEolwsPDkZ+fjzNnzgAAVq1aheDgYPj7+2PJkiV1yvXUtl8fHR2NESNG4Ouvv8bly5eFvqWyrKwsDBo0CP7+/sLx+alTp6p976ZNm3D8+HF07twZSqUSr7/+OrKzs/Hw4cPHlhVohjnNrl27htdeew1JSUk4deoUfvnll0a/3tvExET4WyKRCNfdEhFiYmKQnp4u/FPvBNX2OQC4e/cuYmNjAQC9evXSmDsgOztbeC0+Ph779+9vcNkft2NTW90WLFgg1OvMmTPYtm2b8N6agkvtwIEDwgFSdXX4+eefsWfPHiQmJuL48ePo3r27sN1cXFzQrVs3/Prrr1i4cCFee+21WsusfiwSiSASiVBaWgoigp+fn8a2uXbtGnr27AmgfKdavW4kEkmVnd/HUe9gFhYW4tNPP0VSUhJ69uyJV155pcqP4+uvv641PivurKq3gUgkQnJyMqKjo1FWVga5XI6MjIwa42/8+PHo16+fRvyp6zd+/HgEBQXByckJ6enpcHBwQE5OjhBjvr6+QvyJxWIUFhYKr3l6euLRo0cNvmab46/x40/d5gHA7t270a1bNxQXFwtlUOvRoweuXbuGo0ePok+fPjXGX8X6iMViqFQqmJmZ4dSpU3jhhRdw9epV/Pvvv8jIyKixzbO3t4exsbGwLIlEgrKysmrnKBCJRMjNzRUSIq6urkISRSKR4MqVK4iNjQUR4fXXX8e4ceOqjYGaaCvmKjpw4IDwGxKJRNi/fz/ee+89tGrVCj/88APMzc0hFouxY8cOAC0/5qojlUqrva14UVERFi1ahC+++AKnTp3Cq6++ioKCAqENVL+nIpVKBWtrayHmzp07J7R5NTExMcHcuXNhYmKCW7duYc6cOcI6UdeNiODg4CAki9PT0/Hjjz/CzMwMp0+fhrGxMSwsLBAfH4/jx4+jrKwMo0aNQmBgILp374709HT8/PPPSE9Ph0QiwX//+1/Y2Njg7NmzwpxJjaFizNYWZ4+LbfXrRIRhw4bhq6++Qnp6OrZv3w4fHx/hfaampli/fj0iIyOxdOnSKssZMGAATpw4gWPHjmHz5s1ITEx8bB0qx2h1/RoAPPfcc/jnn39w8uRJPHz4EJaWljXus1Wub8XvEIlENd7WvqysTPj9qU84qFQqof0yMTHBtm3bsGbNGri5uWnss6m/U90+HTp0CO+++y6eeeYZZGRkYNq0aY/dDpWtWrUKe/bswR9//IGTJ09i8uTJuHnzplBGCwsLjT7qzp07wmudO3eucR+wPvGwfv164Tdw/vx5fP/998L7Hte2NSQesrKyhDq0a9dOY44YsViM48ePIzIyEgkJCXjppZdw6tQppKenw8LC4rHxcO3aNfz5558YOXKk0K+pVKo6z4mj7p/Uqvu+im1sxfidPHkyBg8ejIcPH6K4uBhDhgzB7t27NWJI3f5LJBKh/a+pLmqWlpY1tt21qak/8/f3x9atW2FsbIz169cL7ZVYLEZZWZlQxpKSEpiamgplNDY2Fl67e/cu8vPz0adPH9ja2mLNmjUwNzcXYqji91tYWFQ5zqlowIABOHXqFIgIEokEp0+fRnp6uvCbqkvdaqrr0KFDq8TQ4cOHhfUXFxcn1E+93tevX4/ff/8df/75Z7XLfJzKfWhtx50Vf4dAeZuyYMECBAUFVal/5WPaDz74QPicmvp7K6+TijGVl5encRwJAGFhYTh79iwkEonwPnU7py6fUqlEp06dkJ6ejhEjRmDSpElCbBsbGwvvDQ8PR2FhIXbt2gUvLy84OTkhOTkZycnJ6N69u/C+mvoBQHM/yc7OTkgUAeWTu7u6ugIAzM3NNZJYFVlYWNTYDqqfr/i6uh2cP38+unTpgu+//x5EBF9fX3z99ddCbKtvJKL+DqA8hs+dOweVSqXRDtbUPlasX+XjEPXvMjIyEsePH69XbDekv2yK2JZKpTAyMtJ47tChQ0hPT8fgwYOxceNGTJs2rdY8RsW+h4hgZ2en0Q5evnwZo0aNwr59+/DNN99g27ZtOHXqFObNm1enXE9t+/Xr16/HZ599hpKSEvTu3bvGud3Gjx+Prl274uTJk0hPT0fHjh1r/G4iwqhRozTqcPPmTZiamj62rEAjJ83Cw8Nx8eJF4QxHWVkZLl++DKlUCicnJxCRkO1Vs7S0RG5ubq3LDQsLw8mTJ4Ud+l9++QUuLi7CGaGaxMXFITk5Wci8AxBGGdWmR48eOHXqlBDAW7duhaOjIyQSCaysrFBQUAArKysAgLW1tXBA+TiWlpYICgrCsmXLAAAZGRnYt28funXrJrz+uHWh1r9/f8yfPx+FhYUAypNDp0+frtNnASAnJwempqY4ePAg7t27h+zsbCxZsgQ9evRA586d0atXL7zxxhvYtWsX1q9fr3G2FCg/s/n+++/j/v37Qla7PsLDw3H58mWNM/jp6elCcqE2Na2nyvEHABcuXICRkRGKi4uxd+9eYadATd1o1RSfABASEqIRf//973/Rpk0b2NjYICQkBE5OTvDw8BCy7WqV42/btm3C6LatW7cKo2JycnLg7OwMANi7d68w8kIdf2fOnBHiTyaTobi4WIg/FxcXDBkyBB9//LEQV7WtN46/co0Vf7W1eQCwdOlSnDt3DmVlZUhLS0NkZKSw3a2srJCbmwsjIyPhgFEdf+o2T03d5ql3oEpLS1FQUIDY2FhMnDgRRkZGOHPmTIPavF69euHChQvYvn07rl27hiVLlkAkEkEul0MsFsPc3BwFBQXC+x89egQrKyv0798fO3bswO3btwFUHwMtKebUDh48iAsXLsDKygpLlixBWFgY7t+/j5ycHLi6uqJz587YvHkzioqKcO7cOfz1118AWk7MPY56XV+5cgVpaWlwdHQUXuvcuTP++ecflJWVwcPDAyqVSiiPug2sTCQSIS8vD3fv3hVi7tNPP9Vo86rro1u1aoWePXvi2WefRatWrYTEvrm5uXBAHBcXh/v37+O7775DcXExSktL8eWXXyI2Nhb5+fnCTqa1tTVyc3NhZmaGpKQkGBsb48iRI7h+/Tpu3bqFPXv2ICwsDB06dEDfvn3h4uLyRJO/JiUloaSkBA8fPsSqVatq3N69evXCihUrhLIvWbJE2AkHgHXr1uHBgwdQqVRYtmyZsJz+/ftjy5YtMDc3F95b8aCjffv2OH/+PHbv3o39+/ejsLBQI06MjIwwfvx4PPPMM3j22WcbdGcqNze3Km2XelSkTCaDk5MTfv75Z+Tn5wMob68UCgVMTEyQm5uL8+fPY9++fdUuOywsDPfu3cOdO3cAABcvXoSFhQVcXFxgZmaGsrIyoQ+zsLCAWCzGn3/+ibi4OGzdulVob9Rnw1NSUgCUj1xVn41Xj+L97LPP0KNHD2zfvh1///03rl+/jsWLFyM6OlrYZ1P/dvr27YukpCTcvHkTQHkbUlhYiJycHLRq1QpWVlbIz8/HTz/9hJKSEqGM+fn5uHfvnhAP6veqy6HeB3ySePj888+FPj8nJ6fGhLSVlZWQEFJrSDwUFRUJdVCP2FRr1aoVpkyZgt27dyM3N1c4AElKSkJOTo5QjoCAAJw8eRJpaWk4f/480tLShJGTYrEYVlZW1e5XGRsbC4n5ittHrUOHDhCLxTh58iS++uorFBQU4L///S9cXV1BRDh27BiA8pF96ngQiUTo0aMH/vnnH6EPzMzMREFBgUZfql6/6rqLRCJkZ2ejS5cuKC0tRWZmJoDykXJyuRwSiQSBgYEoKChoUNutlpSUJCQOV61aBV9fX7Rt2xZOTk4YOHCg0F4pFAqkpKTg7t27UKlUSE1NhYeHhxBjGRkZuHPnDszNzbFq1SpIpVJhdP5//vMfoQ1Xb6fqqO+y+PPPPwvP3b9/HxMmTIBUKoWbm5sQQ5mZmbh+/Xqd61mdnJwctG3bVlgPOTk5yMvLg6OjIxQKBVJTU5GdnS20KR4eHjh//jzi4uI0+uCa1BRD6oTbb7/9ptGmlJaWCm1KxRiysLBAjx498Nlnnwn7dhs2bMD169eF9jE3N1fjmHbDhg0a31uxf1O3S+p9FnW7FBERgTt37gjt499//w2RSIQJEyZg1KhRuHTpUpVR+0B5f3n8+HE8ePAAQHl7s3z5cmGkWcVEuIODA8RiMRYvXoxRo0bh9OnT2LhxI65cuYJDhw4hOjoaYrEYRkZGNe5bVPydGBkZ4dSpU0K/ePLkSWHkVF5eHvLy8oT2sWfPnlixYgWICCqVSmgHe/XqhUuXLmHNmjW4f/8+fvjhB5SUlAjtoEwmw8cff4yioiIsW7YMERERyMjIQP/+/fHvv/8KiaKcnByhPa64z2JkZIQXX3wRDx48wDPPPCPEcE3tY8X6VT6Wd3R0FPrgy5cvC+1CdeuocmwD/+svk5KSAKDW/rI2jRXbZmZmQmyry/3o0SNcv369znkMLy8vWFlZaRxrZmRkIDs7Gzk5ObC0tIS9vT2Ki4s1TvrUpqb9+tLSUly6dAnBwcGYPHkyBg4cKBzLVF4nOTk5cHd3h0gkwt69e3H8+PEa198zzzyDpKQkYeRxWVmZMFK8Lho1aWZra4sNGzbg3XffhVwuR1BQEPLy8jB06FD4+fkhJCQETz31lMZn3n77bcTExECpVAoNSGWtW7fGypUrMXLkSMjlcixatAhr16597Fm8Dh06YNWqVXj55ZehUCjg4+ODr7766rH18Pf3x4ABAzB9+nTI5XIcPHhQaIiio6MhEonw3HPPYfz48cjNzRWG7NbFypUr8euvv0KhUGDgwIFYsmSJsE7qsi7U3nnnHYSEhKBz586Qy+UICwur12WvNjY2ePjwIcLDw/Hpp59i+/bt2Lx5M77//ntYWFjgww8/xM8//4wTJ07gk08+qXJGLSwsDNbW1nW6DKA6tra22Lp1Kz799FMoFAr4+vri3XffrdPdgGpaT5XjD/jf2Wu5XA4bGxv4+flplPeZZ54R1odSqawSn0D5DqQ6/u7fv4+kpCR8/PHHiI2NxYsvvohz587B398fCQkJGp+rGH+LFy/Gw4cPkZiYCG9vb9y6dUu4ROSzzz7D/PnzkZmZiaVLlyIsLAxhYWGYPn06Nm/ejKtXrwrx99RTTwnxt2XLFhQXF+Opp57Cnj17sHDhQnzzzTe1rjuOv3KNFX8lJSU1tnkA8MMPP+D69esoLCzEN998Aw8PD7z99tsAgFGjRsHR0RHh4eH48ccfNdpHdZsHABEREVXavJKSEsTExEAul6N///6QSqVISEhoUJs3btw4dOvWDUZGRujYsSMOHToEqVSKESNGAChPpojFYsjlcuzatQsmJibIy8vDO++8A3d3dyQlJdUaAy0l5tTCw8Oxe/duvPrqq9i8eTOmTJkCOzs7TJs2DcuWLUN4eDg+/PBD9OzZE35+fkIH3FJi7nHrSaVSITAwELGxsfjmm280zhTa2Njgv//9L6RSKbp16wZLS0sYGxtDIpEIfXRlJiYmGDFiBLp27Yru3btDLpcLlxOr27zq+mgPDw/4+/vjpZdeQnZ2NmbMmAGg/DLa3NxcKJVKXLhwAevWrcP169dhZWUFc3Nz3LlzB2+++SYsLCyEHeLc3FxYW1tDJpNhzpw5+PvvvyGRSNChQwdMnToVt27dQrt27eDn54cZM2bg0qVLyM3Nxfjx4+u6eTT4+PggIiICAQEBiIyMFH7PlY0bNw5BQUEICgqCUqmEh4cH3nzzTeH1kJAQxMXFwcfHBzY2NsJr8+fPh7m5OSZOnAi5XI5hw4ZpjCgwNTWFsbExjI2N8dxzzyEsLAxyuRzBwcHCe8aMGYMbN25UGfVYV6ampkLb9dtvv+Grr77C/v37sXDhQpSUlMDW1hb/+c9/YG1tLXxmxYoVsLe3R8eOHREaGgqFQlFtgqZ169bo3bs3NmzYgM2bN+P06dPo3bu3cBkXAGEf6sGDBzAzM8PEiRMxePBg9OnTB0ZGRkhISIBCoYBCocBbb72FxMRE3L9/HwqFQuN7MjIyMHToUBgZGaFr167o2bMn0tLS8MMPPwAo32crKiqCXC7HqlWrMH36dMTFxUGhUCAqKgp3797FyJEjUVhYCC8vLyQkJCAyMhJSqVQ4GDU2NkZKSooQD506dRJeKywsFPYBnyQeTE1NoVQqIZfLER0djStXrlS73ezs7ITf2ZPEg0wmE+pQVFSkcaZdPULB2NgYL7zwAt566y0EBgbi2LFjGvtIq1atgouLC+Li4hAaGgojIyPcvHkTAQEBaNu2LT7//PNq9/u7dOmCDRs2CPVVbx/173X+/Pn48MMPYWtri88//xwSiQRJSUlYt24d1qxZg4ULF+L8+fNYtWqVRjysXLlSuIw2LCwMnp6e2L59u0ZyCChvB9V1V4+e8PLygomJCeLj4yGXy5GbmyuM/urXrx+8vb3Rv39/2Nvb16vtVvPx8cG8efOwefNmREZG4vLly/Dz88PRo0exatUqzJ07F0D5PoNMJsPMmTPxyiuvwNXVFd7e3kKMOTo6YufOnejfvz+eeuopREVF4ccff8Rff/0FW1tbPHjwQIihmjg4OGDgwIEYP348TE1NoVAosH79eowZMwbFxcVo06YN/P39ERAQgAEDBiArK6vO9azO119/jYEDB2rEkKWlJfLy8rBixQr8/PPP2LFjh9DHt2nTBsbGxhCLxRp9cE0q/sYrxtDEiRMRFBSEY8eOwd7eHkD5b/nXX3/FW2+9hYCAgGpjKCMjQ5jmZeTIkejZsyeCgoKwf/9+BAQEaBzTVjwpBZTv192+fRtKpRIODg744osvkJycjAEDBgjtUuvWrdG3b1+89957kMvl2LhxozAiKyoqCs7OzoiNja1yl+kOHTpg8uTJuHDhAhQKBSZNmgQbGxt06dIFiYmJUKlUmD17tvB+ExMTiMViJCQk4Msvv0RWVhaKi4uxb98+oX1s3bq1sO4WLlyo8X0VfyelpaVwcHAQ+kV3d3eEh4cL63Tv3r1C+/h///d/CAoKwtGjR/HWW28J7eC4ceNgZ2eHzMxMODs7Y//+/QgKChLawaeffhq7du2ClZUV9u7di82bN+PKlSuYP38+xGIx3nrrLSG279+/j5EjR+LIkSOYPHmyUObXX38dQPlVHo9rHyvWr/KxvFQqFfrg9u3bCyeHK6suttVWrFiBRYsWwd/fX4jt+p7gaorY9vf3x4YNG/DWW28hKyurznkMIyMjbNmyBb/99hvkcjn8/PwwZswYPHz4EPHx8fDy8hKmq1En5h+npv16lUqFF198Ef7+/lAqlfjrr78wadIkAMAbb7yBsWPHQqlUIj09HZ999hneffddKJVKLF26VONS6REjRmDNmjUIDAzEkiVLEBkZiTlz5uDZZ5+FQqGAn59f/e5OWufZzwzMsWPHaMSIEURUPlFdxQkzN2zYQB988AERlU++eeDAAa2U8UkcO3aMwsPDqV+/flXqp55EnIjo+eefpz/++KPK569fv05t2rR57MTb2gSAnn/+eTp58qTG87qw/Tj+yul6/OXk5NCoUaN0Lgb1Nf7Uk6rWVr+a4k+XYq4iXYw/Iu3FYFRUlHCH1ydReQLayhoSgxWtXbuWevbs+cTlrI/8/Hxh0uN//vmnyl1lq6Pr8RcVFUXDhg174vh7XDw8qfrGw5PGH1HD4qG5tYT4a0ibUt32UcdQU5W/OdsUdf3y8/Pp448/plWrVgkxdPr0aeF9tcWfrmgJMdgQFWOwbdu29NZbbwmvqcsfFRVFMTExdS5/Te1gY7WPNcVwdc83RhtYm5bSPupq/LVETT6nma5SKpVwdHREZGQkTp8+jeeeew4vv/wygPLJ9a5du4bIyEhhAkRdo1QqYWVlhbS0tCr1W7NmDUJDQxEeHi7M5VPRhx9+iM6dO+Ozzz6r08Tb2rRnzx6MHTsWP/30k05tP44//Yi/QYMGYefOnToXg4YQfzXVr7r406WYq6h37946GX8Ax2BtbWB8fDymTJmCefPmNWuZDxw4AKVSCYVCgb59+2L+/Plwc3Or8f36EH/Hjh3DtWvXWnT8NSQeniT+1OobD81NH+Kv4vZRT9bdFOVv7jZFXb+wsDB8+eWXmD17Njp16oT58+fjwIEDdYo/XaAvMVhQUIAuXbpUKf+xY8cglUpbRPlriuGanm+MNrA2LaF91OX4a4lERNVMYqIl48ePr3YekoMHD9Z5krbmVHEy2IquXr2qcWc1tQULFgiT3jaXmsoYExODL774olnL0hQaM2aedFnp6ekYPXp0ledHjRqFt956q15lqQuOP+170phpzPhtjvjjmNM+bfaTlb/74cOHuHbtGjw9PTUuL2yqNq++dCFeG0tz9z9Ay91nCw4OrjKpsp+fH+bPn8/x0EJ+m2rNEUM1xYN6CoT64Dal5cVQc2hIDLXU9rEmHNsc22oNbR/1XYtKmjHGGGOMMcYYY4wx1hLw5ZmMMcYYY4wxxhhjjFXCSTPGGGOMMcYYY4wxxirhpBljjDHGGGOMMcYYY5UYabsALVFZWRkyMzNhaWkJkUik7eKwJkZEyM/Ph7Ozs8Zk1trEMWg4OP6YNnH8MW3i+GPaxjHItInjj2lTS4y/loqTZtXIzMxsUbfNZs3j33//haurq7aLAYBj0BBx/DFt4vhj2sTxx7SNY5BpE8cf06aWFH8tFSfNqmFpaQmgPICsrKyE50tKSrBz507ExsZCKpVqq3hNxlDrl5eXBzc3N2G7twTVxaChbh99wfHX8ul7Haurn67EH2CY20efNGb8paamYtasWSgrK8Mbb7yB1q1bY8qUKRCLxVi0aBECAgJw69YtjBw5EgUFBZgwYQKGDx9ep2Vz/BlW/XSlDTTU7aMvdD3+AMPdRvpCV/YBWypOmlVDPRTVysqqSmNhZmYGKysrvf0xGXL9WtIQ5Opi0NC3j67j+Gv59L2OtdWvpccfYNjbRx80Vvw9fPgQc+fOxe+//w5jY2MAQFRUFLZu3Yr8/HyMHz8e27Ztw+eff44pU6age/fuiIyMxMCBA2FiYvLY5XP8GWb9WnobaOjbR9fpevwBvI10na7sA7ZUfPFqA/jP2KHtIjDGdIjHu1v1pt3wn7EDHu9u1XYxmAHTl98Sa5iDBw/C1NQUffv2xbPPPoubN29CIpHA1tYWTz31FLKzswEAhw8fRs+ePWFkZITg4GCcOnWqUb6f449pE/fBTNu4DWSGiEeaMcYYY4wxnXD79m1kZGTg0KFDSE5OxvTp0zVGQxgZGaG4uBglJSXCxMbW1tZCMq2yoqIiFBUVCY/z8vIAlJ+VLykpEZ5X/y0Tk8bz+kJdJ32sG1Bz/fS1vowxxhoPJ80YY4wxxphOsLGxQUREBIyNjREdHY3p06drzMdSWloKY2NjSKVSlJWVQSwWIzc3F3Z2dtUub/bs2fjoo4+qPL9z506YmZlVeX5WcBm2bdvWeBVqYXbt2qXtIjSpyvUrLCzUUkkYY4zpCk6aMdZATTkRMWOMMcaqCgkJwdy5c0FESE9Ph6+vLy5fvoz79+8jPz9fSI6FhIQgNTUV3bp1w19//YU5c+ZUu7ypU6di0qRJwmP1xMixsbFV5vPZtWsXPjgqxl8fxjdtJbVAXb+YmBi9nc+nuvqpRxYyxhhjNeGkGWMN0NQTETPGGGOsqlatWuHZZ59FVFQURCIRli5dihs3bqB3794QiUT47rvvAADvvPMORo4ciWnTpmH8+PEwNTWtdnkymQwymazK81KptNrkUVGZSC+TSmo11VtfVK6fPteVMcZY4+CkGWMNUHEiYjMzM3z33XfCRMS2trYaExHPnTsXYrFYmIg4ODi4yvLqMqeKoc43og9kEoJMTAB4PhXGGHtSr776Kl599VXhcfv27XHgwAGN9zg5Oen9pYaMMcYYa3qcNGOsARp7IuL6zKmi7wcB+li/OaH/+5vnU2GMMcZ0G0/RwRhjhkNnk2bcWTFtauyJiOsyp4qhzjeiD/xn7IBMTJgVXMbzqTDGGGM6jKfoYIwxw6KTSTPurJi2NfZExPWZU8XQ5hvRB0UqkfA3z6fCGGOM6S5tTtFR01QPuk6fp+gAaq6fvtaXMX2jk0kzbXRW6sdAeYelj40cd1h119gTETPGGGOMsZZPm1N0zAouAwBs27atsarToujjFB0V8RQdjOkmnUyaabOzAso7LH3trADusOqKJyJmjDHGGDMs2pyi44OjYhSViXBqRlzTVrKZ6fMUHUDN9eMpOhjTDTqZNNNGZwVodlh/fRjfdBXUEu6wGGOMMcYYq5k2p+goKhOhSCXSy/10QD+n6KiIp+hgTDfpZNJMm50VUN5h6XMjxx0WYy0f3wyFMcYYa348RQdjjBkWnUyacWfFGDNkfDMUxhhjTHt4ig7GGDMcOpk0A7izYowZLr5zV9MwxJuh6GtdGWOMMcYYaww6mzRjjDFDxXfualr6frKlYv2e5M5dq1evxhtvvIG7d+9i7dq1mD9/PkxNTbF8+XK4urri3LlzGDduHEpLSzFr1ixER0c3RvEZY4wxxhhrNpw0Y4wxHcN37moahngzlIbeCEWlUmHt2rVwc3NDaWkp5s2bhz/++ANHjhzBrFmz8P333+O9997Djz/+CEdHRyQkJHDSjDHGGGOM6RxOmjHGmI7hO3c1LUO6GUpD67l69WoMGjQIc+fOxcWLF+Hj4wNjY2NERERg8uTJAIDMzEx4enoCAOzs7HDv3j20atWqcSrBGGOMMcZYM+CkGWOM6Ri+GQrTJpVKhTVr1mDjxo2YO3cucnJyNC4PVqlUAICysjLhOfXlwdUlzeoyp576MVA+r54+zsXGc+oxxhirK54igbHmw0kzxhjTQXwzFKYtSUlJGDx4sDBfno2NjcZlnhKJBACE1wHUenlwfebUA8rn1eM59XRXY82pxxhjhoqnSGCseXHSjDHGGGN1dubMGRw7dgxJSUm4ePEiFixYgLNnz6K4uBhHjx6FXC4HUJ60vXTpEhwcHGocZQbUbU49QHNevb8+jG/aSmoBz6nHGGOsLniKBMaaFyfNGGOMMVZnn3/+ufB3cHAwFi1ahF9//RXdu3eHiYkJli9fDgD45JNPMHr0aKhUqmpHkqnVZ049oHxePX1MKqnxnHqMMcZqwlMkNA2eIoHVhpNmjDHGGGuQo0ePAgCGDBmCIUOGaLzm6+uLtLQ0bRSLGQCez4cxZoh4ioSmxVMksOpw0owxxhhjjOkMns+HMWaoeIqEpsFTJLDacNKMMcYYY4zpjMacz4cvTSpniJcmVfeYsZaOp0hoWjxFAqsOJ80YayC+NIQxxhhrXo09nw9fmqTJkC5NAvjyJKbbeIoExpoHJ80YawC+NIQxxhhrfo09nw9fmlTOEC9NAvjyJMYYY4/HSTPGGoBv9cwYY4w1v8aez4cvTdJkSJcmqR83FF9xwBhjhkGnk2bcWTFtaOxLQ4C6zaliqPON6AOZhCATEwCeT4Uxxp5EY8/nw1hD8BUHjDFmOHQ2acadFdOWxr40BKjfnCqGNt+IPpgT+r+/eT4VxhhrHDyfD9OW5r4ZRcUbUVR8rC/0+cQpwDeiYEzX6WzSjO+c1Pi4w6qbxr40BKjbnCqGOt+IPvCfsQMyMWFWcBnPp8IYY4zpMG3ejGJWcPky9fVmFPp44rQiPnHKmG7SyaQZ3zmpaXGHVbumuDSkPnOqGNp8I/qgSCUS/ub5VBhjjDHdpY2bUVS8EUVRmQinZsQ1er20SZ9PnAJ8IwrGdJ1OJs34zklNgzus+uNLQ5i28CXqjDHGWPPT5s0oispEKFLp780o9PHEaUWNeeKUMdZ8dDJpxndOalrcYTHW8vEdXBljjLHmxzejYIwxw6KTSTPurBhjhqyxL1HnSYjLGeK8jvpaV8YYaw58xQFjjOk/nUyaVcSdFWPM0DT2Jeo8CbEmQ5rXkSchZowxxhhjrGY6nzRjjDFD09iXqPMkxOUMcV5HnoSYMcYYY4yxmnHSjDHGdExjX6LOkxBrMqR5HfW5nqzl8Hh3K2QSwpxQbZeEMcYYY6x+OGnGGGM6jC9RZ4wxxhhjjLGmIX78WxhjjDHGGGOMMcYYMyycNGOMMcYYY4wxxhhjrBJOmjHGGGOMMcYYY4wxVgknzRhjjDFWZ4cPH0aXLl3QrVs3PP/88ygpKcHatWsRHh6O6OhoXL9+HQBw7tw5dOvWDeHh4di9e7eWS80YY4wxxlj98Y0AGGOMMVZnbm5u2LNnD0xNTTF16lRs2rQJ8+bNwx9//IEjR45g1qxZ+P777/Hee+/hxx9/hKOjIxISEhAdHa3tojPGGGOMGQyPd7cCAN/B+gnxSDPGGGOM1ZmTkxNMTU0BAMbGxjh//jx8fHxgbGyMiIgInDhxAgCQmZkJT09PWFlZwc7ODvfu3dNmsRljjDGdx6O9GWt+PNKMMcYYY/V29epV7Ny5E5999hnu3r0rPK9SqQAAZWVlwnPW1tbIzs5Gq1atqiynqKgIRUVFwuO8vDwAQElJCUpKSoTn1X/LxKTxvL5Q10kf6yaTEGRiAoBqtyljjLG64dHejDU/TpoxxhhjrF7y8vIwYsQI/PTTT1CpVEKiCwAkEgkAQCz+32D23Nxc2NnZVbus2bNn46OPPqry/M6dO2FmZlbl+VnBZdi2bduTVqHF2rVrl7aL0OgqXhJSsX6FhYUNWt7hw4cxceJESKVSuLi4YMWKFdi4cSPmz58PU1NTLF++HK6urjh37hzGjRuH0tJSzJo1iw8aGWM6z8nJSfi7utHekydPBvC/0d4AhNHefOKqZvp64komKT9hxSeungwnzRhrAN5hZ4wZqtLSUgwdOhTTp0+Hl5cXSkpKcPbsWRQXF+Po0aOQy+UAynfsL126BAcHhxpHmQHA1KlTMWnSJOFxXl4e3NzcEBsbCysrK+H5kpIS7Nq1Cx8cFeOvD+ObtpJaoK5fTEwMpFKptovTqPxn7IBMTJgVXKZRv4rJ1vrgkRaMMUPXWKO9+cSVJn07cVV5HrPGOHFliDhpxlgD8A47Y8xQrV69Gn/++SdmzZqFWbNmYcKECXjzzTfRvXt3mJiYYPny5QCATz75BKNHj4ZKpap2h1xNJpNBJpNVeV4qlVabPCoqE+ldUqmimuqty4pUIuHvivVraD0bc6QFj7Iop6+jLNRqql9D6ssnTpm2NeZobz5xVU5fT1z5z9gBAI164soQ6WTSjDsrpm2NPTSaMcZ0xYgRIzBixIgqzw8ZMkTjsa+vL9LS0pqrWMwANcZICx5loUnfRllUVrl+DRlpwSdOmTY19mhvPnGlSd9OXFU8aQU0zokrQ6STSTPurFhL0ZwTYRvqWWB9UNMk2NU9rgs+ccAYM3SNNdKCR1mU09dRFkDNlwcDDRtpwSdOmTY19mhvxtjj6WTSjCdAbBr6nLQAGndoPqC9ibAN7SywPqhpEmyAz3Izxlh9NeZICx5loUnfRlkANV8erH7cUNo4cVrTCThdx8cgdcejvRlrfjqZNFPjCRCbhj4mLSpqjKSFNibC1uezwIB+14/PcjPGWOPhkRZM27R14nRWcPmxjb4eh/AxCGOsJdLZpBlPgNj49DlpAdRcv4YkLbQ5EbY+ngWuSB/rx2e5dYMhnunW17oy/cYjLZg2afPE6QdHxSgqE+HUjLimr2gz4mMQxlhLppNJM54AsWnpY9KiosZIWvAOO9M2PsvddAzpTDef5WaMsfrR5onTojIRilT6exzCxyCMsZZIJ5NmPCyfMWbI+Cx30zDEM918lpsxxuqHT5wyxphh0cmkGXdWjDFDxme5m5YhnenW53oyxhhjjDH2pHQyacaYIfOfsQPnP+mj7WIwLeITB4wxxhhjjDHW9MSPfwtjjDHGGGOMMcYYY4aFk2aMMcYYY4wxxhhjjFXCSTPGGGOMMcYYY4wxxirhOc0YY4wxxhhjjLFG5vHuVsgkhDmh2i4JY6yhOGnGDAJ3WIwxxhhjjDHGGKsPvjyTMcYYY4wxxhhjjLFKOGnGGGOMMcYYY4wxxlglnDRjjDHGGGOMMcYYY6wSTpoxxhhjjDHGGGOMMVYJJ80YY4wxxhhjjDHGGKuEk2aMMcYYY4wxxhhjjFViEEmzd955B5GRkRgxYgRKSkq0XRxmYDj+mDZx/DGPd7fC492t8J+xQyvfzzHItInjj2kTxx/TNo5Bxp6c3ifNjh8/jhs3biAtLQ3e3t5Yt26dtovEDAjHH9Mmjj+mbRyDTJs4/pg2cfwxbeMYZKxx6H3S7MCBA4iNjQUAxMfHY//+/VouETMkHH9Mmzj+mLZxDDJt4vhj2sTxx7SNY5CxxmGk7QI0tZycHDg5OQEArK2tkZ2dXeU9RUVFKCoqEh7n5uYCALKzszWGsZaUlKCwsBBGJWJkZWU1ccmbn7p+WVlZkEql2i5OozIqLYBRGaGwsKxK/fLz8wEARNTo31uX+APqFoMcf7pL3+JPVSbiGNQhRqUF5f9XE4NNGX8A98H1oa/xB9TcBnL8tRyGGH8A98EtBccft4Hapq8xqM19QH2i90kzGxsb5OXlAShvBOzs7Kq8Z/bs2fjoo4+qPN+2bdsal9vqi8YrI2s+L9TyWn5+PqytrRv1++oSf0D9Y5DjTzfpS/wBQKu5jVNG1rxqisGmiD+A+2CmieOPaRP3wUybmjv+AG4Dmabm7oP1iYj0PLWYnp6OefPmYcWKFfj000/Rtm1bPP/88xrvqZxhLysrQ3Z2Nuzt7SESiYTn8/Ly4Obmhn///RdWVlbNVofmYqj1IyLk5+fD2dkZYnHjXrFcl/gD6haDhrp99AXHX8un73Wsrn5NGX8A98H1YYj14/hrOQy1ftwHtwyGWj9uA1sOQ6xfU8efPtH7kWZKpRKOjo6IjIzEU089hcmTJ1d5j0wmg0wm03jOxsamxmVaWVnp5Y9JzRDr11TZ9brEH1C/GDTE7aNPOP5aPn2vY+X6NeXZRe6D68/Q6sfx17IYYv24D245DLF+3Aa2LIZWPx5hVjd6nzQDgC++4DGkTHs4/pg2cfwxbeMYZNrE8ce0ieOPaRvHIGNPjsfhMcYYY4wxxhhjjDFWCSfN6kEmk2H69OlVhrDqC65fy6br5X8crl/Lpuvlrwt9r6Ou10/Xy/84XL+WTdfL/zhcv5ZN18v/OFy/lk8f6lAbrh+rjd7fCIAxxhhjq70BGgAA1V1JREFUjDHGGGOMsfpqcSPNZsyYgUePHjX796ampkKpVNb7c+np6fjll18av0Coui5Gjx6Nr776qkm+q7F99dVXuHXrlvB48eLFOn1NfX3i0sPDA+np6QCAl156CSkpKQ3+3triMjU1Fdu3b6/2NY7L6ulbXD5OdXHblLGhxu1p41CXubnjtiVuv5aqe/fu2LhxY70/d+XKFSxevLjO72/KfaPvv/8e3t7eUCqVyMrKapLvAP5XhxkzZuDNN9984uW1atUKV65cAQD07t0b58+fr/Z9leP5yy+/xOeff17lfVu2bEH37t2Fx08Sz3369MGff/7ZoM9qkz7Ec2N8108//YT+/fs/0fempqbCzMwMqampAOq+P5ieno633noLEyZMeKLv1yX6GHePU11ZGiPuHqch/btcLseuXbvq1B5++OGHWLlyZbWvffvttxg9enS9vlubDDEuq9OYcZmZmYnIyMgnXk5Dt82TaHFJs48++qhFB05lTXmQ0BTrorS0tFGXV1lZWRnKysqqHOSNHz8e//nPf5r0u5tSQ7fFkiVL0KNHjyYokfaSZhyX2lfX9VXdtnqS2Gjq7cRxq+mjjz5CYWGhzsRtS0iaNfU2aSz13aluyn2jr776CsuWLUN6ejrs7e2b5DuApq3Dtm3b4OXlVaf3bty4sU4HAA2N5wcPHuDs2bMIDQ2t92crqy6eW2KMt6R41uZ31aau+4Pp6elYt25dkydPatMSY6w6LTnuHqcllaU66hjIyMiAhYUFbty4Uaf2cObMmRg2bFiTlqml0+W4bGrOzs5IS0vTdjEahprQgQMHKCIiguRyOQUEBNDGjRuJiOjtt9+m4OBgUigUFBkZSefOnSMiopdffpkAkL+/PykUCrp9+3aNy96+fTsFBgZSQEAAdevWjU6fPk1ERCkpKeTn50cTJkwguVxOvr6+dOTIEY3PRUREUFBQEIWEhNCePXvq9LkVK1ZQQEAABQQEUO/even69et0+/ZtcnNzIysrK1IoFPTyyy/Xex1dvHiRoqOjKSAggBQKBW3YsKHGdTFq1CgaN24c9ezZkzw9PenZZ5+loqIiIiIqLi6md955h0JCQkihUNCgQYMoOzubiIhGjRpFiYmJFBkZSV5eXjWWZdmyZdSjRw/q27cv+fj4UGRkJF2+fJmIiE6cOEEREREUGBhIPj4+NGvWLOFz06dPpwEDBlBsbCz5+PjQjBkzSCqVUseOHUmhUNCxY8do+vTpNHHiROEzX3zxBYWEhFBgYCDFxcXRlStXiIho8+bNwrrw8/MTYqYmW7ZsoeDgYJLL5aRQKOjQoUNEVHN8qKlj08nJiUxNTcnDw4McHR3J29tbiEtnZ2eaPHkytWrVigBQmzZthG3h7u5OkydPpqCgIGrfvj3NmTNH+F5jY2Py9PSkbt26UXBwMG3YsIFSUlLIx8eHvL29SSaTkUwmo759+xIRUXJyMnl7e5O5uTmZmJiQh4eHEJdxcXFka2tLLi4uZGxsTJaWlnTgwAE6duwYOTo6koWFBZmYmJCjoyPHZQuKy8WLF5OnpycFBgbSzJkzqWJTC4BycnKEx/b29kJ9amob1Z9LTEwkCwsLat26NXl6epK3tzc5ODiQiYkJtWnTRvjM9OnTqV27dgSAZDIZWVhY0Llz52qMjY8//phMTU3JxMSErK2tadOmTURENH/+fJLJZOTt7U0mJibk6urK7WkD41ahUAhtlXrbBQYGkp2dHbm4uAjb/NVXX6UBAwaQq6srASAHBwcCQO3atasxbl966SUyMzMjExMTsrW1pV27dhER0SeffEIymYzs7e1JJpORm5tbjdtPLpeTubk5ffjhh+Tp6UlSqZR69+4tbL9vvvlGaFO9vLyoY8eOFBAQQJ07d6aOHTvSqFGjSCqVklgsJk9PT2H71RQj1YmKiqLXXnuNgoODqX379jRp0iQqKysjIqKbN2/SoEGDKCQkhPz9/en9998XPufu7k5TpkyhkJAQeuGFF+jChQsUHh5Ocrlc4735+fmUmJhIfn5+5OfnRzNmzND47rfffpu6du1K7dq104i/lStXUmhoKCmVSpLL5bR582aNz6njrDqFhYU0ePBg8vHxIblcTjExMURE5OXlRSYmJqRQKIS+oD77RqNGjaL58+cL3/P222/T9OnTiah+7dVzzz0ntIvPPfdcrdtM/bseP348tWvXjszMzKhjx45ka2tLTz31FIWGhtLLL79MwcHB5OXlRVZWVuTj40M+Pj4UEhIi1MHR0ZHGjh0rlKFyu7t8+fJq+++PP/6YjI2Nyc7Ojlq3bk0SiURoq9zd3WnhwoUUEBBAJiYmGuVWKBT01FNP0YQJE8jX15dkMhkdOXKEiouLacKECeTg4ECmpqbUqlUrsrOzE9ojOzs7kkgkZGpqSm3atKH09HQiqr1PIiJas2YNTZgwgaKiosjb25tatWpFpqamZGVlRaNHj6Y///yTunTpQubm5uTo6CjE6Ny5cyk4OJikUim1adOGfHx8hHiWyWTUunVrMjU1JT8/P4ON57Zt2wrfXd1xhY+PD0kkEpowYQIFBgaSlZUVASB7e3syMTEhLy8vOnnyJBGVtyndu3enoKAg8vX1pb59+2q0cT169BDi3sXFRehrzczMaOXKlUJZK/5efHx8qH379uTn50dxcXFkYmJCzzzzDMnlcjIzM6MvvviCrl27Rq1bt6alS5cK/aCLiwvNmDFD6AdFIhHJ5XIaOnRotW1ZbX2Uo6Mj+fj4kLm5OUmlUvL19aWCggKhztyONn47un79epJIJBr7Y/v27aOgoCBh3yooKIjOnTtHN2/eJCcnJ2H/zN7enm7evElE5W1Lv379hOWuWLGCOnbsSKampmRmZkadOnWq9nhXKpVSq1atKCAggFxdXWnRokU0adIksrS0JGNjY2rfvr3QHkZHR5OJiQnZ2dmRpaUldezYkY4cOUKXL18ma2trevrpp8nExIScnJyod+/e9P7779MHH3xATk5OJBaLSaFQ0Lhx48jBwYGmTp1KRER//PGH8HupuE7z8vJo8ODB1LFjR7KysiI/Pz+yt7fn/r2Z4lK9TadNm0aBgYHUoUMH2rdvH7355pvC52tqD1999VVSqVREVH1choaGUmBgIEVGRgp9Y0UqlYpeffVV8vb2JrlcTkFBQfTw4UOhTGoA6JNPPqGQkBDy8PCgpUuXCq/t37+fFAoF+fv7U2JiIsnlckpJSamybfLy8uill16ikJAQCggIoLFjxwr7842pyZJmWVlZ5ODgQHv37iWi8pWXlZVFRER37twR3rd69WqKi4v7X4EqHVBWR70zc+LECSIiSkpKIh8fHyorK6OUlBSSSCRC8mTRokUUGxtLRESXLl2isLAwys3NJaLyA6w2bdrQo0ePqv2ch4cHde3alfr06UOOjo50/fp1IiKaOXMmubi4UNeuXSkmJkYjkOorNDSUFi9eTEREFy5cIDs7O+FAvfK6GDVqFIWGhlJBQQGVlpZSeHg4rVq1iojKD45mzpwpvHfmzJn0yiuvCJ+Ty+WUl5en8d1Tpkyhrl270vDhw6m4uJiWLVtGxsbG9NNPP5Grqyu1bduWbG1tiag8IB89ekRE5Y2EUqmkgwcPElF5csLJyYlu3bolLNvd3Z2OHTsmPK54kLdy5Up66aWXqLS0lIjKf3y9e/cmIiK5XE4HDhwgovKYqS0Wzp8/T61bt6azZ88SUflOxP379+n27dtka2tLfn5+ZG5uTrNnzxbig6g8/kxMTIQfljo2t23bJhwEr169mkxMTGjEiBFUVlZGAMjFxYX2798v1E/92t27d8nNzY22bNlCdnZ25OTkRMeOHaOkpCQyMzOj3377jVJSUkgkEtGAAQNIpVLRokWLqHv37kRE9Pfff1Pnzp0pNzeXsrKyyMnJiRwcHGjSpElkY2NDACg5OZlKS0upXbt2FBAQQImJieTo6EhGRkZCXH788ccUHx9PRFUbuPpqjrj09/cnMzMzGjJkCBUXFwvvSUlJIVdXV4qKiqKePXsKcXnmzBkiIvr888+FjqklxuXJkyepTZs2dP78eQoJCSGpVFolaXbv3j1KTEykrl27komJiXDA9bi20dzcXGhT79+/T5mZmXTnzh2h7jNnzqS4uDiaPn06ubu7C9tqyJAh9OmnnxJR1dg4deoUiUQiWrduHRERTZgwgYyNjWny5Mnk4eFBAGjBggVEVN4uxsTEUGJiIgUHB1ObNm3q3J6q2+GTJ09qtKfailt/f3+ysLCgoUOH0sOHD4W4TUlJISsrK/Lw8KCePXsSUd3a08oqxm1WVhZZWFhQcHAwEf1v2xERXb16VYjb1atXU/v27YW4VZe5trj97rvvSCaTCa+PHz+eLCwsKCcnh4yNjQkArVy5klQqFc2dO5diY2OptLSUBg4cSJaWljR+/HgiKv/dAaDVq1dTSkoKicVicnNzIyKiOXPmkJGREZ04cYJOnjxJ1tbW1KFDByorK6MxY8YQADp06BAtW7aMFApFnfrc6qh/88XFxVRQUECOjo7UsWNHGj58OMXExFBqaioRET169IhcXFzI29ubJk6cSO7u7jRmzBihjX/jjTeEeCciYd9jypQp9MILL5BKpaIHDx6QUqmkX375Rfju/v37U0lJCRUWFpKHh4fwm793756w7MuXL5Ojo6NQh8ftVP/222/C+qhYlpSUFHJwcNDog9W//5SUFLKzsyNbW1shBqtrd2vaqa5Pe0Wk2S7WZT9p9+7d5ODgQCNHjiR7e3s6ffo0ZWVl0SuvvCL8Tv79919q27YtSSQSOnLkCCmVSqEOH3zwAfn6+lLXrl0pLi5Oo9399ttvSSqVVtm/u3XrFllaWgrtyvfff08AKDIykoiInJ2dKSAggHJzc8nT05N+++03atOmDZ05c4ZsbW01PhcXF0exsbH07bffUmhoKDk6OtI///xD3bt3Jw8PD4qPj6d9+/ZR69atyc7OjoYPH0579uwhX19fIiL68ccfSSwWU1BQEE2cOFGjTyIiev7552nHjh0UFRVFbdq0odDQUMrKyqKgoCBycHCg/v37U0xMDP3+++/k4OBA6enpFBcXR0uWLBG2R9++faljx45EVB7PNjY2Qow3VjxPmTKFgoODycTEhPLz84XPzZo1S6MPrms8KxQKjffWZ1+/LvGsPq7YunUrEVV/XHH58mUCQAEBARQSEiK0geoD4Tlz5tBzzz1HiYmJFB4eThMmTCAioszMTJJKpfT5558TEdG4cePIwsKCysrK6JtvviGRSCT0vW+//TaZm5sTkebvpaioiNq0aUN2dnb06NEjmjNnDgGghQsXEhGRp6cnKZVKIiLq0qULWVtb0/Xr1yk/P59MTEwoOjpaWL6LiwtNmTKFnJycSKlUCvtIWVlZVFpaSkFBQfTUU08J/UDFPsrR0ZHMzMwoKyuL8vLyyNramgYPHkxERLGxsUI7WlJSQnFxcbRmzRoiomZvR83NzSk8PJyGDx9O3bp1E9rRyvuARNqNO6La29GsrCyyt7cnALRhwwZSqVT01Vdfkbm5ObVv357Mzc1p0qRJ1LlzZ4qLi6OHDx/S/fv3CQB17tyZPDw8aPXq1USkuQ+0b98+6tmzp3C8u3fvXnJxcan2eNfd3Z169uxJsbGxdPHiRTI2NiY3NzfKzc2lw4cPk62trdCOb9iwQfjc7NmzKTIykmJjY4XfjjouiYgiIyNJIpHQM888Q0uWLCFTU1O6evUq/fXXX9S6dWuysLCgiRMn0rRp0+iTTz6psk4nT54sHCtFRESQmZkZjRgxggoKCqhTp05C8llbcTllyhTq3LkzDRs2jIqLi4X+fceOHeTq6krW1tYUEBBQ7XYnatlxef/+fZLL5QSAvvrqKyIiWrJkCZmbm9OuXbsoMTGRPDw8qEOHDkRE9PDhQ6EfKC0tpaeffrrGuExISBD2gfbu3Sv0jRX9/fff5O3tLSTe7t+/TyqVqtqk2ZdffklERGfPniULCwsqKSmhoqIicnV1FZK9e/bsIQDVJs3Gjh1Ly5cvJyIS9kvVg1kaU5Ndnnnw4EF4eXkJ162KxWLY2dkBAHbt2oUuXbrA398fM2fOFOZ/qqs///wTAQEBCAgIAAAMGzYMmZmZuHHjBgCgQ4cO6Ny5MwCgS5cuuHTpEgBg+/btyMjIQLdu3aBUKjFw4ECIxWJcu3atyufs7Oxw7949pKWlgYjg5eUFFxcXAEC7du1w69YtpKamoqioCNnZ2Q1aR/n5+fj7778xZswYAICnpye6du1a67DFZ599FmZmZpBIJAgNDRXqtnHjRiQlJUGpVEKpVGL16tW4fPmy8LlBgwbB0tJSeHz8+HHcuHEDaWlp8Pb2xrp16wAA4eHhcHd3x5AhQ/D333/jwYMHUKlUePjwIV566SUEBAQgLCwMV69e1dhuvXv3hqOjY53qvXHjRiQnJ6NTp05QKpWYM2eOsA2io6MxceJEzJkzBydOnICNjU2Ny9m1axfi4+Ph7e0NAJBKpbC2thbiIyUlBQMHDkSfPn004uObb75B69atceLECfz777/o1KkTunXrhvHjx+P8+fPw8/PDzJkzUVxcjDFjxkAkEgEA+vbti+TkZOH71a+1atUKAwYMwKpVqxAQEABjY2MA5XFZVFQkzBEjFovxxRdfQCwWo0uXLvj3338BAJs3b8axY8fg6OgIFxcX3L59G0VFRbh48SI6d+4MMzMz3LlzBxKJBF26dMHly5fh7OyMwMBAWFhYCOvulVdewZ49e6BSqeq0HWrSHHF5/PhxFBYWYurUqQgICBDiT23IkCFITU3F7t27AZTHpY+PDwBg3LhxSE1NbbFxuWfPHsTHx6Nt27bYunUr+vXrV+U927dvF4YoExH+/vtvAI9vG/38/IQ2taioCFOmTEFQUBBatWqFEydO4LvvvhM+Ex8fL3yuYjtY2S+//AIrKys899xzAIDvvvsOxcXFOHXqFCZOnAixWCxcstWlSxecPHkSzs7OSExMxP379xEcHPzY9rTi96ekpCA+Pl5oT7URt+r4e//99+Hv748NGzZoxK2ZmRmMjY2RlZVVp/a0Juq4PXjwIAICAnD8+HGoVCph2wUEBKBr1644efIkBgwYgJkzZ+LWrVv1ilv13CGjR4+GUqnEvn37UFhYiJycHMTFxUEsFuPo0aM4ceIEoqOjcenSJWzZsgX5+fmQyWRYs2YNOnbsiFdeeQUAEBQUBABo3749MjMzAZTfccnIyEhoV/v374+7d+/ixo0b6NevH0QiEYKDgwEAtra2de5zqzNy5EhIpVJcvHgRrq6uiIiIQLt27bB7925MnDgRSqUS3t7eKCwsxLBhw1BQUICioiKMHj1aaKu7deuGH374Ae+//z527twp/F6Tk5MxduxYiMVimJubY+TIkdi1a5fw3UOGDIGRkRFMTU2hVCqFely+fBkJCQnw9/dH//79kZ2drREPtVEoFDh79ixeeeUV/Prrr5BKpQDKL3kpKSnR6IPVv//ExEQAgLGxsdAG1kd92qvK6rKf9PDhQ3h5eWHAgAHo0KEDfH19YWdnh9DQUPz555/o0qULevXqhXv37qGsrAzDhg3D1atXhe+4cOECzM3NkZaWhgsXLmDr1q0a7a5UKq2yf7d161a0a9dOaFfGjBkDIyMjoVyPHj3Cv//+i27duuHBgwfCdl64cCFiYmKEz23cuBEjRozApUuXsHv3bnh6egrt9YsvvghnZ2fs2bMH33//Pe7fv4/S0lIkJycjMTER2dnZePjwIdLT0+Hm5oa//voLBQUFUCqVQp9UUlKCAwcOCJfhtW/fHs8++yzs7OwwYsQImJiYoHv37tizZw/effddFBQUoH///sjIyEBaWhqioqKQmZmJM2fO4MKFC3j48KFQp+LiYuzatatR4rlr1674/vvvUVJSgpKSkiqXElXug9VqiufqPOm+vpo6nt9++224uLigd+/eAKo/roiLi4NIJMKdO3ewdetWxMXFAYDQb4aGhuKvv/6Cs7Mzdu3ahdTUVHh6eqJLly4AIOzPd+nSBY8ePcKNGzewf/9+SCQSvP7661Aqlfjvf/+LR48e4eHDhxq/F7lcjqysLJiYmODatWsICQmBVCqFr68vAMDKykq4zL5jx46wsLCAi4sL1q5di5iYGKSlpUGlUuHvv/+GjY0Nbty4gQULFuDq1asYNGiQ0JZt2bIFN2/ehImJCX7++Wd07NixSh8VGxsLOzs7WFpa4rnnnsPevXtRUFCg0Y4GBwcjIyNDYy7A5mpH4+Li8PDhQ/z444/w9vbG3bt3NbZ55fjTZtw9rh09ePAg2rVrBxMTE/Tv3x9isRjdu3eHhYUFJk+eDDMzM/z6669IT09Heno6ysrK8PzzzwMo32+5ffs2tm7dWmW5mzZtQnp6OoqKijBixAi8/vrrUKlUNR7vjh49GpcuXUKHDh0gkUjw4MEDdOvWDWPHjkVOTg5EIhGuXbuG5ORkGBkZYcyYMViyZAlycnKEbSQWi9GvXz+4uLjg+PHjsLa2RllZGYKDg3HkyBG0bt0aycnJWLBgAYKCgtC+fXtkZ2djw4YNiI6OrlKH3bt3C8dKRkZGwj6JmZkZhg8fjuTkZK3FZUpKCm7cuIFvv/0Whw8fRrt27YT+/datWxgyZAiUSiVmzpxZY4y05Lg0MzPD0qVLIZFIhG0THBwMCwsLFBQUwNnZGT/99BOys7Nx8OBBlJWV4Z133oFCoUBgYCCOHj1abRk3bdqE48ePo3PnzlAqlXj99deFvrGidu3aobS0FC+++CKWL1+OkpISiMXVp53Ul/N6e3vDyMgIt27dwrlz52BkZCT0pT169ED79u2r/fzGjRvxxRdfQKlUIjAwEGlpacjIyKjL6qyXZp/T7Nq1a3jttdeQlJSEU6dO4Zdffmn063hNTEyEvyUSiXANNBEhJiZGaLjS09Nx48YNeHp6VvncyZMnhcc+Pj64c+eO8Nrhw4eFDR8QENDgpFl11I1CTWqr24IFC4R6nTlzBtu2bRPea2FhobGcAwcOIDY2FkD5wfX+/fs1Xl+/fj0SEhJQVlYGAHjvvffQqlUrHDt2DMePH0f37t01tlvl5deGiDB16lShrCdPnsTJkycBAPPmzcOyZctgZmaGUaNGYc6cOXVebkUikQitW7eu9rWzZ8/C1tYWxcXFSE1NRbdu3bBt2zbk5eUBAPbv349ffvkFVM2NZR+3feqq4rZbtWoV/Pz8UFhYiIcPH0KhUGDIkCEYOHAgAMDc3FzYPkZGRigpKRG2nYWFhfBaY5WtOo0dlwcOHICzszMsLCxqjL/IyEh8/fXXtX5vS45LqVRabQxKJBL8+eefwjYUiUQ4evRondpGiUSiUXdjY2MUFBTgxIkT6NevH0aMGCF8pqZtUlfqHTGpVCpsH4lEgsLCQsTGxoKIEBwcjHHjxj22Pa3t+7URt+r4MzExEeKvYhmzs7MhlUqRmJhYp/a0vtRxu2nTJjx48AA9e/bElClT8Msvv6C0tLTecevu7q4Rt5aWlpBKpZg0aRIsLCxgYmKCUaNGYfny5SgtLcWBAwfg4+ODmJgY/N///R/GjRuHbdu2wdraGh07dgRQvv3UiUyxWFxtewhUXccikajOfW5tDhw4AC8vL4hEIvTq1QtEhEOHDiE9PR2DBw/Gxo0bMW3aNMTHx6OoqEhjnT333HPYv38/vLy88O2336JPnz51KntNMTt06FC89NJLOHXqFNLT02FhYVHn/ZZ27drhzJkzQpz5+/sjJycHp0+fFhKv8fHx2L59u/D7X7ZsGUxMTHDv3r0a20AjIyONRHPF8jxJP1qf/SSJRKLxOCcnBydOnEBSUhK6du2KZ555BkZGRli/fr3GBPv//vsv3N3dAQCOjo4IDw8Xvuvbb79FSEhIjeVTf59621VsV8LCwpCeno6jR49CJBIhIyMDW7duRXx8PExMTJCfn4/z589DLpdX2x6JRCJhuZmZmfD19UWPHj2wZcsW9OnTBzdv3oSpqSkyMjJga2sLoHzbVZzwf8+ePYiIiNA4eKq4jsRiMWQyGQDg0KFDCAsLwxdffIEzZ85g3bp1+PLLL4WDGaD85Mhzzz0HR0dHeHp6Nlo8+/r64ptvvkF6ejrMzc1x5MgRjc/W1AfXFM+V1Xdfvy7xLJPJcOHChSrxXPG7duzYAXNzczx69AitW7eusk7UyYTY2FjMmzcP1tbWePHFF7FgwQI4OjpWW0Yigp2dnRCj69evh6urK0xNTTV+L2vWrIGLi4vG76Xi94tEIqGOcrkc9+7dw82bN/HTTz8JB41EhFOnTkEkEiE2NhbPPfccVq9ejQcPHgjb/sCBA7CyssKCBQuE9rtyH1WZSCQS2nB1O5qeno6MjAxMmzZNeF9ztaNvvvkmTE1N8ejRI8THxyM3N1djGZXjT5txV9d2VP27VtfbyMgI7733Hrp27YpPP/0Urq6uePToEebNm4d//vkHQPnxRnR0tHBDk4qICD179kRwcLCwvW7evKnxnorr2tTUVFjXYrEYkZGRwuckEglOnDiB27dv47fffkOHDh1w6tQpzJs3D8XFxcLnjI2Nhe154MABWFpaQiwWIy4uDhcvXhSSZn/88QdGjBiBHj16wNLSEpcvXxZOnNWmuv0FbcXluXPnEBsbi6FDh2LMmDHo16+f0L8XFxdj/fr1OHbsGP773//WWJ+WHJdSqRT29vYaiSp1n63OAUgkEuEYc968ebhz5w7+/PNPnDhxAi+88EKN7eGoUaM09hHUfWNF1tbWOHXqFF544QWcO3cOcrm8xkTWkx4vEBHWr18vlOf8+fP4/vvvq33vk2iypFl4eDguXrwonOUvKytDdnY2cnNzIZVK4eTkBCLCt99+q/E5S0vLKo1nZWFhYTh58iROnToFoHykhIuLizByoSZxcXFITk7GiRMnhOcOHz5c7Xvz8vKEjdOjRw9cuXJFOOt+8OBBhIaGQiKRwN7eHvn5+bV+b00sLS0RFBSEZcuWASg/87xv3z5069ZNeP1x60Ktf//+mD9/PgoLCwEAhYWFOH36dI3vz8nJgZWVFYDywFYn/g4ePAhLS0ucP38e/fr1g42NDdLT05GTkwNXV1cYGRnh/PnzGtn76lhZWdVY9v79+2Px4sXCd5aUlODYsWMAgHPnzsHPzw+vvfYaJkyYgEOHDtX4HXFxcdixYwfOnTsnLCc3N7dKfPz+++8a8WFubo7MzEzs2bMHpaWlICJcvnwZJSUlAKARl+odVwsLC2zevFnjTIr6NfVZlhdeeAEnT55EcXExgPK4lMlkwigda2trfPnll0IiUv2/kZERLl26hJMnT2Lv3r04fvw4rl69KmwfkUikkZglIlhZWcHf3x937twRzuAvXrwY0dHRkEgkta7/x2mOuMzJyREOKirGH1B+JuT8+fPYvXs3tm/fjitXruDgwYPCdlZPpiuRSFpkXPbo0QM7duwQEu2VO4kOHTrgwoULsLKywm+//YZHjx7h/v37j20bAeCff/4R2tTs7GzIZDJIpVLk5+dj586dVUZVVbetKq+DF154AXl5ediwYQMA4I033oBEIoFc/v/s3XlcVNX/P/AXM8AAsosgAu6GKJsKpiiLoiju5Zq5l1uW9jG/blmaZKalVlrqJ/dQU7TMct8w1xRz3wKX3HcE3ICZOb8/+M39MDAg4MAsvJ6Phw+Z7c45c95zzplzzz0nUHpO7vJRq9VwdHREmzZtcObMGalcgILr07yfz9atW6X61BBx+7L4+/jjj+Hl5YVNmzbh2LFjL61PC6KJ27CwMJw5cwbBwcGQy+V49OgRXF1d8fTpU8hkMhw8eLDQ9rCwz6VPnz64dOmSNLAZHx+PihUrwsvLC9euXYNcLkfv3r0xfPhwqe1LTU1FixYtsHPnTjx8+FDKf0EdlcDAQGRnZ+PMmTNo0aIFNmzYAHd3d3h5eWHjxo2wt7eXyu/JkyfS64rT5mrEx8cjOzsb9+7dw5EjR9CqVSt4enrC3d0dX375pZT+Fy9e4MaNG9KZ8NySk5Ph4eGBfv36YebMmdL3tVWrVli8eDGEEHj69Cl++uknafC6MKmpqahRo4aUPl2d4oLcuHEDFhYW6NSpE77++msIIXD9+nVkZWVJHVInJyfcv39f+v43atQIbdu2hb29PbZu3Ypjx47l+y7Xrl1b+iwfPnyo9YO5OPVVXkUpM03/TvOd0PTvnj9/DgsLC3h6euLRo0dS3XflyhXs2LEDCoUCaWlpePHihTQjOyoqCvv375disFGjRjhx4kS+/l2HDh1w+fJlZGZmAgCWLFmiFa82NjZSZ79KlSoIDQ3F22+/DXd3d6nsNm/ejNjYWKlv16pVK1y6dElqZ5YuXYpbt24hOjoaNWvWxOXLl3H37l04OTnh4cOHSEpKApDTjp07dw4XLlyAk5MTEhMTpTZpw4YNeOONN6R0Xb58WZoVvWrVKri7u8PGxgYtWrSQ4lnzGWVmZqJq1apSvjWSk5Mhl8vRsWNHvcWztbU1HB0dER8fj/T0dOmkIZAzOy53G3zs2DHpsYLiOW8dVdy+flHiefr06ZDL5fjjjz8AFPy7QhMjGhYWFlrvpVKp4OjoiNTUVHh4eCAjIwM1atTArVu38ODBAwA5V7TY2trCy8sLzZo1Q2pqqjSrUa1WS/283N+XunXrQqlU4ocffgAAHDt2THpeXjExMZDJZPi///s/XLp0CZcuXUJ0dDSOHTsGHx8fPH36FI6OjkhOTkaNGjXg7u4ulX1qaipatWqFOXPmQKFQ4NGjR/naqB07diA1NRVPnjzBL7/8goiICNjb2+eLu1u3buHGjRs601ia9eiePXvw9OlTADn1X+7vct4+4LFjxwwady+rR8PCwnD16lUpD2q1GmlpaVCr1bCysoKdnR2EEFIdl5qaKg1WJCcn46+//tI5ONGpUyccPHgQJ0+exJkzZ6BWq/HFF18U6feu5mRn7nr877//RmpqKuzs7GBpaYmsrKx8AwuWlpZS/yw1NRWJiYl4/fXX4erqiqysLCgUCuzatQupqamoX78+WrVqhfXr18PLy0vrpK5Gq1atsHTpUgghoFKpsG3bNqjVaqk+bNWqlcHi8tmzZ1I94Ovri0ePHknte7169XDx4kUEBwfj+PHjWnVgbsYcl4XJPQag6ZOmpqaicuXKsLGxwZ07d5CQkKDztZ06dUJ8fLxWfahpG3O7f/8+nj59ipiYGHzxxReoXr06zp07V+Q0+vr6Ijs7G3v37gUA7N27t8BBty5dumDGjBnSdzA1NbVUZppZ6v2I/5+Liwt+/fVXfPTRR8jIyIBMJkNcXBw6duyIXr16oX79+qhYsWK+3WE++ugjtG7dGnZ2dti+fTvc3d3zHbtSpUpYuXIl+vXrB6VSCRcXFyQkJLx0xkLt2rWxatUqDB06FM+ePUNWVhYaNGiAVatW5Xuug4ODNPrt4eGB5s2bS5c7PX36VBrZrVWrFiwsLBAYGIiwsLBi7ZYB5FxaM2zYMMybNw8WFhZYtGiR1GnK+1kUZty4ccjMzMTrr78ufQ7jxo1D/fr1dT7f2dlZ6iSlpaVJU9zDwsIwdepUpKSkoGLFihg1ahROnjyJSZMmoW/fvli+fDlq1aqFli1bFpqekSNHYvDgwbCzs5MGlzTefvttPHz4UJpyqZm+2aBBA0ycOBEXL16EtbU17OzsMH/+/ALfo3bt2li6dCn69OmD7OxsyOVyLFiwAI0bN5bi4+rVqzhz5oxWfHh4eGDq1KmIi4uDvb09vv/+e2zatAnBwcE4cOAAWrZsKV2qVqlSJTRq1AhWVlbIysrCe++9J5WF5rG0tDS8//77aN++PVauXInOnTujR48e8PT0RL169aT39fLyQmZmJgICAqBUKqUfXt988w0GDBiApk2bQqFQQKFQICUlRSofzVlODQsLC6Snp+O9997DsmXLsGDBAiQkJKBRo0b48ccfAeRM3/3666+NNi6dnZ2lQcrc8Qdon1Hq2LEjkpKSEBYWhnHjxklxuWLFCgAwyrgMCAjApEmT0KxZMzg4OEClUmldyjdnzhz07t0bPXr0QM+ePeHg4ABnZ2cEBAQUWjcCwE8//YTx48cjIyMDmZmZOHbsGFQqlTSzQTP7QUNTVs+ePUOzZs0A6I4NzW5HQggoFAq8++67WoMfuctHJpMhPT0dwcHBGDlyJJYtW4agoKBC69Pc/P398dVXX0n1qY+PT5nH7VtvvVVo/E2cOBFTp07F4sWL0aVLF7i4uBRanxYkd9zWrFkTarUagYGBUtklJCTAyckJz58/x+zZszF06FCt12vS/Pz5cwwcOBCOjo754lYzGNaqVSsAOeXz3nvvwcLCAj/++CPS0tLQvXt3uLq6YvTo0fjoo4/g7OwMZ2dnrFq1CiNGjEBGRgbWrl1b4A88V1dXuLm5SW1u5cqVIYRAUFAQ7Ozs4OPjAyCn/KZOnYpbt25h2LBhWLBgQZHbXA0/Pz80a9ZMOnPdq1cv/P3332jXrh1SUlLg7++Pu3fv4sCBA1i1ahXS0tLyTflft24d4uPjYW1tDbVaLcXRJ598gpEjR0qX/nXv3h09evR4aTl+++236NatG5ydndGyZUsppori9OnTmDBhAoQQUCqV6Nu3LwIDA1G3bl14enrC398frq6uCAwMRJ06dfJ9/zt27IiTJ0/mq3eHDBmCbt26wc/PDzVr1kSTJk2k9yxOfZVXUfpJmv7doEGDcO3aNTRs2BBxcXHw8fFBlSpVUL9+fdja2uLhw4dQKpX45ptv0LJlS2RmZqJ169a4efMmqlSpAiDnB/L58+e16t3u3bvn69+5u7tj7Nix+PzzzxEUFIS2bdvCyclJSpOVlRU++eQTKd0PHz7EzZs3tfL+66+/4t1335VuDx48GGfOnMHly5dRr149VKhQAWq1Gj/++CN+++03qFQqrFmzRvree3l5ISQkBHZ2dqhfvz7GjRuHv//+G1ZWVti9ezeEENi2bRu++uor6T2cnJzw3XffYf78+ejcubM0o2TlypUYPXo0Dh8+jH/++QdVqlTBiBEj0LhxY9y5c0drptq6detw+/Zt9OrVCwqFQi/x/OGHH+KPP/5At27dULlyZekHFJAzY0UzqKmJv0aNGgEoOJ6VSiXq168Pf39/1KxZExs3bixWX7+o8VytWjVkZGQgMDBQ5+8KzcyY3Ozt7aX3iouLg1wuR3p6OkaNGoXo6GgcPnwYR48eRcuWLbF9+3YEBgYiKysLISEhUv++SpUqeOONN6BUKpGRkSEN+OT9vsjlcnz88ceYP38+vL29tWZQ5Obv749x48ZhypQpqFSpEg4cOIAff/wRc+fORf/+/fHNN9/ggw8+QJUqVfDgwQOkp6fj7NmzWLBgAZKSktC2bVskJiZixIgRyMrKwh9//KHVRlWtWhVVq1ZFZmYm6tSpgyVLlmjFnb+/PywsLFChQgUsXLgQ3t7e+dJYmvWoq6urdEI5LS0Nlpb/+zmatw948uRJeHp6GjTuCqtHXVxcsHDhQnTv3l2Ky3fffRfW1tbo3LkzlixZggMHDsDKygqZmZkYNWoUVq1aBSsrK0RFRUEmk+m8bC08PByzZs3CpEmTEBoaCiEEKlWqhG3btr30965cLsc333wjxaVKpcLChQvx888/Y/bs2Th48CDCw8PRqlUrrcEFuVwu9c8ePHiAhw8f4ujRo7hz5w7q1q2Ls2fPIj09Hba2tkhPT0dkZCQePnyo9bnl9sknn+Ddd99F3bp1cefOHVSvXh2bN29GQEAAOnfujF69egEwTFxq8vDtt99K3yNPT09UrVpVWqLDwsICoaGhWnVgboauD4vbvmtoxgBkMhlUKhVcXV3Ro0cPdOvWDfXr10eVKlWkPmVe4eHhmDlzplQfZmVloX379vlmGl6/fh2DBw9Gdna29DslNjZWurT4ZRQKBX7++WeMGDECarUajRo1gq+vr87LUefMmYPx48cjODgYMpkMlpaWmDlzJmrXrl3sz6ZQ+l0izXwcP35c9O3bVwiRs5i5ZmFzIYT49ddfxSeffCKEyFl8TrMgnynRlT/NQn+aBYCFyFnQdu/evYZK5ivr37+/tDOIRlHLL+/i20V9TB/KY/xp5I2/8ePHv9Li8IaQe5H40NBQ0axZM63Hjb0MGX85XqX+e9VNDfTlVerAspZ7YVfGYI7y3AbrU2ZmpqhWrZrIzs4u0vMLK58PPvhAWqQ/d/oPHTok2rdvLz3vZZtEGBLjzzjrj3r16om7d+++Uv3n4eEhBg0aVHaJLoHyEn9CmF4MakyfPl3aHE0fbbCx1YflJQZNNf6E0P4tdeTIEVG5cmVpJ2BDKPM1zUxFcHAwPDw8EB4ejrNnz6Jr167SLIAOHTrg2rVrCA8Ph42NjbSAqCnRlT/NDIa1a9eicePGCAsLg5eXl3R5k6lp164dtm/fjsGDB2PZsmUmVX7lMf40+csbf76+vgZObfFpzng4ODjg5MmTePr0qUnFIOPP9Os/gHWgMSsPMWhM8WdtbY2rV69qzWgpTGHlExQUhOfPn+dLf5MmTaRLB40d488464+zZ8/C3d2d9Z8ZxB9gmjGoMX78eHTs2JExaMIxaMrxB+SsbRgUFITAwEAMHToUP/30E+zs7AyWHgshCljd1wgMGzZM57W6hw4dyrfgnDG4d++ezuun//33X2nR29zmzp0r7ehTVgpKY+vWrbUuKTAmnTp1yrfbmouLC/bs2WOQ9DAu9Y9xWfoYt/pXlnGrr/Irq7g1hfLTJ2P/fhlbfWXsn1dejOccxlI+JY1nY89XXoy7HMZSPqVVjxp7vvNiXOYwlvIxtvbdXBn1oBkREREREREREZEh8PJMIiIiIiIiIiKiPDhoRkRERERERERElAcHzYiIiIiIiIiIiPIo2jZC5YxarcatW7fg4OAACwsLQyeHSpkQAhkZGahSpQpkMuMYR2YMlh+MPzIkxh8ZEuOPDI0xSIbE+CNDMsb4M1YcNNPh1q1b8PHxMXQyqIxdv34d3t7ehk4GAMZgecT4I0Ni/JEhMf7I0BiDZEiMPzIkY4o/Y2WSg2aHDh3ChAkTAOR8sdu3b4/jx49DpVJBLpfjnXfeQd++fXHnzh3069cPT58+xfDhw9GnT58iHd/BwQFATgA5OjpK92dnZ2P79u2IiYmBlZWV/jNmYOU1f+np6fDx8ZHK3RjoisHyWj7mgvFn/Mw9j7ryZyrxB5TP8jEnjD/jVl7zZyoxWF7Lx1yYevwB5beMzIWptMHGyiQHzZo2bYrExEQAwIABA9ClSxccP34cW7Zsgb29vfS8GTNmYOzYsYiKikJ4eDi6desGGxublx5fMxXV0dExX2VhZ2cHR0dHs/0ylef8GdMUZF0xWN7Lx9Qx/oyfueexsPwZe/wB5bt8zAHjz7iV9/wZewyW9/IxdaYefwDLyNSZShtsrEz64tWsrCwcOXIE4eHhkMlkaNeuHTp16oR///0XAHDkyBG0bNkSlpaWCAkJwZkzZ/Tyvv5TtunlOEQlwfgjouKoPn4Tqo/fZDZ1h7nkg0wT448MyX/KNlQfv8nQySAiE2FufUBDMcmZZho7d+5EdHQ0ZDIZEhISULFiRezduxcffPABNm7ciOzsbGlROycnJzx69EjncTIzM5GZmSndTk9PB5AzIpudnS3dr/lbIRNa95sLTZ7MMW9Awfkz1/wSERERERERUcmZ9KBZQkICBg4cCACoWLEiACAyMhIfffQRAMDKygpqtRoymQxpaWlwdXXVeZzp06fjs88+y3f/9u3bYWdnl+/+uBA1Nm/erK9sGJ0dO3YYOgmlKm/+nj17ZqCUEBEREREREZGxMtlBs+zsbBw9ehSLFy8GkDM7zNHREefOnYOLiwsAIDQ0FImJiYiIiMCxY8cwc+ZMnceaMGECRo8eLd3WLIoXExOT71ruHTt24JMkGY592rYUc2cYmvy1bt3abK/l1pU/zcxCIiIiIiIiY7d69WqMHDkS9+/fR0JCAubMmQNbW1ssX74c3t7euHDhAoYMGQKlUom4uDhER0cbOslEJstkB8127tyJli1bSpdftmzZEra2tgCA77//HgAwbtw49OvXD5MmTcKwYcOkx/NSKBRQKBT57reystI5eJSptjDLQSWNgvJtLvLmz5zzSkRERERE5kOlUiEhIQE+Pj5QKpWYPXs29u7di6NHjyIuLg4LFy7ExIkTsXjxYnh4eCA2NpaDZkSvwGQHzWJjYxEbGyvdTkpKyvccT09Ps7/UkIiIiIiIiMqH1atXo3v37pg1axaSk5Ph5+cHa2trNGvWDGPGjAEA3Lp1C3Xq1AEAuLq64sGDB3Bzc8t3rOKu7W2ua0Gba/4UcpHzvyznf11lSi9nsoNmREREREREROWFSqXC2rVrsWHDBsyaNQupqalaywmpVCoAgFqtlu7TbIina9CsuGt7m/uEFHPL38zG2rdz54/rehcdB82IiIiIiIiIjFx8fDx69OghLVHk7OystT6zXC4HAOlxAIVuiFfctb3L29rXps5/yjYAOTPN4kLUWvnjut5Fx0EzIiIiIiIiIiN37tw5HD9+HPHx8UhOTsbcuXNx/vx5ZGVlISkpCYGBgQBylim6dOkS3N3dC5xlBhR/be/ytva1qctUWWjdzp0/c8pnaeOgGRGRCUpMTERcXBzUajVGjhyJSpUqYezYsZDJZJg/fz4CAgJw584d9OvXD0+fPsXw4cPRp08fQyebiIiIiEpoxowZ0t8hISGYP38+1qxZg6ioKNjY2GD58uUAgGnTpmHAgAFQqVQ6L78koqLjoBkRkYl5/vw5Zs2ahS1btsDa2hoAEBkZiU2bNiEjIwPDhg3D5s2bMWPGDIwdOxZRUVEIDw9Ht27dYGNjY+DUExERmTaeuCJjoNkIr2fPnujZs6fWY/Xq1cO+ffsMkSwis8NBMyIiE3Po0CHY2tqiY8eOsLOzww8//AC5XA4XFxe4uLjg0aNHAIAjR45g1qxZkMlkCAkJwZkzZxASEmLg1BMREZkunrgiIipfOGhGVEI8y0iGcvfuXaSkpODw4cPYuXMnJk+erLVYq6WlJbKyspCdnS0tBKvZOUmXomw3bq5bcedmrnnkduNkbtj+kiHp+8RVcdpgXfW4OTDX9lejoPyZYn79p2zDxWkdDJ0MojLFQTOiEuBZRjIkZ2dnNGvWDNbW1oiOjsbkyZPh4OAgPa5UKmFtbQ0rKyuo1WrIZLJCd04qznbj5rYVty7mlkduN07mhO0vGZq+T1wVpw2OC1EDADZv3qyv7BgVc2t/88qbP7bBRKaBg2ZEJcDL48iQQkNDMWvWLAghcOLECdSrVw9XrlzB48ePkZGRIQ2OhYaGIjExERERETh27Bhmzpyp83hF2W7cXLfizs1c88jtxsmcGGKWj+Y2kPM9MsXZIS/DmT5Fp+8TV8Vpgz9JkiFTbYEzU9oUO93GzFzbX42C8sc2mMg0cNCMqAQMeXkcO+ymSZ8ddjc3N7zxxhuIjIyEhYUFlixZgps3b6Jdu3awsLDADz/8AAAYN24c+vXrh0mTJmHYsGGwtbXVebzibDdubltx62JueeR242RODDnLB8iZ6WOus3wAzvQpCn2fuCpOG5yptkCmysJs625za3/zyps/c84rkTnhoBlRCRjy8jh22E2bvqbmjxgxAiNGjJBu16pVCwcPHtR6jqenp9l/nkRUvhhilg+gPdPn2KdtSy+DBsKZPkWn7xNXRERk3DhoRlQChrw8jh1208Sp+WRuVq9ejZEjR+L+/ftISEjAnDlzYGtri+XLl8Pb2xsXLlzAkCFDoFQqERcXh+joaEMnmcyAIWf5ADkzfcyxjdLgTJ+i4YkrIqLyg4NmRCVgyMvj2GE3bZyaT+ZApVIhISEBPj4+UCqVmD17Nvbu3YujR48iLi4OCxcuxMSJE7F48WJ4eHggNjaWg2akF5zlQ0RERGWJg2ZEJcSzjERUXq1evRrdu3fHrFmzkJycDD8/P1hbW6NZs2YYM2YMAODWrVuoU6cOAMDV1RUPHjyAm5tbvmNxIfYc5XFdx5Lmle0vERERlRUOmhEREVGRqVQqrF27Fhs2bMCsWbOQmpqqtfaTSqUCAKjVauk+zULsugbNuBC7NnMf6Mmdv5Ku6UhERERUVjhoRkREREUWHx+PHj16SDsTOjs7a63NJ5fLAUB6HAAXYi+C8riuI9d0JCIiImNnkoNmV69eRWhoKOrXrw8ASEhIQGJiIhchJiIiKmXnzp3D8ePHER8fj+TkZMydOxfnz59HVlYWkpKSEBgYCCDn8rhLly7B3d29wFlmABdiz6s8retozvkkIiIi82CSg2YAEBkZiXXr1gEAFyEmIiIqIzNmzJD+DgkJwfz587FmzRpERUXBxsYGy5cvBwBMmzYNAwYMgEql0nn5JRERERGRsTPZQbMDBw4gPDwc4eHh6Nu3Lxch1oPyuAixrttERFQ0SUlJAICePXuiZ8+eWo/Vq1cP+/btM0SyiIiIiIj0wiQHzTw9PZGSkgI7OzsMHjwYv/zyCxch1qPytAgxwIWIiYiIiIiIiCg/kxw0y73+yZtvvolly5bB3t5eepyLEJdMeVyEGOBCxERERERERESUn0kOmmVkZMDBwQEAsG/fPrRv3x4LFizgIsR6Up4WIdbcJiIiIiIiIiLKzSQHzfbv349JkybBzs4ONWrUQFxcHGxsbLgIMRERERERERER6YVJDprFxsYiNjZW6z4uQkxE5c3q1asxcuRI3L9/HwkJCZgzZw5sbW2xfPlyeHt748KFCxgyZAiUSiXi4uK4gzAREREREVExyF7+FCIiMjYqlQoJCQnw8fGBUqnE7NmzkZiYiKlTpyIuLg4AMHHiRCxevBhbt27Fp59+auAUExERmY/Vq1ejUqVKAICEhASEhYUhOjoaN27cAABcuHABERERCAsLw65duwyZVCIiegUcNCMqIXaWyJBWr16N7t27QyaTITk5GX5+frC2tkazZs1w6tQpAMCtW7dQp04dODo6wtXVFQ8ePDBwqomI9INtMBkST1wREZUfJnl5JpGh6eos7d27F0ePHkVcXBwWLlwodZY8PDwQGxvLS+NIb1QqFdauXYsNGzZg1qxZSE1N1drpV6VSAQDUarV0n5OTU4EbomRmZiIzM1O6rdlRNjs7G9nZ2dLfuf83R+aaR4Vc5Pwvy/k/d/7MLa9UPrANJkPTnLiaNWtWvhNXY8aMAfC/E1cApBNXr9oG66rHzYG5tr8aBeXPXPNLZG44aEZUAvrsLBEVV3x8PHr06AGZLGeysLOzs9TJBgC5XA4A0uMAkJaWBldXV53Hmz59us7NUrZv3w47Ozut+3bs2PHK6Td25pbHmY21b+fO37Nnz8o4NUSvjm0wGZK+T1wVpw2OC8k55ubNm/WSF2Njbu1vXnnzxzaYyDRw0IyomPTdWQKKf5bRHM9M8Sxj0Z07dw7Hjx9HfHw8kpOTMXfuXJw/fx5ZWVlISkpCYGAgAMDT0xOXLl2Cu7t7ofE3YcIEjB49Wrqdnp4OHx8fxMTESLGdnZ2NHTt2oHXr1rCysip2mk2BuebRf8o2ADl1R1yIWit/uQdbiUyBIWbaam4DbINNlT7bYH2fuCpOG/xJkgyZagucmdKm2Ok2Zuba/moUlD+2wUSmgYNmRMWk784SUPyzjOZ6hhHgWcaimDFjhvR3SEgI5s+fjzVr1iAqKgo2NjZYvnw5AGDatGkYMGAAVCqVzvjSUCgUUCgU+e63srLK13nVdZ+5Mbc8ZqostG7nzp855ZPKB0POtAXYBps6fbTB+j5xVZw2OFNtgUyVhdnW3ebW/uaVN3/mnFcic8JBM6Ji0ndnCSj+WcZjn7Yt3UwaAM8ylkxSUhIAoGfPnujZs6fWY/Xq1cO+ffte6fhERMbEEDNtAbbBpk6fbbC+T1wRFceRI0cwatQoWFlZwcvLCytWrMCGDRswZ84c2NraYvny5fD29saFCxcwZMgQKJVKxMXFcV1HolfAQTOiYiqNzlJxzzKaY4dWg2cZiYioIIacaQuwDTZ1+m6DeeKKypqPjw92794NW1tbTJgwAb/99hs3QyEqZRw0I3oF7CwREREZBttgIipvPD09pb+tra1x8eJFboZCVMo4aEZERERERERkIv79919s374dX375Je7fvy/dz81QSsZcN0NRyEXO/7Kc/3WVKb0cB82IiIiIqNRUH78JCrnAzMaGTgkRkelLT09H3759sWzZMqhUKm6GokfmthlK3nY3d/5KshFKecVBMyIiIiIiIiIjp1Qq0atXL0yePBm+vr7Izs7mZih6YK6bofhP2QYgZ6ZZXIhaK3+vuhlZecJBMyIiIiIiIiIjt3r1avz111+Ii4tDXFwchg8fjg8//JCboeiJuW2Gkqmy0LqdO3/mlM/SxkEzIiIiKjJud09ERGQYffv2Rd++ffPdz81QiEqP7OVPISIiIsqh2e7+zz//RPXq1aXt7hMTEzF16lTExcUBgLTd/datW/Hpp58aONVERERERMXHQTMiIiIqMk9PT9ja2gLQvd39qVOnAPxvu3tHR0dpu3siIiIiIlNikpdn6ro0pF69evDy8gIAfPzxx2jdujUvDSEiIiol3O5ev8x1u3sgZ8t7bndPREREpsgkB800l4bY2tpiwoQJ+O233+Dk5ITExESt52kuDfHw8EBsbCwHzYiIiPSA292XHnPb7h7Q3vKe290TERGRKTHJQTNPT0/pb2tra8hkMjx58gSRkZHw8vLCvHnz4OrqKl0aAkC6NIRnuQtmzme5gYLzZ675JSIqDdzuvnSY63b3QM6W99zunoiIiEyRSQ6aaWguDZk0aRIiIyNRsWJFrFixApMnT8bcuXOLfGkIz3JrM8ez3LnlzR/PdJOp4e6FZEjc7r50mdt294D2lvfc7p6IiIhMickOmuW+NMTKygoVK1YEAHTr1g2LFi0CUPRLQ3iWO4c5n+UGCs5fSc50c9CCDEnXJeqzZ8/G3r17cfToUcTFxWHhwoW8RJ1KBbe7J0NjG0yGxPgjIipfTHLQLO+lIVlZWRBCQKFQYN++fahduzaAol8awrPc2szxLHduefNXkrxy0IIMKe8l6nl3LxwzZgwAFPkSdSIiU8I2mAyJ8UdEVL6Y5KCZrktDZs6ciQoVKkChUGDJkiUAin5pCFFxcdCCjEFZ7l5o7mseAuabR4U8Z9dC7l5I5oJtMBmSvuOvOG2wrnrcHJhr+6vBdZWJTJtJDprpujQk72UhAC8NodJniEELbkRhmvTdYTLU7oXmvuYhYH55zL1zIcDdC8l86KMN5mZQOdgGF5+++oDFaYPjQnKOaa5rK5tb+5sX11UmMk0mOWhGZAwMNWjBjShMmz46TIbYvdDc1zwEzDeP/lO2AQB3LySzoq82mJtBaWMbXDT67AMWpw3+JEmGTLUFzkxpU6J0GytzbX819LmuMhGVPQ6aEZWAIQctuBGFadJnh8mQuxea+5qHgPnlMffOhQB3LyTTp882mJtB5WAbXHT67gMWpw3OVFsgU2W+ayubW/ublz7WVSaissdBM6ISMOSgBTeiMG366DBx90IiKs/02QZzMyhtbINfTt99QCIiMm4cNCMqAQ5aEBERGQbbYDIkxh8RUfkie/lTiIiIiIiIiIiIyhcOmhEREREREREREeXBQTMiIiIiIiIiIqI8OGhGRERERERERESUBwfNiIiIiIiIiIiI8uCgGRERERERERERUR4cNCMiIiIiIiIiIsqDg2ZERFRk/lO2ofr4TYZOBhERERERUamzNHQCiIiIiIiIiMxN9fGboJALzGxs6JQQUUlxphkREREREREREVEeHDQjIiIiIiIiIiLKg5dnEhGVMk7NJyIiMgy2wURE9CrKxUyzcePGITw8HH379kV2drahk0MGUH38JvhP2WaQ92b8kSEx/sjQGINkSIw/MiTGHxkaY5Do1Zn9oNnJkydx8+ZN7Nu3D3Xr1sW6desMnSQqRxh/ZEiMPzI0xiAZEuOPDInxR4bGGCTSD7O/PPPgwYOIiYkBALRt2xZLly7FW2+9pfWczMxMZGZmSrfT0tIAAI8ePdIakc/OzsazZ89gmS3Dw4cPyyD1ZUuTv4cPH8LKysrQydErS+VTWKoFnj1T58tfRkYGAEAIoff3LUr8AUWLQcaf6TK3+FOpLRiDJsRS+TTnfx0xWJrxB7ANLg5zjT+g4DqQ8Wc8ymP8AWyDjQXjj3WgoZlrDBqyD2hOzH7QLDU1FZ6engAAJycnPHr0KN9zpk+fjs8++yzf/TVq1CjwuG5f6S+NVHZ6F/JYRkYGnJyc9Pp+RYk/oPgxyPgzTeYSfwDgNks/aaSyVVAMlkb8AWyDSRvjjwyJbTAZUlnHH8A6kLSVdRtsTsx+0MzZ2Rnp6ekAckbOXV1d8z1nwoQJGD16tHRbrVbj0aNHqFixIiwsLKT709PT4ePjg+vXr8PR0bH0E1/Gymv+hBDIyMhAlSpV9P6eRYk/oGgxWF7Lx1ww/oyfuedRV/5KM/4AtsHFUR7zx/gzHuU1f2yDjUN5zR/rQONRHvNX2vFnTsx+0CwsLAyzZ89Gv379sG3bNjRr1izfcxQKBRQKhdZ9zs7OBR7T0dHRLL9MGuUxf6U1ul6U+AOKF4PlsXzMCePP+Jl7HvPmrzTPLrINLr7ylj/Gn3Epj/ljG2w8ymP+WAcal/KWP84wKxqz3wggODgYHh4eCA8Px9mzZ9G1a1dDJ4nKEcYfGRLjjwyNMUiGxPgjQ2L8kaExBon0w+xnmgHAV1/xwmsyHMYfGRLjjwyNMUiGxPgjQ2L8kaExBolendnPNNMnhUKByZMn55vCai6YP+Nm6ul/GebPuJl6+ovC3PNo6vkz9fS/DPNn3Ew9/S/D/Bk3U0//yzB/xs8c8lAY5o8KYyG4xygREREREREREZEWk59pNmXKFLx48aLM3zcxMRHBwcHFft2JEyfw888/6z9Br+CPP/5AVFQUACApKQk9e/YEADx+/BhffvnlS19/9erVAheMnDJlCgIDA5GRkVHg64v6PsVR0vJZtGgRqlWrBqDwfJXUp59+ipUrV+p8bN68eRgwYEC++0uzfICc9Q7Kunz0ZePGjfjPf/4DICefCxYs0Hq8Xbt2uHjxYomPHxUVhQ0bNki3s7OzUb16dWRlZQHQXf/o+rzyHudVvUr9M2bMGOkzK+z4W7dulW7funUL4eHhxX6/otB3fIeEhCAxMREA8O6776J27drIyMjAhg0bcPjw4XyvN3R8P3nyRGt3qoKURn2koSuOx4wZgylTppTK+2kUFMcvi7dXaUeXLVuGLl26lOi1hWE9XbT+WFHq5LzHKazdBMq2P1ZY+rt164Zly5YVOx25MY6Mx6uWhca6deswfPjwQp+Tu14qaTwX5t1338WePXt0PjZ69GhERUWhbt268Pf3R3BwMIYMGYLHjx+/clp09ctexb1799C2bVvUqVMH/v7++PPPP4v1+pfVJUV53ZQpU/Dhhx8W+xiAdv2xbNkyXLhwoUTHMSdKpRKfffaZScTfF198AV9fX8hkshL16Rl/ps3kB80+++wzgwyalVRZDJoplcoSvzYkJARr1qwBoJ/O02effYY///wTDg4OBT7nVd7nVfKqS0pKCh4/fqzXY+Y2depUqeNVEvouHyAnJk2lfPIeu1OnTpgzZw4A3Y3j5s2b4evrq7f3TExMRNOmTWFtbQ1Ad/1TWj869PFZnjhxAikpKdJnVpC8g2ZVqlTBvn37ivQe+q5/XuV4ixYtQkpKChwcHEpl0Kw047sslVY7WtLP52XxVt7aUcA06umixFFR6uS8x5k6dSrefvttvaQxt5LEUXHbFMaR8TBUWfz666+lMlBfHIsWLUKLFi10PrZt2zbcvn0bhw4dwpkzZ3D8+HG0bt0ajx49euX3fZVBC13lNX78eDRp0gTJyclYunQpevfujezs7CIfs6R1yavWQWq1Gmq1Wqv+4KBFjnfeeQdJSUkmEX+tWrXCli1bEBERUaJjMv5MnDAiBw8eFM2aNROBgYEiICBAbNiwQQghxEcffSRCQkJEUFCQCA8PFxcuXBBCCDF06FABQPj7+4ugoCBx9+7dAo+9detW0aBBAxEQECAiIiLE2bNnhRBC7NmzR9SvX18MHz5cBAYGinr16omjR49qva5Zs2aiYcOGIjQ0VOzevbtIr1uxYoUICAgQAQEBol27duLGjRvi7t27wsfHRzg6OoqgoCAxdOjQYn9GN27cEF27dhX+/v4iICBATJo0SQghRP/+/cXAgQNFeHi48PX1ldLQuHFj0aBBAxEeHi5OnDghhBAiKytLDB8+XNSuXVuEhoaK0aNHi8jISClfQUFBQggh2rRpI2QymQgKChKNGjUqsHzeffddIZfLhYeHh7C3txc+Pj5i06ZNUvloyuj27dtixIgRom7duiIwMFA0bNhQPH/+XLRp00ZYWFgIW1tbYWtrKyIiIsTmzZtFdHS0qFGjhrCxsRFt27aVPmcA4tNPPxUhISGia9euws7OTvj6+krls3DhQhEVFVVg+WRnZwt/f39ha2srFAqFqFKlivjnn3/E3bt3hbu7u5Tn3r17CycnJ53lMHnyZPHmm2+KFi1aCF9fX9GhQwfx4MEDcePGDREeHi5sbW2FjY2NcHNzE4sWLZLKx93dXbi7uwshhFi4cKFwdXUVCoVCODo6iq5du4r+/fuXuHwKcuXKFeHk5CQ+/fRT0bBhQ1GrVi2xadMm6XEAIjU1VahUqgLLJ+/7JCcni+joaBEQECCCgoLEr7/+qnU8TfmMGTNG+Pv7iwMHDkiPL1y4UPTo0aOwMBd//PGHCAkJEYGBgSIoKEgcPnxY57GXLl0qOnfuLIQQwtfXV9jY2IigoCDRsWNHIYQQ1apVE8ePHxdCFPzdWblypWjcuLGoU6eOsLOzE9WrV5fiOzIyUnTu3Fmqfzw9PcWcOXOEEEK8+eabAoBQKBTC1tZW/P777wWWS0BAgHB3dxd2dnbCxsZGyv+ePXuEn5+f8PPzk+KxS5cuIjMzU3qdp6enqFChgnBwcMhX/7zzzjvC1dVV2NjYiFq1aonBgweLzMxMMWPGDGFpaSl8fX1Fu3btxKRJk0RUVJTw8fERtra2wtHRUQwdOlT8888/IiwsTAQGBgp/f3/x8ccfi+PHjwsPDw/h5uYmgoKCxGeffSbFkOYzBCA8PDyEp6enqF69umjWrJlU/1StWlUEBQUJLy8v4ebmJmxsbERgYKDO+qdGjRqiUqVKokKFCqJRo0Zi9erVws/PTzg5OYkaNWoIAKJKlSoiMDBQdO/eXYSGhkrp1Fi3bp2QyWSiUqVKomLFisLa2lpMnz5dCCFEZGSkACDWrFkjnJ2dhZ2dnVAoFMLb29ug8b1gwQJRu3ZtERwcLKZOnSpyN8VHjhwRLVq0EI0aNRLBwcFi7dq1Qoj/fY81evfuLRo1aiS1Mbdv3xZCCNG1a1dRrVo1qZ7+/vvvhbe3t/jPf/4jGjVqJNzd3YW9vb3w9fUV3bt3FwMGDBAARN26dYWDg4OoU6eOiI6OFj179hSTJ0/WSremHQUgqlatKurWrSvq1KkjJk6cKNW3AESlSpWEn5+fGDNmjEhPTxdt27YV9vb2wsbGRlSqVEls27ZN7N+/X1SvXl2rnrazsxNff/21lFdNO1q9enXh6Ogo6tatK5o0aSIqV64sHB0dRa1atUTFihWl9J0+fVpUq1ZNCCHEvXv3ROvWraXv+4ABA4QQQixdulTExMSIrl27CgcHB+Hj4yPVBa1btxZubm6l2o4WxNjq6bFjxwo3N7dC+2P+/v6iQYMGomHDhqJSpUpSW+/q6ioGDhwohBDi4cOHombNmmLv3r1CiJw6+fvvvxcNGjQQHh4ewtbWVvj6+oqgoCCxevVq4eLiIgAIGxsboVAoxLZt20T//v3FnDlzxNatW4WPj49wcXERTk5OomrVqqJFixbit99+E/Xr1xdDhw6V6oD69etLZaKrP9a0aVNhb28v5HK5sLa2FoGBgUKtVgshhJg1a5aUx5CQEHHw4EHpc8ndpiQmJmq14TVq1BBLly4tk/6YqcTRy+rDl6W3oL57XobsGwshRFxcnKhbt64ICgoSQUFB4urVq9Ixq1WrJrKyssTt27dFVFSUaNiwoahXr54YMWKEUKlUYsOGDcLT01M4OjqKsWPHCkdHR+Hn5yeEEOKff/4R7dq1EyEhISIgIEA0btxYDBw4UDRt2lRUqVJFuLq6Cn9/fxERESFmzJghGjduLIKDg0VgYKDYuHGj9PlERkZK5Xbr1i0RExMj/Pz8RNOmTYVcLhf/93//p/NzzZ3vvG1QRkaG1HY9e/ZM9OjRQ/j5+YnAwEDRunVrIYTuflnePM2dO1c6Zt54yqtChQpSWyeEEKGhoWLHjh35nnfo0CHRsGFDERQUJOrXry9++OEHKR40fbjJkyeL7t27iw4dOog6deqI9u3bi9OnT4uYmBhRp04d0atXL6FSqXS+btSoUUIIIU6dOiWaNWsmGjRoIPz8/ERcXJyUBs3vBM1nfePGDan++PHHH0WFChVE9erVRVBQkNi0aVOJ+hL6sGHDBqkeGDt2rKhYsaK4cuWKEKLwsirqd7MwycnJwtbWVty/f1/n48YWfxq5v0+6MP7Mk9EMmj18+FC4u7uLP//8UwghhEqlEg8fPhRC5HR+NVavXi3atGkj3dZ0AApz9+5d4erqKk6dOiWEECI+Pl74+fkJtVot9uzZI+RyufSjfP78+SImJkYIIcSQIUOEg4OD6NGjh8jKyhLJycmicuXK4unTp6Jt27YCgOjZs2e+150+fVp4eHiIGzduCCGE+Pzzz0Xbtm2FEELrR35JREVFiS+++EK6rfls+vfvLwIDA0V6eroQQoj9+/eL2NhY8eLFCyGEEH/++aeoV6+eEEKIefPmiZYtW4qPPvpIhIWFCQ8PDxERESGEyKmgatWqJby9vcXrr78uLC0thRCFl09SUpIAINatWydWr14tGjVqJF577TUhhJAGzVJTU8Xff/8t6tatK1UCjx8/FiqVShw9elRYWFholY+NjY2YP3++VD6Ojo7i6tWrYv78+QKA+Oyzz8SlS5dEkyZNxLfffiveeustqXz8/f3Fl19+KXW8T58+rVU+x48fFzY2NiI0NFSMHDlShIaGSukdN26ccHR0FELkr6Bzmzx5sqhUqZLUeA8fPlwMHjxYREVFiU8++UQolUohhBAXL14UVatWFd26dROBgYGid+/eYs6cOWL//v2iRo0aonfv3kKtVovNmzcLa2tr0b9//5eWT1BQkBg4cKAICQkR1tbWL42ZK1euSOUjhBBbtmyR8qspo8LKR9fn0LhxY7FgwQIhRE4D5OrqKnUSNeWj8d///le89dZb0u3AwEDx559/irFjx4rmzZuLPn36iKysLOnxFStWCJlMJkJDQ0XLli1FVlaWePz4sc5j5/4+5W5cNXL/wCnou/PgwQPx4MED4e7uLtasWSM8PDzEs2fPxMOHD0VkZKRYtmyZEEIItVot3N3dRXR0tMjMzBQ+Pj7SZ7dv3z7h4eEhMjIypM/r8ePHIjQ0VNjZ2Qm5XC5atGghsrOzxZIlS4SVlZVo166d8Pf3lxppIYT44YcfhJeXl5g5c6a4dOmScHBwENHR0Vr1z4sXL6TvRefOncXy5cul+H7nnXfE6NGjhYeHh5g3b55o1KiReOedd4Stra24d++eWLp0qWjQoIH0mY0cOVLrM9F8pydPniwaNWoklc8///wjnJycRFRUlPj8888FAFG9enUxePBgcf78eWFpaSkCAgLEgwcPhLe3t/juu+9EbGys2Lp1qwAgvv3223z1z/bt20W1atVEWFiYiIyMFE+fPhVbtmwRfn5+AoD4+uuvpbiLiYkRiYmJQoicQe82bdqItWvXiszMTFG5cmUpvrdt2yYACB8fHyHE/wbNUlNTRYcOHYS7u7sU36NGjRLNmjUTXbp00YrvPXv2CCsrK1GnTh3RsmXLEsd3QTTtw61bt4QQQkyYMEHq+KWmporg4GDpsfv37wsfHx9x48aNfN/D3O3i9OnTxdChQ8XDhw9FxYoVhYeHh3j48KEICQkRcrlcjBgxQty7d09MmzZNTJ06VaxcuVJ4eXmJqlWrisDAQAFAdOrUSfqheePGDeHm5qY1aJa7HQUgOnfuLPz8/ERKSoo0+HD48GEBQHTo0EGqb3v27Clq1aol0tLShFqtFt27dxcODg7ixYsXwsvLS3rdpUuXhKOjo2jdurW4cuWKsLe3FxUqVBDBwcHC2tpavPfee6Jt27YiPj5eeHp6Ch8fH+Hv7y/c3Ny0PlvNoNns2bPFkCFD8sX20qVLhZubm/jiiy/E9u3bRdOmTaXP0sfHR/j4+BSrHc3MzBSZmZkiKipK5w/swtqR3HXgP//8I8Xxnj17hJubm7C1tRUtW7YUQpRtPf3w4UNhaWkp/ZjQ1R+7dOmSqF27toiOjhZC5Pz4ASDu3Lkjnj9/Lvz9/UWtWrWEXC6XOvhCCOHt7S2cnJxEu3bthFwuF61atRJ+fn7iyZMnYtu2bUIul0t51dRr/fv3Fx9//LFo0qSJGD9+vKhWrZo4cuSIqFy5sujWrZt04m7MmDGiVatWYt68eaJ169aiVatWIiQkRGd/LDIyUvj5+YmOHTuKp0+fikaNGomVK1dq5XHs2LEiMDBQODo6Sm1UtWrVRFJSkhg4cKCQy+WiVq1aQoicHy9WVlbSoFlx+2OvEke5Fae9d3V1Fc2aNRN9+vQR9+/fl+KoQoUKwtvbW0RGRoqWLVvqpb0vSXoL67vnlbd9T0lJEaGhocLS0lK89tprWmXRtm1b0a9fP9G8eXPRrVu3Vy6LR48eCScnJ/Hs2TMhhBBPnz4Vz58/F0IIsW3bNunzeP78ucjIyBBCCKFUKkX79u3FggULhKurq/jiiy9E586dxZIlSwQA6aSDvb296Nixo8jKyhJPnz4VLi4uokaNGqJbt27C0tJSeHl5iWnTpon4+Hjx2muvSfXClStXhIeHhxRzuX/kd+vWTarr58+fL+Ryeb4TJBpFHbT45ZdfpDpfiP/Vt3n7ZUqlUjRq1EicP39ejB07VjRt2lQ4OztLA9MAxIABA7TiT+PBgwf5+rzdu3cXixcvzpfuTp06iVWrVmmVkRD5Bx9q1KghHj16JNRqtYiIiBCvv/66SE9PF9nZ2SIoKEj88ccfOl+nqdPS09Olz/jZs2ciODhYHDp0SAiR83vCyspK2NnZidOnTwsh/tcnVSqVonLlysLPz086VnG/O/qg+Y6dP39eCCGk+Lty5YpQKpXCw8NDNGzYUPTp00c8fvxYBAQEiCNHjohbt24JhUIhGjRoIEaNGlXod7Mwa9asEYGBgQU+XprxJ4QQ//nPf4SdnZ1o27atyMrKkuqzPXv26IxBjZcNmjH+zJPRXJ556NAh+Pr6SuuYyGQyuLq6AgB27NiBpk2bwt/fH1OnTsWJEyeKdey//voLAQEBCAgIAAC8/fbbuHXrFm7evAkAqF27Nl5//XUAQNOmTXHp0iWcPHkSx48fh0KhwJ9//ok6deqgW7dukMlkWL58Odzc3ODr6wsHBwccOnRIeh0A7NmzB23btoWXlxcA4L333sPu3buhUqle6TN68uQJ9u/fj48++ki6r1KlStLf3bt3l6bd//bbbzh58iRef/11BAcH44MPPsCjR4/w/Plz7Nq1C1FRUbhz5w4OHDiA5s2b4/79+1rv1bNnT/z888+oUKECgMLLZ//+/bCwsMDkyZMxdepUXL9+XfoscqtZsyaUSiUGDRqE5cuXIzs7GzKZDCdOnIBMJpPKp1OnTnjx4gXatm0rlU9UVBT27duHpk2bAgAGDRqErVu3IiUlBYsWLUJCQgK6dOkCpVIJpVKJRo0aoXbt2tKlkLnLZ968eXB0dER2djZ++uknXL16FSkpKcUun/bt26Ny5coAgCFDhmD79u3Yv38/evXqhe7du8Pf3x89evTAw4cP8fjxY3Tv3h1WVlZS+dy4cQN//fUXGjRogAkTJsDa2hpKpfKl5ZOWloYqVaogISFBKpuXsbGxwZtvvpnvsyhK+eSVkZGBv//+G++88w4AoE6dOmjevLnW5VSDBg2S/u7Tpw/27NmDu3fvSrHi6OiImzdvYt++fahbty7WrVsnPT8pKQl169bFkSNHsGvXLlhZWcHJyUnnsYuqsO/OlStXEBMTgydPnuCLL77Ao0eP8O+//0rxffLkSTRt2hS1a9fGs2fPcObMGVy8eFHrs2nevDk8PDy06iY7Ozts2rQJTZo0ga2tLYYMGQJLS0sMHDgQarUa2dnZmDt3LiwsLBAfH4/g4GB88803uHfvHlJSUrB161a8ePECycnJCA0Nleqfa9euAcj5Xhw8eBBfffUV5syZg8TEROzbtw+HDx9G27ZtMWLECDRs2BArVqyAUqmU8pNbREQEfvzxR3z88cfYvn27tBbOnTt38OTJE6l8tmzZAiEE9u/fL03tPnToEIQQSE1NhYWFBdq0aYObN2/C0tIS169fx8mTJzFu3DhYW1vju+++06p/+vXrh+3bt6Nfv34YNmyY9HnZ2NgAAKysrPDGG28AAJ4+fYpdu3Zh1KhRCA4ORkhICFJSUnDx4kVcuHABcrlciu+YmBhUr15dqttzc3BwgFqtxqBBgxAXF4cbN25g//79qFWrltblHc+ePYNKpcK5c+ewa9euEsV3Yetx7d69G7GxsfD09AQArbVuDh48iMuXLyM2NhbBwcFo1aoVAOhcQ2nVqlUICQmBv78/Fi1ahBMnTuDQoUOoV68eGjZsiJ07dyIhIQFWVlbo1q0bduzYgS+++AKff/45BgwYgNu3b6NChQrSJRB79+7Fu+++CwDw8vJCp06dtN4vbzv67bff4tatW1AoFAgKCoK7u7vUjn7wwQdSHbNp0yZcvXoVHh4esLOzw8aNG6FSqXDt2jXExsbCwcEBr7/+OpYvX47OnTvj8uXLAICsrCx4eHhg6tSpqFKlCoKDg7F792706tULDx8+hKWlJebOnQu1Wq2zDmzSpAm2bNmCjz76CL/99pvUlr148QIPHz7ERx99hNatWyMtLQ03btzAv//+i/v372PQoEFFbkf79esHa2trWFtbF7teOnnypFYduGXLFq16ulevXsjKysKuXbu0XlcW9fSiRYugUCjw/vvvA9DdH2vevDmuXLmCvXv3Ijg4GN26dQMA3LhxAzY2NlizZg3u3bsHDw8PKQ0AkJmZCTc3NwQFBaFBgwZISUnB5cuXce7cOVhbW6N27drSc3O3VRcuXEBKSgqWLl2KJ0+eYPDgwZDJZPD19cWtW7dQu3ZtXLp0CX369EHz5s1x+fJl9O/fH48fP9bZHxNCICwsDDKZDHZ2dujTpw927twJADh+/DhCQkKwcOFCWFhYID09HatWrZLSpVlyQqVSoXnz5jh06BACAgK0vvfF6Y+9ShzpUpT2/unTp8jOzpbatfXr12vFUc+ePZGYmIgNGza8cnv/svUwC0rvy/ruGrra96pVq2LTpk2oVq0aYmJitMriyJEj2Lp1KzIyMrB7927cvXv3lcrC0dERderUQZ8+fbBw4UI8evRIass2bNggtWVqtRrjxo2TYj8pKQnbt29HYGCg1B70798fVlZWePHiBS5cuAClUomjR4+iTp06CAsLQ3Z2Ntzd3WFhYYFmzZrBz88P69evlz6bli1bwt/fH126dMGjR49w5cqVfOndtWuXVNe7urrC0dGxSPksTFBQEM6fP4/33nsPa9askfq6eV28eBFnz55F586d8eOPP+LZs2cAgBUrVkjPadeunRR/eeu/omrRogXi4uIwdepU7N+/Hy4uLjqfFxMTAxcXF1hYWKBhw4aIioqCg4MDLC0t0aBBAyQnJxf6Ps+fP8e7776LgIAANGnSBP/++6/UD7SyskKPHj3QvXv3fK/7448/YG1tjS+++AJPnz7FoUOHSvTdeVWHDx9GYGAg6tatCyAn/jTLj/z222948OABVCoVdu7ciaCgIGRkZODcuXOYN28ePDw88Pfff+Pp06eoWbOmzu9mWSlu/PXq1Qu+vr5YtmwZ3N3dYWdnJ/0O0XzvXyUGGX/myWgGzQpy7do1vP/++4iPj8eZM2fw888/633tFU3jBgByuRxKpRIHDx5E3bp10bp1a/zxxx/o0KEDTpw4gZs3b+Lq1asICQmBjY0N2rZtiwMHDkiv06Uoizzrg729vfS3EAL9+/fHiRMnpH+3b9+Gra0tgJy1u2JiYgDkVDZpaWlax1q/fj26d++OzMzMQt/z2rVrmDx5MipUqCCVT2Zmps4BKCcnJ5w5cwa9e/fGhQsXEBgYiJSUlJfmy8bGRvoM5XK5lFchBFq3bo1Tp05hzJgx6NGjB1q1aoXRo0fD0tJS+pGkeZ2mfA4dOoSsrCzs3bsXP/74I0JDQ1+ahqLQpPH9999H8+bNcfr0aZw4cQKvvfYaVCpVvvJxdXXF4sWLpfIZN24cLC0tARRePs+ePZMes7S0xIEDB16aNoVCofUZ6rN8cuddI3debW1tMWDAACxcuBDff/89RowYgYMHD0p50HyHcvv3338RHh6Ob7/9Nt975T62PvTq1QsxMTFo1KgRTpw4AXt7e6mOefHiBX788UfEx8ejR48eGD16dIH1T97PwMrKSmtQO3c9I4RAnTp1pNd169YNJ06cwPr161GlShUsXLgQQgi4uLjg22+/lWLk5s2b0utsbGwghMD69euxfv16eHp64uLFi+jVqxeAnPUYzpw5AxcXF4gCNknu2rUrDhw4AF9fX8ybNw8dOnQAAFy/fl3aEKNt27ZISkqSXqNZFyx3/SeTyaS6RZM/Tf1TtWpVLFq0SKv+KexzA3IG0DQ/4DRpP3z4sPQ5pKSkYNKkSdLzc8e3hYUF1Gp1vmNaW1tj3Lhx6N27N/bv349du3YhJSUFkZGR+eputVqNqKgoKf6KG9/FkfvYQgjUr19fq96+du0aWrZsqfWa/fv347vvvsPmzZtx5swZzJ49WysuR40ahQULFmDbtm3w9PREdnY23n//fdSoUQO///473n77bdjZ2eHcuXP45ptvXpquosjdSbW3t5c+UyEEYmNj8fz5czx//hwvXrzA06dPUadOHbRp0wYZGRl4/vw5VqxYgTfffFPrdVWqVAGQ88PuyJEj0vHVajXc3NxgaWkJe3t7qf7I/Rk0bdoUJ06cwOuvv45ffvkFoaGhOuu9kSNHYu7cuViwYAHq1KmjtYj6y9rRV/m88taBSUlJWnGsGVzMWweWRT29adMm6bPPLXd/7OOPP0br1q1RoUIF6bMBgFq1agEALl++LNWledccevLkCdq2bYvDhw9j0KBBUKlU6NSpE06dOqWzP6bRunVrDBs2DH369JHqw8qVK0OlUul8na4yKaycLCwskJWVhTfffBPR0dH47rvvpIXG9+/fLz3v5MmTiIyMBJDzo0cTf7mPXZz+WFHTV1RFae9Pnz6NOXPmoHfv3nB0dMTo0aO14mj9+vUIDw/HDz/88NI0vmp9WJT0Flfu9tfOzk66XwiB1157DQkJCThx4gT++9//Yvz48a9UFnK5HIcPH8aHH36Ie/fuoUmTJti3bx+EENi2bRtiY2MBALNnz8a9e/fw119/4dSpU4Wux/X8+XM0bdoUrq6uWr9BunbtKv2gBoCGDRsiPT0dQE7f8K233sKZM2fy9WUKonm9ZvCqMJaWllplk/vYNWvWxLlz56Q+gb+/P1JTU/MdQ9P3/fDDD/HNN9/gxIkT2Llzp9S3B3JiSBN/ueu/ihUrwtLSEnfu3JHuu3r1KqpWrZrvfT788ENs2rQJnp6emDhxIt577z2decpbbxRW/+gyceJEuLm54fjx4zh58iSioqKkz0Uul8PNzU3n6w4ePCidiNB8Zq/al9C348ePw8HBASdOnJBi8MqVK+jfvz8uXLiQL/0l0bBhQyQnJ+Phw4cvfa4+4+/EiRNSDF65cgUTJ06U8qCpz3TFYFEx/syT0QyahYWFITk5WTp7pVar8ejRI6SlpcHKygqenp4QQmDevHlar3NwcMg34JNXkyZNcPr0aZw5cwYA8PPPP8PLy0s686hLamoqmjZtip07d+L27dvS2fgjR44gNTVVGpBxcnLKt1hhixYtsHXrVty6dQsAsGDBAkRHR0Mul8PR0fGl6S2Ivb09IiIiMGvWLOm+vDPENDp16oT4+HhpVoparZZ++LZq1QoHDhyAnZ0dsrKysGXLFq0vpp2dHS5evIjff/8dmZmZOHz4cKHlY2lpCZlMVmD55E7r06dPERMTgy+++ALVq1fHuXPnEBERAZVKhePHjwPImZlga2uL7du3A8g5M71///58Cy+2adMGO3fuxKlTpzBixAj897//xbZt2166WKKVlRWePn2KJ0+ewMrKCn/++Sfc3Nwgl8thZ2dX5A7b5s2bcffuXQA5i6zGxMQgIiICFy9eRLVq1WBhYYGNGzfi5MmT+V7bqVMnPHv2DHPnzoUQAo8fP8aSJUuKVD5KpRKOjo7SbLmCYqC4CiofR0dHPH/+XNox0sHBAQ0bNsTSpUsB5Azw6Sqf3DTls3v3brz99ttITU2Vzm7m/Q4NGjQIFSpUwPfff4+tW7fir7/+KtJ3prDvVmHfndTUVMTExCA5ORmffPIJUlNTpfhWKpWQy+Xw9PTEb7/9Js2C8fX1hVqthp2dHdLS0nDw4EHcuXMHwcHB+T6vSpUq4fnz5/j3338B5NQ/crkc7u7uAIAKFSpg69atUhmr1WqkpKSgTZs2SE1NxdWrV6U05x44AIAuXbpgxowZ0mtTU1NRq1YtbN26Fe+//z58fX3Rr18/WFpa4sqVK1LaNJKTk+Hh4YF+/fph5syZ0oCYEEI6ppOTEx4/fgwLCwtERERo7VYlhNAqO19fX2RnZ6Nq1aqIj4/H2rVrkZKSkq/+iY+PR2xsLH766ScsXLgQQE6HP3cnSJNWa2trtGjRQmvh5Vu3buHGjRuoW7cuVCqVlNadO3fqPLMO5Hzv79+/L31PPTw8cO7cOXh6ekKpVErlFRERgdDQULz99tvYunUrNmzYUOz4LkzLli2xdetWqeOfe5HasLAwXLlyRZrxAuQs3K1Jm0ZqaiocHBxQsWJFZGVlSZ+hpp62tbXFnTt3EBcXh5o1ayIjIwNWVlZ48803MXv2bBw8eFD6zFNTU2FtbY1mzZpJddDt27exceNGrffM245+9NFH8PLyQnZ2Nk6dOqV1kiK3du3aYefOnfj777+ltK9fvx4A4ObmBjs7O/znP/+Bu7u71iwjCwsLnDlzBtWrV8fly5fx559/Ijo6GgkJCbC2toYQAjVr1sSDBw9w48YNAMBPP/0kvf7KlSuwt7dHjx49MHfuXPzzzz948uQJbGxsULFiRaku6Nu3LzZv3oylS5fmW+D9Ze1ofHw8srOzkZWVJdWHeeWtD3KXYe46MPcGNCEhIVJ7uHXrVhw7dkx6rCzq6XPnziE9Pb3Q/lhMTAwSExO1Bqg19eG1a9fw3nvvYefOnahUqRJmzJghPUehUCA1NRX//vsv7t69ixcvXsDZ2RkRERHSWXVd/TpfX1/s3LlTanOB/PVhy5YtsWrVKmkgYsWKFXB2dtbZH7OwsMDRo0fx+PFjPH/+HKtWrUKrVq3w4sULZGVlSf21uXPnSuWlkZGRgZo1a8LBwQHz58/Ho0ePcPbs2QI3ryjNOCqpGzduwNraGjExMZg4caI0iK7pV5w+fRq7du1CYmIifH19X6m9L6mi9t2L2zc+ffo0nj59CiAn1s6ePQug5GWRkZGBu3fvIjw8HJ988gmaN2+O48eP48iRI/Dz85N+hKempqJy5cqwsbHBnTt3kJCQgMqVK+PUqVO4ffs2AEjvr1Qq4evrC0dHR+zYsUNqZ9PT03HmzBkEBgbi9OnT2LZtG7y8vPDzzz/DwsICjRo1ko6ja9BAk09NXV+hQgVYWlpi48aNUh2kORmn6e9oVK5cGUIInDt3DoD27LAbN27AwsICnTp1wtdffw0hBK5fv56vX6bJ086dO6X67/Hjx1L+gZwdXi9evIhdu3blq/+6d+8utZlHjx7FzZs3pQHs3C5evIgaNWpg8ODBmDhxos5NgPQhNTUV3t7esLS0xMWLF7Fjx44iv07z2eTuA+vru1NUTZo0walTp6TZ7PHx8VJcW1tbo0KFCli6dKmUxpSUFDx69Ah2dna4cuUKzpw5AycnJ+zfv/+lv6t1qV27Nrp27Yp33nmnTONv6dKlUhmkpKTk68uGhIQUGINFwfgzT5aGToCGi4sLfv31V3z00UfIyMiATCZDXFwcOnbsiF69eqF+/fqoWLFivh1oNJdY2NnZYfv27dIP0dwqVaqElStXol+/flAqlXBxcUFCQkKhZ5GcnZ1RoUIFrFq1CqNHj8atW7fg5+eHBg0awNvbW2pw09LS8l325O/vj6+++kq6vNDHxwc//vgjACA6Ohpff/01AgMDERYWVuxdPX766Sd88MEHqF+/PqysrNC5c2d89tln+Z4XHh6OmTNn4o033pB+ELZv3x4hISEYPHgwfv75Z7z//vuYOXMmXnvtNekHPZBzKYa1tTUqV66Mpk2b4o033oCXl1eB5dOxY0csX74coaGhOstnypQpCA8Px5w5czBu3DhkZ2dDpVKhWbNmiI2NhZWVFdq0aYOmTZtCJpMhNDQU69evx6xZs3DlyhXcunULq1evRtWqVaXOE5BT2a5atQpDhw7Fs2fPkJaWBi8vL62zirpERkYiLS0NNWrUgIWFBezs7KQtjRs2bAghBAIDA6VLAgoSHh6O3r17S7N/li1bhszMTPTs2RO9e/eWLh3RXLKU97XffvstPvroI9jY2EAul0sziF5WPnK5HOnp6XB1dUVERASWLl2KXbt2ac0GKonr169j8ODBOsunX79+CAwMhL29PZKSkrBy5UoMGzYM8+bNg4WFBRYtWqTzbJ+Gt7c3GjRogNdeew12dnZwdnaWzo7m/Q4FBQVh6dKlGDRoEO7cuYMePXogISEBjRs3LjT9gYGBqF+/Pvz9/VGzZs18P/oL+u58++230iVZS5YsgZWVFd566y18/fXXsLe3R/PmzfHaa6/h8ePH6N27N37//XdYW1vjl19+wRtvvIHXXnsNcrkca9askTrHuT8vTXq+++47LF68GC4uLqhSpYo0QOTi4gJ7e3sEBwcjKysLd+/exdWrV9GqVSv4+flh/vz5WLp0KbKystCgQQOty4TmzJmD8ePHo1u3brh16xaio6Mxc+ZM9O7dGz/88ANq166NatWq4ZtvvkGPHj2wZcsWZGdnY8+ePRg2bBiqVauG+Ph4WFtbQ61WS/VRkyZNsHjxYgQHB6NJkybS7JuffvoJUVFRAHJi2M/PT6tcFAoFfv75Z4wYMQJCCAwfPhwKhQLvvvsu3nzzTan+OXPmDIYOHQoLCwtcuHABz58/R2RkpNblNa6urtLnaGNjg5SUFPj7+8PCwgIVKlTAwoUL4e3tjXnz5qFHjx4ICAhAaGgo/P39teoJjYiICAwZMgTfffcdFAoFgoODERsbi1OnTqFu3bpa8b1q1SoMGzYMFy5cwMiRI4sd34Xx9/eX6kR7e3vpsiRNLGzatAljxozBRx99JA1A5t3avG3btoiPj4evry8qVqyIVq1a4ebNm1rt6OPHj/H48WNpZkWvXr2wYsUKZGZmIi0tDS9evECTJk3QokULREZG4sKFC9i3bx9Wr16NatWq5ZvdlrsdBXJ+sNjZ2aFNmzb44IMPsGnTJp35XbJkCd5++200b94cImcdVYSFhaFr165SOS9cuBDz58/Xep1cLkePHj2kWXFXrlyBTCbD/Pnz8cYbb+DEiRNo27YtPD09sWLFChw+fFia0QHk7AA7e/Zs6WztV199JV3m3bBhQyQlJUl1QdWqVeHj45Nv4O9l7eiZM2dQr149uLi4IDw8XGfHOncca+ILQL46MPcMN3t7ezx58gQA0LFjR60TL2VVT3fv3v2l/bFevXrhp59+QlBQELKyslCzZk20bt0aN2/exJdffol69eohNDQUx44dw5o1a9CzZ0/I5XKEh4dj8uTJGDp0KNRqNRwcHKBUKtGmTRvs27dP6tflnjVaqVIlrFq1CgMHDsSLFy+wbds2NGjQAGFhYVL6hw4ditOnT6Nz5864e/cu2rVrB1tbW4wcOTJff+ztt99GkyZNsHLlSjg7O6Nu3bro1asXLCws8Pnnn+PLL7/EqlWrpLPtuS+vsbe3R3p6Ov744w+0b98ehw4dwvfffw9vb2+dn2tpxlFJqVQqfPLJJ/jqq6/w9OlTuLu7a8VRSEgI7O3tMWjQIKSlpWHNmjUlbu9Lqjh99+L0jVu0aIEPP/wQ48ePx8OHD6XZkSUtiw0bNqBbt254+vQpLCwsUKdOHfTv3x/Tp0/X6hOPGjUK3bp1Q/369VGlShW0atUKdnZ2WLRoEUaMGCHNwLW1tZVO7v7xxx8YOHAgzp07h/r16+Phw4eoXLkyxo8fDxsbG1y6dAn29vaYP38+pk2bhu7du8PZ2RktW7YssIy+/fZbDBgwAPXq1ZMuxb979y5ef/11WFpaQq1WIyIiAtHR0dJALwDpkvgOHTqgYsWK0iXZQM7MxQkTJkgn3Pr27YvAwEAolcp8/bI//vgDXbp0wQcffICPP/4YNjY2Wv1te3t76RJBTf2nGQycMWMG+vbtizp16sDa2hrx8fE6L8WbN28edu/eDWtra8jlcq0BVX2aNGkS+vbti+XLl6NWrVr52s2CODs7IzY2Fl988QWysrKk1+nru1NU7u7uWLRoEbp06QKFQoHWrVvD3t4ezs7OqFixIkaNGoVffvkFX3zxBR4+fIibN29i1apV8PT0xMcff4x+/fpJyyX8/vvvJZopu2TJEnz++edlGn8ffvihdCKvdu3aGD9+vNbvkNyzZ3PH4Oeff44FCxbg/v37OHPmDN5//30cP35c66oSgPFntkpttTQTd/z4cdG3b18hhBDTpk3TWtDv119/FZ988okQQojBgwdr7axkKgrLX1pamvT3W2+9Je16ZcyePHkivLy8xOXLl7Xu79+/v7QAooY+yi/3QoylwdziL2/5mFr8TZ8+XXz11Vclem1pxWBpepX40yy4LISQFut++vRpGaS66PQdfwXVP4bUvn17sWLFilKJPxRhA55XVVZ1oFKpFEFBQWW+4K0x1oGlEcdlXf9p6p+srCzRvXt38eWXX+p83ssWcja3Njiv0ow/Y6oPDdn+1qtXT9y9e/elz8vdZv7666+ibt26BZZP//79xaBBgxh/JqQ4MWiI746u+BOCdaCGqcegscefKTGayzONTXBwMDw8PBAeHo6zZ8+ia9euGDp0KACgQ4cOuHbtGsLDw2FjYyMtTm9KCsvf2rVr0bhxY4SFhcHLy6vQafjGYMGCBahbty7ee+891KhRQ7q/Xbt22L59OwYPHoxly5aZVPmZU/zpKh9Ti7/x48djzJgxxX6dqcbgq8Tf+vXrERQUhMDAQAwdOhQ//fST0Z2x0mf8FVT/GEpSUhJq164NmUyGVatWmWT8AWVTB27cuBG1atVC06ZNy3zBW2OrA0sjjg1R/7Vq1QrBwcEICAiAo6MjRo4cWaLjmFMbrEtpxZ8x1YeGbn/Pnj2r8+qXvObOnYugoCD4+/tj5syZWLlyZaHlU69ePcafiShODBrqu6Mr/gDWgeYQg6YQf6bEQogCVog2QcOGDdN53fChQ4d0LvJpaPfu3ZMWAs7t33//lRbhzm3u3LkG3cmiU6dOWlNlgZxLFvbs2VOk15d1+bxqestb+ZQ1U0vvy7D+KV2mFi+mll4NxnHpMrW4KGl6GUelq7zEERXP5s2bMXHixHz3a9YuystYv49kmhh/ZM7MatCMiIiIiIiIiIhIH3h5JhERERERERERUR4cNCMiIiIiIiIiIsqDg2ZERERERERERER5WBo6AcZIrVbj1q1bcHBwgIWFhaGTQ6VMCIGMjAxUqVIFMplxjCMzBssPxh8ZEuOPDInxR4bGGCRDYvyRIRlj/BkrDprpcOvWLfj4+Bg6GVTGrl+/Dm9vb0MnAwBjsDxi/JEhMf7IkBh/ZGiMQTIkxh8ZkjHFn7HioJkODg4OAHICyNHRUbo/Ozsb27dvR0xMDKysrAyVvFJTXvOXnp4OHx8fqdyNga4YLK/lYy4Yf8bP3POoK3+mEn9A+Swfc8L4M27lNX+mEoPltXzMhanHH1B+y8hcmEobbKxMctDs0KFDmDBhAoCc0fD27dvj+PHjUKlUkMvleOedd9C3b1/cuXMH/fr1w9OnTzF8+HD06dOnSMfXTEV1dHTMV1nY2dnB0dHRbL9M5Tl/xjQFWVcMlvfyMXWMP+Nn7nksLH/GHn9A+S4fc6DP+CvNfiDjr3zmz9jrwPJePqbO1OMPYBmZOlPpAxork7x4tWnTpkhMTERiYiLCwsLQpUsXAMCWLVuQmJiIvn37AgBmzJiBsWPHYu/evfj+++/x4sULvby//5RtejkOUUkw/siQ/KdsQ/XxmwydDCIyIdXHb9Jr22XIfiDbYDIktsFkaKwDqTwyyZlmGllZWThy5AiWLFkCmUyGdu3awdnZGXPnzkW1atVw5MgRzJo1CzKZDCEhIThz5gxCQkLyHSczMxOZmZnS7fT0dAA5I7LZ2dnS/Zq/FTKhdb+50OTJHPMGFJw/c80vEVFpWr16NUaOHIn79+8jISEBc+bMga2tLZYvXw5vb29cuHABQ4YMgVKpRFxcHKKjow2dZDIz+ugHsg+Yg31AIiIi3Ux60Gznzp2Ijo6GTCZDQkICKlasiL179+KDDz7Axo0bkZ2dLe0E4eTkhEePHuk8zvTp0/HZZ5/lu3/79u2ws7PLd39ciBqbN2/Wb2aMyI4dOwydhFKVN3/Pnj0zUEqIiEyTSqVCQkICfHx8oFQqMXv2bOzduxdHjx5FXFwcFi5ciIkTJ2Lx4sXw8PBAbGwsB81I7/TRD2QfUBv7gERERNpMetAsISEBAwcOBABUrFgRABAZGYmPPvoIAGBlZQW1Wg2ZTIa0tDS4urrqPM6ECRMwevRo6bZmUbyYmJh813Lv2LEDnyTJcOzTtqWVLYPR5K9169Zmey23rvxpzioTEVHRrF69Gt27d8esWbOQnJwMPz8/WFtbo1mzZhgzZgyAnLWm6tSpAwBwdXXFgwcP4ObmZshkk5nRRz+QfcAc7AMSERHpZrKDZtnZ2Th69CgWL14MIKfRc3R0xLlz5+Di4gIACA0NRWJiIiIiInDs2DHMnDlT57EUCgUUCkW++62srHR2HDLVFmbZodAoKN/mIm/+zDmvRET6plKpsHbtWmzYsAGzZs1Camqq1uCCSqUCAKjVauk+zSwfXYNmxb08zlwvpzLn/CnkAgqZAACdZVoS+uoHsg+ojX3AouMl6kRE5YPJDprt3LkTLVu2lKbdt2zZEra2tgCA77//HgAwbtw49OvXD5MmTcKwYcOkx4mIiKhk4uPj0aNHD6n9dXZ21pqtIZfLAUB6HEChs72Le3lcebt8zBzMbPy/v3Pn71UujWM/kAyJl6gTEZUfJjtoFhsbi9jYWOl2UlJSvud4enqaZeeTiIjIUM6dO4fjx48jPj4eycnJmDt3Ls6fP4+srCwkJSUhMDAQQE4bfOnSJbi7uxc4ywwo/uVx5e3yMXPgP2UbFDKBuBC1Vv5e5dI49gPJkHiJOhFR+WGyg2ZERERU9mbMmCH9HRISgvnz52PNmjWIioqCjY0Nli9fDgCYNm0aBgwYAJVKpXMmmUZxL48rb5ePmYNMlYX0d+78mVs+qXwwxCXquXdvzX3bXJjz5ekAd28lMnUcNCMiIqIS0czu6dmzJ3r27Kn1WL169bBv3z5DJIuIqNQY8hL1uJCcgThz3cHV3GeGcvdWItPEQTMiIiIyGf5TtuHitA6GTgYRlVOGuEQ99+6tmWoLnJnSpvQzWobM+fJ0gLu3Epk6DpoREREREREVgSEvUc9UWyBTZb47uJrj5em56XP3ViIqOxw0IyIyQYmJiYiLi4NarcbIkSNRqVIljB07FjKZDPPnz0dAQADu3LmDfv364enTpxg+fDj69Olj6GQTERGZDV6iTkRk/jhoRkRkYp4/f45Zs2Zhy5YtsLa2BgBERkZi06ZNyMjIwLBhw7B582bMmDEDY8eORVRUFMLDw9GtWzfY2NgYOPVERERERESmgYNmREQm5tChQ7C1tUXHjh1hZ2eHH374AXK5HC4uLnBxccGjR48AAEeOHMGsWbMgk8kQEhKCM2fOICQkJN/xuHNXjvK4e5e55pWIiIiISB84aEZEZGLu3r2LlJQUHD58GDt37sTkyZO1tru3tLREVlYWsrOzpd27NNvd68Kdu7SVp927uHMXEREREVHBOGhGRGRinJ2d0axZM1hbWyM6OhqTJ0+Gg4OD9LhSqYS1tTWsrKygVqshk8kK3e6eO3flKI+7d3HnLiIiIiKignHQjIjIxISGhmLWrFkQQuDEiROoV68erly5gsePHyMjI0MaHAsNDUViYiIiIiJw7NgxzJw5U+fxuHOXtvK0e5c555OIiIiI6FXJDJ0AIlO1evVqVKpUCQCQkJCAsLAwREdH48aNGwCACxcuICIiAmFhYdi1a5chk0pmxs3NDW+88QYiIyMxduxYfPrpp/j888/Rrl079OrVC9OnTwcAjBs3DtOnT0dERASGDRsGW1tbA6eciIiIiF4Vf4cQlR3ONCMqAZVKhYSEBPj4+ECpVGL27NnYu3cvjh49iri4OCxcuBATJ07E4sWL4eHhgdjYWERHRxs62WRGRowYgREjRki3a9WqhYMHD2o9x9PT0+zX5yIiIiIqT/g7hKhscaYZUQmsXr0a3bt3h0wmQ3JyMvz8/GBtbY1mzZrh1KlTAIBbt26hTp06cHR0hKurKx48eGDgVBMRERERkSnj7xCissWZZkTFpFKpsHbtWmzYsAGzZs1Camqq1s6FKpUKAKBWq6X7NDsXurm56TxmZmYmMjMzpduaxbmzs7ORnZ0t/Q0ACpmQ/jYnefNpbgrKn7nml4iIiIj0S9+/Q4ryG0RzG+DvEFOlK3/mmtfSwEEzomKKj49Hjx49IJPlTNR0dnbW2oFOLpcDgPQ4gEJ3LgSA6dOn47PPPst3//bt22FnZ6d1X1yIGps3b36lPBgzc7+cMG/+nj17ZqCUEBEREZEp0ffvkOL8BgH4O8TU5c4ff4MUnUkOml29ehWhoaGoX78+gJzFDxMTEzFnzhzY2tpi+fLl8Pb2xoULFzBkyBAolUrExcXxWm7Si3PnzuH48eOIj49HcnIy5s6di/PnzyMrKwtJSUkIDAwEkLOe1KVLl+Du7l7oLDMAmDBhAkaPHi3dTk9Ph4+PD2JiYqSzR9nZ2dixYwc+SZLh2KdtSzeTBqDJX+vWrc1yR7+C8pe7o0NERC/HfiARlVf6/h1SlN8gAH+HmDpd+eNvkKIzyUEzAIiMjMS6desAgAsgUpmaMWOG9HdISAjmz5+PNWvWICoqCjY2Nli+fDkAYNq0aRgwYABUKpXOMzi5KRQKKBSKfPdbWVnlq7gz1RZmWZlr6MqzOcmbP3POKxFRaWE/kIjKI33/DinObxCAv0NMXe78mXM+9c1kB80OHDiA8PBwhIeHo2/fvloLII4ZMwbA/xZABCAtgFjYbB+i4kpKSgIA9OzZEz179tR6rF69eti3b58hkkVERGTW2A8kovKOv0OIyoZJDpp5enoiJSUFdnZ2GDx4MH755RcugKgH5XEBRF23iYiIyHjpsx/IPmAO9gGJiIh0M8lBs9zTSN98800sW7YM9vb20uNcAPHVlKcFEAEugkhERGRK9NkPZB9QG/uARERE2kxy0CwjIwMODg4AgH379qF9+/ZYsGABF0B8ReVxAUSAiyASERGZEn32A9kHzME+IBERkW4mOWi2f/9+TJo0CXZ2dqhRowbi4uJgY2PDBRD1pDwtgKi5TURERKZBn/1A9gG1sQ9IRESkzSQHzWJjYxEbG6t1HxdAJCIiIjJ/7AcSERFRWZG9/ClERGSMVq9ejUqVKgEAEhISEBYWhujoaNy4cQMAcOHCBURERCAsLAy7du0yZFKJiIiIiIhMjknONCMiKu9UKhUSEhLg4+MDpVKJ2bNnY+/evTh69Cji4uKwcOFCTJw4EYsXL4aHhwdiY2MRHR1t6GQTERERERGZDM40IyIyQatXr0b37t0hk8mQnJwMPz8/WFtbo1mzZjh16hQA4NatW6hTpw4cHR3h6uqKBw8eGDjVREREREREpoMzzYiITIxKpcLatWuxYcMGzJo1C6mpqVq7vKlUKgCAWq2W7nNycipw97jMzExkZmZKtzW7iWVnZyM7O1v6GwAUMqF125zkzau50ZU/c80rEREREZE+cNCMiMjExMfHo0ePHpDJciYLOzs7SwNdACCXywFAehwA0tLS4OrqqvN406dP17mz3Pbt22FnZ6d1X1xIzkDc5s2bXy0TRmzHjh2GTkKpyp2/Z8+eGTAlRERERETGjYNmREQm5ty5czh+/Dji4+ORnJyMuXPn4vz588jKykJSUhICAwMBAJ6enrh06RLc3d0LnGUGABMmTMDo0aOl2+np6fDx8UFMTIw0gy07Oxs7duzAJ0kyZKotcGZKm9LPaBnT5LF169awsrIydHL0Tlf+cg+2FtWRI0cwatQoWFlZwcvLCytWrMCGDRswZ84c2NraYvny5fD29saFCxcwZMgQKJVKxMXFcU09IiIiIjI5HDQjIjIxM2bMkP4OCQnB/PnzsWbNGkRFRcHGxgbLly8HAEybNg0DBgyASqXSOZNMQ6FQQKFQ5Lvfysoq3+BRptoCmSoLsxxU0tCVb3OSO38lyaePjw92794NW1tbTJgwAb/99hs3oiAiIiIis8SNAIiITFhSUhIAoGfPnjh48CB2794NHx8fAEC9evWwb98+HDx4EK1btzZkMsmMeHp6wtbWFgBgbW2NixcvciMKIio3jhw5gqZNmyIiIgJvvfUWsrOzkZCQgLCwMERHR+PGjRsAgAsXLiAiIgJhYWHYtWuXgVNNREQlxZlmREREVGz//vsvtm/fji+//BL379+X7i+NjSg0t4GczSjMcQMDc96IQiEXOjcRMce8kvnjbFsiovKFg2ZERERULOnp6ejbty+WLVsGlUpVZhtRADmbUXAjCtMys/H//uZGFGTqPD09pb91zbYdM2YMgP/NtgUgzbblDta6mfNJA6Dg/JlrfonMDQfNiIiIqMiUSiV69eqFyZMnw9fXF9nZ2aW+EQWgvRnFsU/blm4mDcCcN6Lwn7INCplAXIj6lTeiIDIW+pptyx2s/8ccTxrkljd/PHFAZBo4aEZEVMqqj98EhVxozbYgMlWrV6/GX3/9hbi4OMTFxWH48OH48MMPy2QjCiBnMwpzG1TKzRw3oshUWUh/v+pGFETGQJ+zbbmDtXmfNAAKzh9PHBCZBg6aERERUZH17dsXffv2zXd/z549tW5rNqIgIjIn+p5tyx2s/8ccTxrkljd/5pxXInPC3TOJSoA7JxERERGVP7ln20ZFReGXX36RZttOmjQJkyZNAvC/2bZt2rTBlClTDJtoMhv8DUJU9jjTjKgEuHMSERERUfnD2bZUHPpeooO/QYjKHmeaEZWAp6cnbG1tAejeOenUqVMA/rdzkqOjo7RzEhERERERUXHxNwhR2TPJmWZHjhzBqFGjYGVlBS8vL6xYsQL16tWDl5cXAODjjz9G69atceHCBQwZMgRKpRJxcXEcYSe909fOScXdbtwct6g25+3GFXJR4Dbx5phfIqLSxH4gEZV3ZfkbRHMb4O8QU6Urf+aa19JgkoNmuqalOjk5ITExUet5nJZKpUmfOycVd7txc91qHDDP7cZzT8nnduNERK+G/UAi41R9/CYAwNUv2xs4JebNUL9BAP4OMXW588ffIEVnkoNmnp6e0t/W1taQyWR48uQJIiMj4eXlhXnz5sHV1VWalgpAmpbKEfaClccRdl23i0LfOycVd7vxY5+2LXaajZ05bzfuP2UbFDKBuBA1txsnInpF+uwHsg+Yg31AItNgiN8gAH+HmDpd+eNvkKIzyUEzDc201EmTJiEyMhIVK1bEihUrMHnyZMydO7fI01I5wq6tPI2wAyUbZc+9c1JcXByGDx8u7ZxkY2OD5cuXA/jfzkkqlUpnjGkUd7txc6zMNcxxu/FMlYX0N7cbJyLSD330A9kH1MY+IJFxM+RvEIC/Q0xd7vyZcz71zWQHzXJPS7WyskLFihUBAN26dcOiRYsAFH1aKkfYc5THEXagZKPs3DmJDEnXej4bNmzAnDlzYGtri+XLl8Pb25vr+RCR2dJXP5B9wBzsAxKZBv4GISp7JjlolndaalZWFoQQUCgU2LdvH2rXrg2g6NNSOcKurTyNsGtuE5kSbjdOROWZPvuB7ANqYx+QiIhIm0kOmumaljpz5kxUqFABCoUCS5YsAVD0aalERKYk73o+ebcbHzNmDADodV3H3Ov55L5tTsrjmj7mmlcyb+wHEhERvZxmcw6FXGhtTEbFY5KDZrqmpeadkgpwWioRmTd9bTde3N1bAXBNHxPGnZPI1LEfSERERGXFJAfNiIjKO31uN17c3Vsz1RY4M6VNaWTLoMrjmj5cz4eIiIiIqGAcNCMiMjH63m68uLu3Zqq4po8p485JRERERERFw0EzIiITo+/txomIiIiIiCg/DpoREZkYbjdORERERERU+mQvfwoREREREREREVH5wkEzIiIiIiIiIiKiPDhoRkRERERERERElAcHzYiIiIiIiIiIiPLgoBkRERERERERmazq4zeh+vhNhk4GmSEOmhEREREREREREeXBQTMiIiIiIiIiIqI8OGhGRERERERERESUBwfNiIiIiIiIiIiI8rA0dAKIiIio9GgWxVXIBWY2NnBiiIiIiIhMCGeaERERERERERER5cFBMyIiIiIiIiIiojzKxaDZuHHjEB4ejr59+yI7O9vQySEDqD5+E/ynbDPIezP+yJAYf2RojEEyJMYflfc+YPXxm6TL9Kn8MYYYJDJ1Zj9odvLkSdy8eRP79u1D3bp1sW7dOkMnicoRxh8ZEuOPDI0xSIbE+CNDYvyRoTEGifTD7DcCOHjwIGJiYgAAbdu2xdKlS/HWW29pPSczMxOZmZnS7bS0NADAo0ePtEbks7Oz8ezZM1hmy/Dw4cMySH3Z0uTv4cOHsLKyMnRy9MpS+RSWaoFnz9T58peRkQEAEELo/X2LEn9A0WKQ8We6zC3+VGoLxqAJsVQ+zflfRwyWZvwBbIOLw1zjDyi4DmT8GY/yGH9A+WiDNW1AWcRtSd+L8We+deDr03cBAP6aEF3ifBTVq8S6ucagIfuA5sTsB81SU1Ph6ekJAHBycsKjR4/yPWf69On47LPP8t1fo0aNAo/r9pX+0khlp3chj2VkZMDJyUmv71eU+AOKH4OMP9NkLvEHAG6z9JNGKlsFxWBpxB/ANpi0Mf7IkMp7G1yW7Tb7CPmVdfwBxlUHMv4Mr6zbYHNi9oNmzs7OSE9PB5Azcu7q6prvORMmTMDo0aOl22q1Go8ePULFihVhYWEh3Z+eng4fHx9cv34djo6OpZ/4MlZe8yeEQEZGBqpUqaL39yxK/AFFi8HyWj7mgvFn/Mw9j7ryV5rxB7ANLo7ymD/Gn/Eor/ljG2wcymv+WAcaj/KYv9KOP3Ni9oNmYWFhmD17Nvr164dt27ahWbNm+Z6jUCigUCi07nN2di7wmI6Ojmb5ZdIoj/krrdH1osQfULwYLI/lY04Yf8bP3POYN3+leXaRbXDxlbf8Mf6MS3nMH9tg41Ee88c60LiUt/xxhlnRmP1GAMHBwfDw8EB4eDjOnj2Lrl27GjpJVI4w/siQGH9kaIxBMiTGHxkS448MjTFIpB9mP9MMAL76iotPkOEw/siQGH9kaIxBMiTGHxkS448MjTFI9OrMfqaZPikUCkyePDnfFFZzwfwZN1NP/8swf8bN1NNfFOaeR1PPn6mn/2WYP+Nm6ul/GebPuJl6+l+G+TN+5pCHwjB/VBgLwT1GiYiIiIiIiIiItHCmGRERERERERERUR4cNDMwpVKJzz77DHXr1oW/vz+Cg4MxZMgQPH78GImJiQgODi7xsa9evYoFCxboLa0DBw7Ea6+9hqCgIDRr1gxHjx4t1usXLFhQouvqc79u2bJl6NKlS7GPAQDvvvsu9uzZAwDYsGEDDh8+XKLjmBtTi8HAwEAEBwcjNDQUu3btKtbrP/30U6xcubLY75v7dVOmTMGHH35Y7GMAQLt27XDx4kUAObF84cKFEh3HnJhS/Gns3r0bcrkc33zzTbFexzrQ+JhS/EVFRaFGjRoIDg5GcHAw5syZU6zXs/4zPqYUf0IITJkyBa+99hoCAgLQokWLYr2e8WecTCkGX3/9dan+8/f3h4WFBU6dOlXk17MNNj6mFH9HjhxBkyZN0KBBA/j5+WHmzJnFej3jz8QJMqh+/fqJDh06iEePHgkhhFCr1WLt2rXi0qVLYs+ePSIoKKjEx36V12dnZ+e777fffpPu//3330W1atVKnLaSWrp0qejcuXOxX6dUKrVu9+/fX8yZM0c/iTJxphSDqamp0t9///23cHFxESqVqoSpK5nJkyeLUaNGFes1KpUqXzojIyPFr7/+qr+EmShTij8hhHj8+LEIDQ0VHTp0MEgdwjpQv0wp/oyhzmD9p1+mFH/ffPONeOONN0RmZqYQQojbt2+XOG0lxfjTP1OKwdwSEhKEv79/iY79KtgG65cpxV9QUJD47bffhBBCPHz4UFSqVEmcPXu2xOkrCcaf4ZT7QbMNGzaIunXrisDAQDF27FhRsWJFceXKFSGEEO+++65wcXERrq6uwt/fX8ydO1cIkRN4MTExokKFCqJixYoiIiKiRF+a5ORkYWtrK+7fv6/z8dxf9itXrggnJyfpsYyMDKEZ83z27Jno0aOH8PPzE4GBgaJ169ZCCCF8fX2FjY2NCAoKEh07dhRCCPHPP/+Idu3aiZCQEFGpUiVRs2ZN0adPH5GVlSUAiE8//VT4+voKe3t7ERkZKVq2bKkzbffv3xeWlpY6K5V//vlHhIWFicDAQOHv7y8+/vhjIYR2Z2fp0qUiOjpa9OrVS/j5+YmmTZuKs2fPii5duoi6deuK1q1bi4yMDJ2v01QWt2/fFlFRUaJhw4aiXr16YsSIEVLH6PvvvxcODg5CLpeL2rVri4MHD0qdpI0bNwpra2thbW0t3NzcxI8//ijat28vVq5cKeVh27ZtonHjxoUVn94UFINjx44VjRo1ElWqVBGNGjUSAQEBYu7cuUKpVIqBAweK+vXri0qVKomAgACTjMHKlSsLOzs7ERoaKrKysoQQQgAQ/fr1E1ZWVsLb27vA+NOkraBBs0OHDomGDRuKoKAgUb9+ffHDDz8IIbQbicmTJ4vu3buLDh06iDp16oj27duL06dPi5iYGFGnTh3Rq1cv6dh5X6eJx1OnTolmzZqJBg0aCD8/PxEXFyelYdy4ccLFxUXI5XJRs2ZNcePGDVGtWjVx/PhxsXDhQmFpaSkUCoVwc3MTmzZtEv7+/uLAgQPS6xcuXCh69OhRYP71hfFXvPjr06eP+O233wrtcLAOLDq2wUVvg4v6Q5/1X9Ex/ooef15eXuLixYsvzRfjr3jYBpesD9i2bVu2wXrAOrDodWBwcLBYvny5EEKIa9euCS8vL50nDxh/5snSELPbjMW9e/cwaNAgHDhwAHXr1sXSpUvx8OFDAMDff/+NX375BQcPHsT69evh5eWF2bNn4/XXX8e5c+ewb98+/PXXX/juu+/g5eWFbt264ezZs7CwsCjy+//999+oU6cO3NzcXikfW7duxePHj3Hu3DkAwKNHjwDkTOf88MMPceLECQCASqXCW2+9hfj4eGRmZmL69Ok4d+4c7O3tsW7dOgCAXC7HggUL8Mcff+Drr78u8D2//fZbtGvXDpaW+UNo3rx56NChAyZMmKCVnryOHj2K06dPo2rVqujbty86duyIgwcPwsPDAx06dMDy5csxYsSIAtPg7OyM33//Hfb29lCpVOjcuTPWrl2LXr16QaFQQKlUomPHjoiLi4O/v7/0OrVajbp162LgwIE4e/Ys6tevj2rVqmHy5Mno3bs3AOD777/H+++/X+B760tBMXju3Dlcv34dANC9e3e8/vrr6Ny5M5o0aYIXL17AyckJt2/fRlRUFMaMGYPLly+bVAxOmjQJ69atw3//+1/UrFkTM2bMwKRJkwAAMpkMI0eOLDD+xo8fj4SEBKSmpmL9+vWQyfJfZT59+nSMGTMGb731FgAgNTVV57GSkpJw7NgxODs7IyoqCu+++y527NgBW1tbhISEYMuWLWjfvn2B+a5evTp27doFhUKB58+fIywsDK1atUKTJk1gZWUFa2trdO3aFZ988gm8vLyk13l4eKBKlSr49ttvsWnTJri4uGDkyJGYN28ewsLCAOTE4Lx58wr72F8Z46948bdu3TrIZDJ06tQJv/zyS4HpYR1YNGyDi98Gjx8/Hp988gnq1auH6dOno2bNmvmew/qvaBh/RY+/9PR03L17F7/99pv03NGjR6Nnz5750sP4Kzq2wcXvAwLA9evXsXfvXvz00086H2cbXDSsA4vXBi9duhSdO3fGpEmTcP/+fSxcuBCVK1fOlx7Gn3kq12uaHT58GIGBgahbty4AoH///rC2tgaQc61vRkYGevXqhRUrVmDMmDHIyMjAuXPnsHHjRvj6+iIgIABt27aFvb09bt26hZs3bxokH0FBQTh//jzee+89rFmzBlZWVjqfd/HiRZw9exa9evVChw4dcOjQIWRkZKBy5co4cOAAAGDQoEEAgPXr1yM8PBzffvttvuPEx8dj7dq1+O9//6vzfSIiIvDjjz/i448/xvbt2+Hs7KzzeU2bNkXVqlUBACEhIQgNDYWHhwcAIDQ0FMnJyYXmW61WY9y4cQgKCkKDBg2QlJQkVYxyuRxhYWFwcnLK97qDBw+iSpUqAIC2bdviwIEDaN26NdLS0nD8+HH8+++/OHLkCHr06FHo++tDQTH4999/IyAgAGfPnsWWLVvwwQcfICwsDBkZGdizZw/c3d0REBCA3r1748CBA3j77bdNKgY/+OAD7Nu3D2FhYZDL5UhMTJSe065du0Lj78svv8SlS5ewdu1ajB07FllZWfme06JFC8TFxWHq1KnYv38/XFxcdKYnJiYGLi4usLCwQMOGDREVFQUHBwdYWlqiQYMGL43B58+f491330VAQACaNGmCf//9VysGO3ToAFtb23yvO3jwIFxdXQH8Lwb79OmDPXv24O7du9i/fz8sLCwQHh5e6Pu/KsZf0ePvzp07+Pzzz3XGZF6sA4uGbXDx2uCffvoJFy5cwKlTpxAeHo4OHTrofB/Wf0XD+Ct6/CmVSiiVSjx//hx//fUX1qxZg//85z84efJkvvdh/BUd2+Di9wGBnHWdOnToUOBgC9vgomEdWLw2+Msvv8T06dNx7do1nD17Fh9//LE0UJcb4888letBs8KkpaXBwcEBJ06cwKZNmxATE4MrV66gf//+ePLkCeRyOQDAycmpwBHkl2nYsCGSk5OlUf3CWFpaQqVSSbdfvHgh/V2zZk2cO3dOCnx/f3+dZ/aEEHB1dcWJEycwfPhwfPvtt7hy5QrefvttKQ/29vYICQnBxYsXsWvXLmzduhXHjh2TjrFmzRp89tln2LFjh/TFzqtr1644cOAAfH19pdF2XWxsbKS/5XJ5vttKpbLQz2T27Nm4d+8e/vrrL5w6dQq9e/fW+lzs7e11vi41NVWqUHOX38iRIzF37lwsWLAAgwYNgkKhKPT9S1NaWhoqVKgAV1dXKf5OnDiBK1euwMvLC3Z2dvnSXxKGikFN/J04cQJ79+6Fu7u79JymTZsWGH+5tWrVChkZGTh9+nS+xz788ENs2rQJnp6emDhxIt577z2dx3jVGJw4cSLc3Nxw/PhxnDx5ElFRUUWOwbx1iK2tLQYMGICFCxfi+++/L/TsUmlj/OWPv2PHjuH27dsIDg5G9erVsW7dOkydOhUff/xxvvdhHfjq2Abnj0EfHx8AgIWFBd5//31cvnxZZ9pZ/706xp92/Lm6usLe3h59+vQBkDPLq6ANoRh/r45tcMF9QCEEli5dinfeeafAtLINfnWsA7Vj8MGDB/j111+lmVg1a9ZEkyZNpMG23Bh/5qlcD5o1adIEp06dknbTiY+Pl2at1K5dGwqFAkuXLkVaWhpcXV2RkpKCR48eoU6dOkhOTsaZM2eQlpaGa9euwcvLS2v6d1HUrl0bXbt2xTvvvIPHjx8DyPlCr1+/HpcvX9Z6buXKlSGEkEa0V6xYIT1248YNWFhYoFOnTvj6668hhMD169fh6OiItLQ06Xm+vr5wdHTE0qVL4ezsjPT0dKSkpODatWvSWTcg50tmbW0Na2trdOzYUTqTuHbtWkyaNAk7d+6URsZ1SU5OhoeHB/r164eZM2eW2u4cqampqFy5MmxsbHDnzh0kJCQU6XXOzs6wsrJCWlqaVLYA0LdvX2zbtg1Lly7FsGHDSiXNeRUUg46OjnBwcICjoyN++uknKY0pKSlQKBSoUaMGTp8+jdOnT8PV1RU///yzScXg+fPnkZ6eDgA4c+aM1AEECo6/7OxspKSkSM87cuQI7t27p/PypIsXL6JGjRoYPHgwJk6cWKox6O3tDUtLS1y8eBE7duwo0uucnZ2hUCjyxeCIESPw3//+F7t378bbb79dKmnOjfFX9Phr37497t69i6tXr+Lq1avo1q0bPv30U0ybNi1fvlgHFg3b4KK3wUqlEnfv3pWes379enh4eKBixYr58sX6r2gYf8XrA7711lvYunUrgJzLjY4cOYLAwMB8+WL8FR3b4KK3wRq7d++GUqlE69atC8wX2+CiYR1Y9DrQxcUFFSpUwO7duwEADx48wF9//aV12aMG4888les1zdzd3bFo0SJ06dIFCoUCrVu3hr29PZydnREeHo7du3fjl19+wYQJE2BpaYnTp09j1apVaNWqFR49eoR+/frh33//RdWqVZGQkFCs67g1lixZgs8//xyvv/46LC0toVarERERgejoaFy7dk16nqWlJebOnYsOHTqgYsWK6Natm/TY6dOnMWHCBAghoFQq0bdvXwQGBkKpVKJ+/frw9/dHzZo1sXHjRvzxxx/48MMPcfHiRdy7dw9BQUFo0qQJmjVrhu+//x5AztoVjo6OAID9+/dLX5y3334blStXRufOnaX33rVrV75O+7p16xAfHw9ra2uo1Wq9bveb26hRo9CtWzfUr18fVapUQatWrYr0urCwMDx69Ahr167F3LlzMXjwYACAnZ0d3nzzTdy6dUs6o1/aCorBqKgo/Pjjj/jjjz/Qvn17PHv2DHv27IGbmxv69euHpKQkrFy5En369IGzszO8vLxMKgYHDRqEUaNGYcaMGXj27BnGjBkjHSs9PV2aypw7/rKzs9G/f3+kpaXB0tISFSpUwLp163Re+jFv3jzs3r0b1tbWkMvlmDVrVrE/l6KYNGkS+vbti+XLl6NWrVpo2bJlkV4XFhaGa9eu4YsvvsCDBw8wceJEAIC3tzcaNGiA1157TasTWVoYf0WPv+JgHVg0bIOL3gZnZmaiffv2yMzMhEwmg5ubGzZu3KgzT6z/iobxV7w+4PTp0zFw4ED88MMPAIBx48ahcePG+fLE+Cs6tsHFb4MXL16MgQMH6lzPVoNtcNGwDix6HSiXy7F27Vr83//9H5RKJbKzs/Hhhx+iadOm+fLE+DNTpbvPgPFLT0+X/v71119F3bp1pdtjxowRzZs3F7179xaZmZliyJAhQoicbWj79+8vmjdvLj744IMyT7O+FJS/H3/8UYSGhoqmTZuKMWPGGDiVJRcbGys8PT1FkyZNxNKlS19afkqlUgQFBYk///yzTNNZUAwy/kw7/oQofgw+efJEeHl5icuXL5dZGhl/jD8NQ9SBbIPNNwZNuf4TgvHH+CsbbIPNM/6EYBts7Mw9Bk0h/kxJuR80mzZtmggMDBT169cXTZs2FceOHTN0ksgAfvvtN1GtWjUxbNiwMn9vxiAJIcT8+fOFt7e3mDZtWpm+L+OPhDBcHcj4IyFY/5FhGSr+hGAMUg62wWRIhvwdbCoshBDC0LPdzMXmzZulad65paWl6dy94tChQzp39iEqKcYgGRLjjwyJ8UeGxPgjQ2MMkiEx/siccdCMiIiIiIiIiIgoj3K9eyYREREREREREZEuHDQjIiIiIiIiIiLKw9LQCTBGarUat27dgoODQ4m2zyXTIoRARkYGqlSpUugW1mWJMVh+MP7IkBh/ZEiMPzI0xiAZEuOPDMkY489YcdBMh1u3bsHHx8fQyaAydv36dXh7exs6GQAYg+UR448MifFHhsT4I0NjDJIhMf7IkIwp/owVB810cHBwAJATQI6OjtL92dnZ2L59O2JiYmBlZWWo5JUqc8+jrvylp6fDx8dHKndjoCsGjaVsmA79poHxZ5zKS35NJf6A8lEm5S2Pz58/Z/wZifKaP1OpA8tr+ZgLU48/oPyWkTnJm0djjD9jxUEzHTRTUR0dHfNVFnZ2dnB0dDTrL5M557Gw/BnTFGRdMWgsZcN0lE4aiht/iYmJiIuLg1qtxsiRI1GpUiWMHTsWMpkM8+fPR0BAAO7cuYN+/frh6dOnGD58OPr06VOstBhj/JWV8pZfY6//gPJRJuU1j4w/wyvv+TP2GCzv5WPqTD3+AJaROSgoj8YUf8aKg2ZERCbm+fPnmDVrFrZs2QJra2sAQGRkJDZt2oSMjAwMGzYMmzdvxowZMzB27FhERUUhPDwc3bp1g42NjYFTT0REREREZBq44lsJ+E/ZZugkUDnG+KNDhw7B1tYWHTt2xBtvvIHbt29DLpfDxcUFVatWxaNHjwAAR44cQcuWLWFpaYmQkBCcOXPmld/bf8o2VB+/6ZWPQ1RSrAPJkBh/ZEhsg8nQWAdSecSZZkQltHr1aowcORL3799HQkIC5syZA1tbWyxfvhze3t64cOEChgwZAqVSibi4OERHRxs6yWQm7t69i5SUFBw+fBg7d+7E5MmTtabQW1paIisrC9nZ2dJuOE5OTtJgWl6ZmZnIzMyUbqenpwPImcadnZ0t/Q0ACpnQum2u8ubbXJl7/oiIiIiIXgUHzYhKQKVSISEhAT4+PlAqlZg9ezb27t2Lo0ePIi4uDgsXLsTEiROxePFieHh4IDY2loNmpDfOzs5o1qwZrK2tER0djcmTJ2st4qlUKmFtbQ0rKyuo1WrIZDKkpaXB1dVV5/GmT5+Ozz77LN/927dvh52dndZ9cSFqAMDmzZv1mCPjtWPHDkMnoVQ9e/bM0EkgIiIiIjJaJjtoVpqLYBO9zOrVq9G9e3fMmjULycnJ8PPzg7W1NZo1a4YxY8YAyNmyuU6dOgAAV1dXPHjwAG5ubjqPV9yZPoacHWIsM3CMIR36SENJXhsaGopZs2ZBCIETJ06gXr16uHLlCh4/foyMjAxpcCw0NBSJiYmIiIjAsWPHMHPmTJ3HmzBhAkaPHi3d1uymExMTo7UI8Y4dO/BJkgyZagucmdKmBLk1HZr8tm7d2mwXhAX+V9cQEREREVF+JjloxkWwyZBUKhXWrl2LDRs2YNasWUhNTdW6NE6lUgEA1Gq1dJ/m0riCBs2KO9PHGGb5GMsMHGNIx6ukoSQzfdzc3PDGG28gMjISFhYWWLJkCW7evIl27drBwsICP/zwAwBg3Lhx6NevHyZNmoRhw4bB1tZW5/EUCgUUCkW++62srPINGGWqLZCpsjDrgaTcdH0G5sSc80ZERERE9KpMctAs9yLYdnZ2+OGHH6RFsF1cXLQWwZ41axZkMpm0CHZISEi+4xVllo/mNmD4mT6lyRhm75QmXfkrbl7j4+PRo0cPaa0oZ2dnrdkacrkcAKTHARR6aRxQ/Jk+xz5tW6w065OxzMAxhnToIw0lnekzYsQIjBgxQrpdq1YtHDx4UOs5np6eRjGoSEREROZPs0nB1S/bGzglRET6Y5KDZvpeBLs4s3wA45npU5rM/Yd27vwVd6bPuXPncPz4ccTHxyM5ORlz587F+fPnkZWVhaSkJAQGBgLIGbC4dOkS3N3dC51lBhR/po8xzA4xlhk4xpCOV0mDodNOREREREREupnkoJm+F8EuyiwfwHhm+pQmY5i9U5p05a+4M31mzJgh/R0SEoL58+djzZo1iIqKgo2NDZYvXw4AmDZtGgYMGACVSqVzUJaIiIiIiIiIjJdJDprpexHs4szyAYxnpk9pMobZO6Upd/5eJZ9JSUkAgJ49e6Jnz55aj9WrVw/79u0reSKJiIwUN+MhQ1u9ejVGjhyJ+/fvIyEhAXPmzIGtrS2WL18Ob29vXLhwAUOGDIFSqURcXBx3sCYiIqISMclBM30vgk1ERERFw814yNBUKhUSEhLg4+MDpVKJ2bNnY+/evTh69Cji4uKwcOFCTJw4EYsXL4aHhwdiY2M5aEZ6c/fuXbzxxhuwsrKCXC7HypUrcenSJZ44ICIyUyY5aAZwEWwiIiJD4GY8pcfcN+MBtPNY0nyuXr0a3bt3x6xZs5CcnAw/Pz9YW1ujWbNmGDNmDADg1q1bqFOnDgDA1dUVDx48KHRtUaKicnNzw/79+yGTybBs2TIsXrwYO3bs4IkDIiIzZbKDZkRERFT2uBlP6SsPJ/x27NhR7I14gJxZZmvXrsWGDRswa9YspKamasWfSqUCAKjVauk+TfzpGjTjoG0Ocx+wLSh/JcmvZpd0AMjIyECtWrWwe/fuEp84ICIi48ZBMyIiIioybsZTesx9Mx5AO4/Pnz8v9uvj4+PRo0cPaUDW2dlZa0MfzYCG5nEAhcYfB221mfuAbd78lWTgFgBOnDiBoUOH4vHjx9i+fTvWrFkjPVbcEwdFGbjNPWib+7axUchLlj4O2hKRMeOgGRERERUZN+Mpfea+GQ+Qk0elUlns1507dw7Hjx9HfHw8kpOTMXfuXJw/fx5ZWVlISkpCYGAggJwlOi5dugR3d/cCZ5kBHLTVMPcB24LyV9wd1DWCg4Px119/Ye3atZg2bZrWcYp74qA4A7dxITkzKI114HZm45z/S5o+DtoSkTHioBkREREVGTfjIUOaMWOG9HdISAjmz5+PNWvWICoqCjY2Nli+fDkAYNq0aRgwYABUKpXOAQkNDtpqM/cB27z5K0les7KypE1QnJycYG9vD6VSWeITB0UZuM09aJuptsCZKW2Kne6y4D9lGwAUO30ctC0+7iBMVHY4aEZERETFws14yBgkJSUBAHr27ImePXtqPVavXj3s27fPEMkiM3fixAmMGTMGcrkcNjY2WLJkCZKTk0t84qA4A7eZagtkqox34DZTZQGgZIORmtcZa970QR+DtgB3ECYqaxw0IyIiIiIiKoLGjRvjzz//1LrP09OTJw6ozHAHYaKyxUEzIiIiIiIiIiPHHYRLh7lvRgHkz6M551XfOGhGREREREREZOS4g3DpKg+zQzV55EYURcdBMyIiIiIiIiIjxx2ES4e5b0YB5M/jq2xEUd5w0IyIiIiIiIjIyHEH4dJl7ptRAP/Lo7nnU584aEZERERERERkQriDMFHZkL38KUREZIxWr16NSpUqAQASEhIQFhaG6Oho3LhxAwBw4cIFREREICwsDLt27TJkUomI6P+1d+/BUdXnH8c/AZIAhiSGmwFjUUHCPVgukhBAgRgEGSlowE7QgYqlY5GhdBDsT6QZi5dCy6AVr4DDlEIQEIRCEBsJFxEQwiWYRiqCUMolIQmCy2azvz8yWZPNxWyym7Pn7Pv1D9lzdk+eZ8/DLjzn+/0eAABgOow0A0yo03NbXD+ffnm0gZHAKA6HQ+np6YqJiVFJSYkWL16szz77TAcOHFBaWpreeustzZs3T++9957at2+vUaNGafjw4UaHDQAAAACmQdMMAExo9erVevTRR7Vo0SLl5eWpW7duCgkJUUJCgmbPni1JOn/+vLp06SJJioqK0uXLl+t9u/GKtxqv+NiqAuV23FbPDwAAAGgIUzfNVq9erRkzZujSpUtKT0/XX/7yF7Vo0UIrV67U7bffrq+++krTpk1TSUmJ0tLSGGUBwBIcDofWrl2rjRs3atGiRSooKKh0hyOHwyFJKi0tdW2LiIio8e5JntxuPK1f2TGtfLvxiqx+63FuNw4AAADUzLRNM6YmAQhUq1at0mOPPaYmTcqWpYyMjKx02+imTZtKkmu/JBUWFioqKqra49XlduMVbzVuKw3S8Rcf9Hpe/iQQbj0uiduNAwAAALUwbdOssacmlT+WyqYnWXVKi9WnJFWXn1VzhXXl5OTo8OHDWrVqlfLy8rR06VKdPHlSN2/e1MGDB9W7d29JUnR0tE6dOqV27drVOMpM8ux247bSINkc1r7deEVWvyW3lXMDAAAAGsqUTTMjpyZJZdOTrD41yepTkirmx/QkmM0rr7zi+rlfv3568803tWbNGg0bNkzNmzfXypUrJUkvvfSSnnzySTkcjmo/44CGYIkEAAAAWJ0pm2ZGTE2SKk9POvRCsldz8hdWn5JUXX5MT4KZHTx4UJKUkpKilJSUSvu6d++urKwsI8KCxbFEAoBA9cUXX+jZZ59VcHCwOnbsqA8++EAbN27kwgEAWJQpm2ZGTk2SyqYnWbGhVFEgTEkqz8/KeQKAL3hziQQAMJOYmBh9+umnatGihebOnauPPvqICwcAYGGmbJoxNQkAAGN4e4kE1hX9kdXXFZUq52jlPGFd0dHRrp9DQkKUm5vLhQMAsDBTNs0qYmoSAACNx9tLJLCuaFVWX1dUKsuRNUVhZt9++60yMjL08ssv69KlS67tvrhwUPGiQcXH/ia0af3is/oFg5rys2q+gNWYvmkGAAAaj7eXSGBd0R9ZfV1RqXKON27cMDocoF6KioqUmpqqFStWyOFwNNqFg7R+ZY04f71w8OqAsj/rG5/VLxi458eFA8AcaJoBAIA68/YSCawrWpXV1xWVynIsKSkxOgzAYyUlJZo4caLmz5+vrl27ym63+/zCQcWLBrbSIB1/8UHfJ1oPPV/cLkkex2f1CwY15cfNyABzoGkG1AN3TgIAlkgAEHhWr16t/fv3Ky0tTWlpaZo+fbpmzpzZKBcObKVBsjn898KBzREkqf432bL6BQP3/KycK2AlNM2AeuDOSQAAAIEnNTVVqampVbZz4QAArImmGVAP3r5zUn0WgS3X2IuI+stirf4QhzdiMPp9BAAAAABUj6YZ0ADeunNSfRaBLWfUYrD+slirP8TRkBhYBBYAAAAA/BNNM6CevHnnpPosAluusReD9ZfFWv0hDm/EwCKwAOAZ1hUFAACNhaYZUA/evnNSfRaBrfgcI/jLYq3+EEdDYjA6dgAwG9YVBQAAjYWmGVAP3r5zEgAAqBtvritalzVFyx9LZeuKWnEtSn9YJ9SXasrPqvnCuhhpCzQ+mmZAPXDnJAAAjOWNdUU9WVNUKltX1Ki1RBuDP6wT6kvu+bGuKMyGkbZA46NpBgAAAFPx1rqidVlTVKq8ruihF5J9kVKD9Xxxu+tnT9c79Yd1Qn2ppvxYVxRm482RtgDqhqYZAAAATMOb64p6sqaoVLauqL82lbyx3qk/rBPqS+75WTlXWJs3RtpKTFEvZ/Up6lLVHK2cq7fRNAMAk2E9CwCBjHVFAQQyb420lZii7s7qU9SlH3Nkenrd0TQDAJNhPQsAgYx1RQEEKm+OtJWsOUXd0+npkvWnqEtVc2R6et2ZsmnGKAsAgYz1LAAAAAKPt0faWnGKekPis/oUdenHHK2epzeZsmnGKAsAaNz1LCquZVHxsVUFynoP9cmPC1cAABiDkbZA4zNl08zboyxYAPFHVv+PYnX5WTVXWJtR61mk9StrxFl5PYuKrL62RX3Ws+DCFYBAVlhYqJEjRyonJ0eff/65evbsqfT0dC4cAIBFmbJpVs5boyxYALEqq/9HsWJ+LIIIszFiPYuKa1nYSoPqtV6EmQTC2haS6rWeBdODAQSyli1basuWLfr9738vqew7mQsHAGBdpm2aeXOUhZUWQGwoq/9Hsbr8WAQRZmPkeha20iDZHP67noW3WX3Nh4bkxu3uvc/qo72lyjlaOU9YV3BwsNq2bet6nJeX5/MZL2ZZIiG0af3is/pnX035WTVfwGpM2TTz9igLKy2A6C2B8B/F8vysnCesifUsYDRud+9bVh/tLZXlyEhvWEFBQUGli+y+nPHi70skvDqg7M/6xmf1zz73/PgMBMzBlE0zb4+yAAAAdcPt7n3H6qO9pco53rhxw+hwgAaLjIz0+YwXsyyR0PPF7ZLkcXxW/+yrKT9muwDmYMqmGaMsAAAwBre79z2rj/aWynIsKSkxOgygwbp06dJoM178fYkEmyNIUv1ncVj9s889PyvnCliJKZtmAADAGFy4AmAmnZ7botCmTtfUQW946KGHdOTIEeXm5urpp59mxgsAWBhNMwAAAACoo+rW7OLCAQBYU5OffgoAAAAAAAAQWGiaAQA81um5Ler03BajwwAAAAAAn6FpBgAAAAAAALihaQYAAAAAFsTIcABoGJpmAAAAAAAAgBuaZgAAAAAAAIAbmmYICAxNBwAAAAAAnqBpBgAAAAAAALihaQYAAAAAAAC4aWZ0AAAAAAAAWE2n57YotKlTrw4wOhIA9UXTDAB+QsX18E6/PLrer6/PawEAAACr4N/FMBuaZgAAWFhDm74AAAAwn4r/BsxLSzIwEnMLiDXN5syZo8TERKWmpsput3vlmNyNEXXli/oD6or6g9ECqQb5t0H1jHxfAqn+4H+oPxiNGgQazvJNs+zsbJ07d05ZWVmKjY3VunXrjA7Jp/gHu38JtPqDf6H+YDRqEEai/mAk6g9GowYB77D89My9e/cqKalsKGJycrKWL1+uSZMmVXqOzWaTzWZzPS4sLJQk5efnV+rI2+12Xb9+Xc3sTeQoDZIkXblyxdcpeKRZyfeSao9r4MKdkqT9c4dX2Vee45UrVxQcHOybIA1Q8X1xz6+4uFiS5HQ6vf5761J/Ut1qsLr6K8+potrOrzfUtUZqiqN8e0NjbMxaLa8fqfL7XdcYavt7abX68wVf13R1vyu0iVN/6Ftqic/CmupX8m39Sb77Du48e62kxqkJT9TlO7g2gfAdXDHHH374QZL56s+T8+ut77y6qO3v+k+xau1JZe9Ls1Knrl+v+plu1u9gT85vQz+XPFHf30X9meszsKHfwb6qyYYc16o16P69UDFHX9eflVi+aVZQUKDo6GhJUkREhPLz86s8Z+HChVqwYEGV7XfeeedPHr/NoobH6At1ictfY/el6FpyLi4uVkREhFd/X13qT/JNDfrL+a0tDn+J0RMNibm211qt/nyhsevl8cb9dY2ipvfQF/Un8R2MyqxSf21eq188Vv68NIPaPtPN9h1cn/NL/RmrsetP8v/vYF/VCfVXvZr+H+yr+rMSyzfNIiMjVVRUJKmscx4VFVXlOXPnztWsWbNcj0tLS5Wfn6/WrVsrKOjHERVFRUWKiYnR2bNnFR4e7vvgDWD1HKvLz+l0qri4WB06dPD676tL/Ul1q0F/OTfE4d0YqD//FCj5+rL+JL6DPRVoObZq1Yr68xOBmh/fwf4hUPPjO9h/WD0/qWqOvq4/K7F80yw+Pl6LFy/W5MmTtX37diUkJFR5TmhoqEJDQytti4yMrPGY4eHhlv3LVM7qObrn56vuel3qT/KsBv3l3BCH92Kg/vxXIOTry6uLfAfXTyDlSP35l0DMj+9g/xGI+fEZ6F+snp9UOUdGmNWN5W8EEBcXp/bt2ysxMVEnTpzQ+PHjjQ4JAYT6g5GoPxiNGoSRqD8YifqD0ahBwDssP9JMkl57rZ6LTwBeQP3BSNQfjEYNwkjUH4xE/cFo1CDQcJYfaeZNoaGhmj9/fpUhrFZi9RzNnJ+/xE4c/hVDYwmkXKXAy9cMAuGckKP/MmvcdUV+/s3s8f8U8vN/VsihNlbPTwqMHH0lyMk9RgEAAAAAAIBKGGkGAAAAAAAAuKFpZrCSkhItWLBAsbGx6tmzp+Li4jRt2jRdvXpVmZmZiouLq/exT58+rWXLlnkt1gMHDighIUF9+vRRXFycPv30U49ev2zZsnrNq6/4uhUrVuiRRx7x+BiS9Ktf/Ur/+te/JEkbN27U559/Xq/jAAAAAAAA6wuIGwH4s6lTpyo/P1/79u3TrbfeKqfTqXXr1ik/P7/Bxy5vmv3617/2+LUlJSVq1uzH8nA6nRo3bpxWrFihESNG6N///rdGjBih3NxctWjRok7HrE8cDXldRQ6HQ++++67r8caNGxUXF6f77ruvwccGAAAAAADWE/AjzT766CN169ZNffr00Zw5c9SmTRudPn1akvTUU08pKipKrVu3Vq9evfT6669LKmvAPPjggwoLC1ObNm00dOhQ5eTkePy7v/76a6Wnp2v58uW69dZbJUlBQUF69NFHddddd1V67unTpxUZGel6fO3aNQUFBUmSbty4oZSUFHXv3l19+vRRUlKSpLJmU25uruLi4jR27FhJUl5enkaPHq3+/furXbt2uvvuu5Wamiq73a6goCDNnz9fsbGxuvXWWzVs2DANHz5cknTlyhVdunRJI0aMkCTdc889ioyM1D//+c8qeeXl5blGpPXq1Ut/+MMfJEkvvviiZs6cKUmu5tukSZPUvXt3xcfHKycnR+PGjVO3bt2UlJSka9euVXldRRcuXND999+vn//85+rRo4eeeeYZlZaWuo6fmJioqKgoNWnSRKtXr9awYcO0ceNGbd26VZs2bdLzzz+vsLAwjRgxQmPGjNHf//5317EzMjI0cODAup9ML5kzZ44SExNd56Scw+HQlClTlJiYWOm9WLJkiRISEjR27FgVFRX5PI7Nmzdr4MCBGjx4sJ599lnX9latWmnYsGEaNmyYjh075vM4MjMzFRMTU6lGpcZ/PzZs2ODK+4477tCSJUskSV26dHFt37Fjh9fiaEye1qLZeVpr8L1AqEGr111hYaEGDBigsLAwHT9+vNI+fz+PVq8/K9eemeuuHPVn3vqTqEEzoAbNfw4bU0A3zS5evKgpU6Zow4YNys7OVmxsrK5cuSJJ+vLLL7V+/Xrt3btXs2bN0u9+9zu9/fbbOnDggFatWqWsrCzt27dP48aN0/33368JEybI03sqfPnll+rSpYvatGnToDy2bdumq1evKicnR9nZ2frHP/4hqWxaY9euXXXkyBFt2rRJDodDkyZN0qJFi/Tuu+/qgQce0C233KKwsDCtW7dOktS0aVMtW7ZMTz/9tDIzM7Vz505JUps2bRQdHa21a9dKKpuqmZub62owVvT6669rzJgxys7O1rFjxzRr1qxq4z5w4IBeeeUV5eTk6O6779bDDz+sZcuW6eTJkwoJCdHKlStrzTsyMlKbN2/WoUOHdPToUZ0+fdoVnyQdOnRIW7Zs0eTJkytNc33ooYcUFxenhIQEXbt2TXfeeadGjhzpaopK0htvvKFnnnnmp998L8rOzta5c+eUlZWl2NhY1zmRpI8//lgdOnRQVlaWvv/+e+3bt0+XL1/Wpk2btHv3bqWkpOiNN97weRx9+vTRnj17tHv3bl28eFEHDx6UJHXt2lWZmZnKzMxUr169fB6HJKWkpFSqUSPej3Hjxrnyvvvuu11ThyMiIlzbR44c6ZU4GpOntWh2ntYafC8QajAQ6q5ly5basmWLJkyYUGWfP59Hq9ef1WvPrHVXjvozd/1J1KC/owbNfw4bW0A3zT7//HP17t1bsbGxkqQnnnhCISEhksqm7xUXF2vixIn64IMPNHv2bBUXFysnJ0ebNm1S165d1atXLyUnJyssLEznz5/XuXPnDMmjT58+OnnypH7zm99ozZo1Cg4OrvZ5ubm5OnHihCZOnKgxY8Zo3759Ki4u1m233aY9e/ZIkqZMmSJJ+vDDD5WYmOgaOSOVjcp7//331bdvXy1ZskSDBw+uNIWz3JAhQ/TOO+/o+eefV0ZGRqURchUNGjRId9xxhySpX79+6t+/v9q3by9J6t+/v/Ly8mrNu7S0VHPmzFGfPn3Ut29fHTx4UEeOHHHtj4+P16BBg6p97aVLl9S1a1dJUnJysux2uwoLC3X48GF9++23+uKLL/TYY4/V+vu9be/eva5RgsnJya5zUtO+AwcOaOjQoQoKCqryfF/Fcccdd7jOeUhIiJo0KfsIOXXqlIYMGaLp06frhx9+8HkcUtUaNeL9KHfhwgXZbDb97Gc/k1Q2EnTo0KF6/PHHvTLVurF5Wotm52mtwfcCoQYDoe6Cg4PVtm3bavf583m0ev1ZvfbMWnflqD9z159EDfo7atD857CxBXTTrDaFhYVq1aqVjhw5oi1btigpKUnffPONnnjiCV27dk1NmzaVVDaipL7/Kb733nuVl5fnGt1Wm2bNmsnhcLgeV2xM3HXXXcrJyXEVfc+ePVVQUFDlGE6nU1FRUTpy5IimT5+uJUuW6JtvvtEvf/lLVw5hYWHq16+fcnNztXPnTm3btk2HDh2SVNac27Ztmw4fPqxVq1bp/Pnz6tGjR5XfM378eO3Zs0ddu3Z1jTqrTvPmzV0/N23atMrjkpKSWt+TxYsX6+LFi9q/f7+OHj2qxx9/vNL7EhYWVuNrbTab6/eVn8MZM2Zo6dKlWrZsmaZMmaLQ0NBaf7+3FRQUKDw8vFJMte2r7fm+iqPcgQMHdPHiRd17772SyqYa79q1S9HR0V4b4VVbHNXVqJHvx/r16zV+/HjX4z179uizzz5TcnKy5s+f75U4GpOntWh2ntYafC8QajDQ686fz6PV6y+Qa88M54/6s279SeY4h9QgNYjKArppdt999+no0aPKzc2VJK1atUo3b96UJHXu3FmhoaFavny5CgsLFRUVpa+//lr5+fnq0qWL8vLydPz4cRUWFurMmTPq2LGjOnbs6NHv79y5s8aPH6+pU6fq6tWrksoaWx9++KH+85//VHrubbfdJqfT6Vo77YMPPnDt++677xQUFKSxY8fqz3/+s5xOp86ePavw8HAVFha6nte1a1eFh4dr+fLlioyMVFFRkb7++mudOXNGUVFRrueFhYUpJCREISEhevjhh5WdnS1J+u9//+t6zjvvvKNbbrlFDzzwQJW88vLy1L59e02ePFmvvvqqz+5SWVBQoNtuu03NmzfXhQsXlJ6eXufXtmrVytWsLD+/qamp2r59u5YvX+6Vmw94qvycVIyptn21Pd9XcUhl9TZz5sxK02dbt24tSZowYYKrXnwZR3U1atT7IUnr1q2rNPzZF+9HY/K0Fs3O01qD7wVCDQZ63fnzebR6/QVy7Znh/FF/1q0/yRznkBqkBlFZQDfN2rVrp3fffVePPPKI4uLidOzYMYWFhSkyMlKJiYkaOHCg1q9fr9GjR2vjxo2aOnWqbty4oREjRmj06NGaPHmypk2bphMnTig9Pd21ML8n3n//ffXp00cDBw5Ujx491L17d2VkZFQp3mbNmmnp0qUaM2aM+vfvX2nBwmPHjrkW3u/bt69SU1PVu3dv9e7dWz169FDPnj01duxYNWvWTB9//LHWr1+vv/71r5oxY4amTp2q7du3KyEhwXW8iguo7969W507d5Ykvf3227rnnnvUpUsXbd68WRs2bKg253Xr1qlXr17q27evUlJStGzZMo/fl7p49tlntX//fvXo0UOpqamumxTUxcSJE7V582bXVNOEhAS1bNlSv/jFL5SQkKCYmBifxFyb+Ph4ffLJJ5JU5ZxUt69///7atWtXtc/3VRzlU5bfeusttWvXTpL0/fffu0ZBZmVluerFl3FUV6NGvB+S9L///a/S1MybN2/KZrNJ8u770Zg8rUWz87TW4HuBUIOBXnf+fB6tXn+BXHtmOH/UXxkr1p9kjnNIDZahBuHiDHBFRUWunzds2OCMjY11PZ49e7Zz8ODBzscff9xps9mc06ZNczqdTqfdbnc+8cQTzsGDBzt/+9vfNnrM3lJTfu+8846zf//+zkGDBjlnz55tcJQNM2rUKGd0dLTzvvvucy5fvrzWc1hSUuLs06ePc9euXYbF62nNLV682BkfH+8cPXq08+rVqz6P449//KOzQ4cOzqFDhzqHDh3qzMzMdB4+fNjZt29fZ2JionPs2LGNEkdNNdrY74fT6XT+7W9/cy5atMj1+MKFC857773XmZiY6BwxYoTzzJkzXoujMVn9889dIHwemk0g1GAg1J0n38P+xOr1Z/XaM2vdlaP+zF1/Tic16O+oQfOfw8YU5HR6eMtHi/nTn/6kNWvWyOFwKDw8XK+//rprnSYEjk2bNmnGjBkaNWqU3nzzTaPDAQAAAAAABgv4ppk3bd26VfPmzauyvbCwUBEREVW279u3Ty1atGiM0AAAAAAAAOABmmYAAAAAAACAm4C+EQAAAAAAAABQHZpmAAAAAAAAgJtmRgfgj0pLS3X+/Hm1atVKQUFBRocDH3M6nSouLlaHDh3UpAl9ZAAAAAAAQNOsWufPn1dMTIzRYaCRnT17VrfffrvRYQAAAAAAAD9A06warVq1klTWRAkPD3dtt9vtysjIUFJSkoKDg40Kz2esnp9UfY5FRUWKiYlxnXcAAAAAAACaZtUon5IZHh5epWnWsmVLhYeHW7KpZPX8pNpzZCouAAAAAAAoxwJOAAAAAAAAgBuaZvXQ88XtRocAAAAAAAAAH6JpBgAAAAAAALihaQYAAAAAAAC4oWkGAAAAAAAAuKFpBgAAAAAAALihaQYAAAAAAAC4oWkGAAAAAAAAuKFpBgAAAAAAALihaQYAAAAAAAC4oWkGAAAAAAAAuDF102z16tVq27atJCk9PV3x8fEaPny4vvvuO0nSV199pSFDhig+Pl47d+40MlQAAAAAAACYSDOjA6gvh8Oh9PR0xcTEqKSkRIsXL9Znn32mAwcOKC0tTW+99ZbmzZun9957T+3bt9eoUaM0fPhwo8MGAAAAAACACZi2abZ69Wo9+uijWrRokfLy8tStWzeFhIQoISFBs2fPliSdP39eXbp0kSRFRUXp8uXLatOmTZVj2Ww22Ww21+OioiJJkt1ul91ud20v/zm0ibPSdqsoz8mKuZWrLkcr5wsAAAAAAOrHlE0zh8OhtWvXauPGjVq0aJEKCgoUHh5eab8klZaWurZFREQoPz+/2qbZwoULtWDBgirbMzIy1LJlyyrb0/qVauvWrd5IxS/t2LHD6BB8rmKO169fNzASAAAAAADgj0zZNFu1apUee+wxNWlStiRbZGSka3SYJDVt2lSSXPslqbCwUFFRUdUeb+7cuZo1a5brcVFRkWJiYpSUlFSpGWe327Vjxw7938EmOvRCsldz8gfl+Y0cOVLBwcFGh+MT1eVYsXYAAAAAAAAkkzbNcnJydPjwYa1atUp5eXlaunSpTp48qZs3b+rgwYPq3bu3JCk6OlqnTp1Su3btahxlJkmhoaEKDQ2tsj04OLja5pGtNMiyTSWp5rytpGKOVs8VAAAAAAB4zpRNs1deecX1c79+/fTmm29qzZo1GjZsmJo3b66VK1dKkl566SU9+eSTcjgc1U6/BAAAAAAAAKpjyqZZRQcPHpQkpaSkKCUlpdK+7t27Kysry4iwAAAAAAAAYGJNfvopAAAAAAAAQGChaQYAAAAAAAC4oWkGAAAAAAAAuKFpBgAAAAAAALihaQYAAAAAAAC4oWkGAAAAAAAAuKFpBgAAAAAAALihaQYAAAAAAAC4oWkGAAAAAAAAuKFpBgAAAAAAALihaQYAAAAAAAC4oWkGAAAAAAAAuKFpBgAAAAAAALihaQYAAAAAAAC4oWkGAAAAAAAAuKFpBgAAAAAAALihaQYAAAAAAAC4oWkGAAAAAAAAuKFpBgAAAAAAALihaWYxnZ7bok7PbTE6DAAAAAAAAFOjaQYAAAAAAAC4MWXT7IsvvtCgQYM0ZMgQTZo0SXa7Xenp6YqPj9fw4cP13XffSZK++uorDRkyRPHx8dq5c6fBUQMAAAAAAMAsTNk0i4mJ0aeffqpdu3apU6dO+uijj7R48WJlZmbqj3/8o9LS0iRJ8+bN03vvvadt27bphRdeMDhqAAAAAAAAmIUpm2bR0dFq0aKFJCkkJES5ubnq1q2bQkJClJCQoKNHj0qSzp8/ry5duig8PFxRUVG6fPmykWEDAAAAAADAJJoZHUBDfPvtt8rIyNDLL7+sS5cuubY7HA5JUmlpqWtbRESE8vPz1aZNmyrHsdlsstlsrsdFRUWSJLvdLrvd7tpe/nNoE2el7f4ktKlTkuoVX/lr/DU3b6guRyvnCwAAAAAA6se0TbOioiKlpqZqxYoVcjgcrkaXJDVt2lSS1KTJjwPpCgsLFRUVVe2xFi5cqAULFlTZnpGRoZYtW1bZntavVFu3bm1oCj7x6oCyPxsS344dO7wUjf+qmOP169cNjAQAAAAAAPgjUzbNSkpKNHHiRM2fP19du3aV3W7XyZMndfPmTR08eFC9e/eWVDaN89SpU2rXrl2No8wkae7cuZo1a5brcVFRkWJiYpSUlKTw8HDXdrvdrh07duj/DjbRoReSfZtkPfV8cbsk6fiLD3r82vL8Ro4cqeDgYG+H5heqy7FiwxUAAAAAAEAyadNs9erV2r9/v9LS0pSWlqbp06dr5syZGjZsmJo3b66VK1dKkl566SU9+eSTcjgc1Y4kKxcaGqrQ0NAq24ODg6ttHtlKg/y2qWRzBElSg+KrKW8rqZij1XMFAAAAAACeM2XTLDU1VampqVW2p6SkVHrcvXt3ZWVlNVZYAAAAAAAAsAhT3j0TAAAAAAAA8CWaZgAAAAAAAIAbmmYAAAAAAACAG5pmAAAAAAAAgBuaZgAAAAAAAIAbmmYAAAAAAACAG5pmAAAAAAAAgBuaZgAAAAAAAIAbmmYAAAAAAACAG5pmAAAAAAAAgBuaZgAAAAAAAIAbmmYAAAAAAACAG5pmAAAAAAAAgBuaZgAAAAAAAIAbmmYAAAAAAACAG5pmAAAAAAAAgBuaZgAAAAAAAIAbmmYAAAAAAACAG5pmAAAAAAAAgBuaZo2s03Nb1Om5LUaHAQAAAAAAgFrQNAMAAAAAAADc0DQDAAAAAAAA3ARE02zOnDlKTExUamqq7Ha70eHAAEyLBQAAAAAAnrB80yw7O1vnzp1TVlaWYmNjtW7dOqNDAgAAAAAAgJ9rZnQAvrZ3714lJSVJkpKTk7V8+XJNmjSp0nNsNptsNpvrcWFhoSQpPz+/0sg0u92u69evq5m9ia5cuVKveJqVfC9J9X69L49fnt+VK1cUHBzs7dAMVfF9cc+xuLhYkuR0Og2LDwAAAAAA+BfLN80KCgoUHR0tSYqIiFB+fn6V5yxcuFALFiyosv3OO++s8bhtXmtYXG0WNez1Rh/frKJreV+Ki4sVERHReMEAAAAAAAC/ZfmmWWRkpIqKiiSVjSCLioqq8py5c+dq1qxZrselpaXKz89X69atFRQU5NpeVFSkmJgYnT17VuHh4b4PvpFZPT+p+hydTqeKi4vVoUMHg6MDAAAAAAD+wvJNs/j4eC1evFiTJ0/W9u3blZCQUOU5oaGhCg0NrbQtMjKyxmOGh4dbtqkkWT8/qWqOjDADAAAAAAAVWf5GAHFxcWrfvr0SExN14sQJjR8/3uiQAAAAAAAA4OcsP9JMkl57rYELkAEAAAAAACCgWH6kmTeFhoZq/vz5VaZyWoXV85MCI0cAAAAAANBwQU6n02l0EAAAAAAAAIA/YaQZAAAAAAAA4IamGQAAAAAAAOCGphkAAAAAAADghqZZLebMmaPExESlpqbKbre7tjscDk2ZMkWJiYmaOXOmcQE2UE35ZWZmKiYmRsOGDdPw4cMNjLBhCgsLNWDAAIWFhen48eOV9lnlHAIAAAAAAN+gaVaD7OxsnTt3TllZWYqNjdW6detc+z7++GN16NBBWVlZ+v7777Vv3z4DI62f2vKTpJSUFGVmZmrnzp0GRdhwLVu21JYtWzRhwoQq+6xwDgEAAAAAgO/QNKvB3r17lZSUJElKTk7Wnj176rTPLH4qhw8//FCJiYlasmSJEeF5RXBwsNq2bVvtPiucQwAAAAAA4Ds0zWpQUFCg8PBwSVJERITy8/PrtM8sasuhX79+ys3N1c6dO7Vt2zYdOnTIqDB9xgrnEAAAAAAA+A5NsxpERkaqqKhIUtnaWFFRUXXaZxa15RAWFqaQkBCFhITo4YcfVnZ2tlFh+owVziEAAAAAAPAdmmY1iI+P1yeffCJJ2r59uxISEuq0zyxqy6G8mSRJu3fvVufOnRs9Pl+zwjkEAAAAAAC+Q9OsBnFxcWrfvr0SExN14sQJjR8/Xk8//bQkacyYMTpz5owSExPVvHlzDRo0yOBoPVdbfmvXrtWAAQMUHx+vjh07asiQIQZHW38PPfSQMjIy9NRTT2nFihWWOocAAAAAAMB3gpxOp9PoIAAAAAAAAAB/wkgzAAAAAAAAwA1NMwAAAAAAAMANTTMAAAAAAADADU0zAAAAAAAAwA1NMwAAAAAAAMANTTMAAAAAAADADU0zAAAAAAAAwA1NMwAAAAAAAMANTTMAAAAAAADADU0zAAAAAAAAwA1NMwAAAAAAAMDN/wMjugpVVADvmAAAAABJRU5ErkJggg==\n"
          },
          "metadata": {}
        }
      ],
      "source": [
        "plt.rc('font', size=8)\n",
        "plt.rc('axes', labelsize=8, titlesize=8)\n",
        "plt.rc('legend', fontsize=8)\n",
        "plt.rc('xtick', labelsize=6)\n",
        "plt.rc('ytick', labelsize=6)\n",
        "\n",
        "X_train_prep_df.hist(bins=50, figsize=(14, 10))\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Let's also check the correlation between the new columns."
      ],
      "metadata": {
        "id": "U5nC6aHdvDNp"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "9e33nFko6U2T"
      },
      "outputs": [],
      "source": [
        "from sklearn.preprocessing import LabelEncoder\n",
        "\n",
        "le = LabelEncoder()\n",
        "\n",
        "corr_train_prep_df = X_train_prep_df.copy()\n",
        "corr_train_prep_df.insert (0, \"class\", le.fit_transform(y_train_copy.reset_index(drop=True)), True)\n",
        "\n",
        "corr_train_prep_matrix = corr_train_prep_df.corr(numeric_only=True)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "collapsed": true,
        "id": "R9aTDFVa6sNg",
        "outputId": "7267d067-1cbd-4579-ae58-d0f0e13a1592"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "=====monthly_credit_burden__ratio=====\n",
            "monthly_credit_burden__ratio     1.000000\n",
            "log__credit_amount               0.664788\n",
            "dependents_per_credit__ratio    -0.603197\n",
            "installment_per_credit__ratio   -0.737331\n",
            "Name: monthly_credit_burden__ratio, dtype: float64\n",
            "\n",
            "=====installment_per_credit__ratio=====\n",
            "installment_per_credit__ratio    1.000000\n",
            "dependents_per_credit__ratio     0.832216\n",
            "log__installment_commitment      0.674063\n",
            "monthly_credit_burden__ratio    -0.737331\n",
            "log__credit_amount              -0.891710\n",
            "Name: installment_per_credit__ratio, dtype: float64\n",
            "\n",
            "=====dependents_per_credit__ratio=====\n",
            "dependents_per_credit__ratio     1.000000\n",
            "installment_per_credit__ratio    0.832216\n",
            "monthly_credit_burden__ratio    -0.603197\n",
            "log__duration                   -0.630935\n",
            "log__credit_amount              -0.948341\n",
            "Name: dependents_per_credit__ratio, dtype: float64\n",
            "\n",
            "=====log__duration=====\n",
            "log__duration                   1.000000\n",
            "log__credit_amount              0.635610\n",
            "dependents_per_credit__ratio   -0.630935\n",
            "Name: log__duration, dtype: float64\n",
            "\n",
            "=====log__credit_amount=====\n",
            "log__credit_amount               1.000000\n",
            "monthly_credit_burden__ratio     0.664788\n",
            "log__duration                    0.635610\n",
            "installment_per_credit__ratio   -0.891710\n",
            "dependents_per_credit__ratio    -0.948341\n",
            "Name: log__credit_amount, dtype: float64\n",
            "\n",
            "=====log__installment_commitment=====\n",
            "log__installment_commitment      1.000000\n",
            "installment_per_credit__ratio    0.674063\n",
            "Name: log__installment_commitment, dtype: float64\n",
            "\n",
            "=====log__existing_credits=====\n",
            "log__existing_credits                                        1.000000\n",
            "cat_onehot__credit_history_critical/other existing credit    0.544207\n",
            "cat_onehot__credit_history_existing paid                    -0.619858\n",
            "Name: log__existing_credits, dtype: float64\n",
            "\n",
            "=====cat_ordinal__checking_status=====\n",
            "cat_ordinal__checking_status    1.000000\n",
            "geo__Cluster 9 similarity       0.708038\n",
            "geo__Cluster 1 similarity       0.641539\n",
            "geo__Cluster 8 similarity       0.614203\n",
            "geo__Cluster 6 similarity      -0.691783\n",
            "Name: cat_ordinal__checking_status, dtype: float64\n",
            "\n",
            "=====cat_ordinal__savings_status=====\n",
            "cat_ordinal__savings_status    1.000000\n",
            "geo__Cluster 2 similarity      0.604368\n",
            "geo__Cluster 4 similarity     -0.504253\n",
            "Name: cat_ordinal__savings_status, dtype: float64\n",
            "\n",
            "=====cat_onehot__job_high qualif/self emp/mgmt=====\n",
            "cat_onehot__job_high qualif/self emp/mgmt    1.000000\n",
            "cat_onehot__job_skilled                     -0.544443\n",
            "Name: cat_onehot__job_high qualif/self emp/mgmt, dtype: float64\n",
            "\n",
            "=====cat_onehot__job_skilled=====\n",
            "cat_onehot__job_skilled                      1.000000\n",
            "cat_onehot__job_high qualif/self emp/mgmt   -0.544443\n",
            "cat_onehot__job_unskilled resident          -0.655148\n",
            "Name: cat_onehot__job_skilled, dtype: float64\n",
            "\n",
            "=====cat_onehot__job_unskilled resident=====\n",
            "cat_onehot__job_unskilled resident    1.000000\n",
            "cat_onehot__job_skilled              -0.655148\n",
            "Name: cat_onehot__job_unskilled resident, dtype: float64\n",
            "\n",
            "=====cat_onehot__housing_for free=====\n",
            "cat_onehot__housing_for free                        1.000000\n",
            "cat_onehot__property_magnitude_no known property    0.777798\n",
            "cat_onehot__housing_own                            -0.547001\n",
            "Name: cat_onehot__housing_for free, dtype: float64\n",
            "\n",
            "=====cat_onehot__housing_own=====\n",
            "cat_onehot__housing_own         1.000000\n",
            "cat_onehot__housing_for free   -0.547001\n",
            "cat_onehot__housing_rent       -0.742249\n",
            "Name: cat_onehot__housing_own, dtype: float64\n",
            "\n",
            "=====cat_onehot__housing_rent=====\n",
            "cat_onehot__housing_rent    1.000000\n",
            "cat_onehot__housing_own    -0.742249\n",
            "Name: cat_onehot__housing_rent, dtype: float64\n",
            "\n",
            "=====cat_onehot__other_payment_plans_bank=====\n",
            "cat_onehot__other_payment_plans_bank    1.000000\n",
            "cat_onehot__other_payment_plans_none   -0.821578\n",
            "Name: cat_onehot__other_payment_plans_bank, dtype: float64\n",
            "\n",
            "=====cat_onehot__other_payment_plans_none=====\n",
            "cat_onehot__other_payment_plans_none    1.000000\n",
            "cat_onehot__other_payment_plans_bank   -0.821578\n",
            "Name: cat_onehot__other_payment_plans_none, dtype: float64\n",
            "\n",
            "=====cat_onehot__property_magnitude_no known property=====\n",
            "cat_onehot__property_magnitude_no known property    1.000000\n",
            "cat_onehot__housing_for free                        0.777798\n",
            "Name: cat_onehot__property_magnitude_no known property, dtype: float64\n",
            "\n",
            "=====cat_onehot__other_parties_co applicant=====\n",
            "cat_onehot__other_parties_co applicant    1.000000\n",
            "cat_onehot__other_parties_none           -0.665163\n",
            "Name: cat_onehot__other_parties_co applicant, dtype: float64\n",
            "\n",
            "=====cat_onehot__other_parties_guarantor=====\n",
            "cat_onehot__other_parties_guarantor    1.000000\n",
            "cat_onehot__other_parties_none        -0.712188\n",
            "Name: cat_onehot__other_parties_guarantor, dtype: float64\n",
            "\n",
            "=====cat_onehot__other_parties_none=====\n",
            "cat_onehot__other_parties_none            1.000000\n",
            "cat_onehot__other_parties_co applicant   -0.665163\n",
            "cat_onehot__other_parties_guarantor      -0.712188\n",
            "Name: cat_onehot__other_parties_none, dtype: float64\n",
            "\n",
            "=====cat_onehot__personal_status_female div/dep/mar=====\n",
            "cat_onehot__personal_status_female div/dep/mar    1.000000\n",
            "cat_onehot__personal_status_male single          -0.737872\n",
            "Name: cat_onehot__personal_status_female div/dep/mar, dtype: float64\n",
            "\n",
            "=====cat_onehot__personal_status_male single=====\n",
            "cat_onehot__personal_status_male single           1.000000\n",
            "cat_onehot__personal_status_female div/dep/mar   -0.737872\n",
            "Name: cat_onehot__personal_status_male single, dtype: float64\n",
            "\n",
            "=====cat_onehot__credit_history_critical/other existing credit=====\n",
            "cat_onehot__credit_history_critical/other existing credit    1.000000\n",
            "log__existing_credits                                        0.544207\n",
            "cat_onehot__credit_history_existing paid                    -0.709401\n",
            "Name: cat_onehot__credit_history_critical/other existing credit, dtype: float64\n",
            "\n",
            "=====cat_onehot__credit_history_existing paid=====\n",
            "cat_onehot__credit_history_existing paid                     1.000000\n",
            "log__existing_credits                                       -0.619858\n",
            "cat_onehot__credit_history_critical/other existing credit   -0.709401\n",
            "Name: cat_onehot__credit_history_existing paid, dtype: float64\n",
            "\n",
            "=====geo__Cluster 1 similarity=====\n",
            "geo__Cluster 1 similarity       1.000000\n",
            "cat_ordinal__checking_status    0.641539\n",
            "geo__Cluster 8 similarity       0.622303\n",
            "geo__Cluster 9 similarity       0.586287\n",
            "Name: geo__Cluster 1 similarity, dtype: float64\n",
            "\n",
            "=====geo__Cluster 2 similarity=====\n",
            "geo__Cluster 2 similarity      1.000000\n",
            "cat_ordinal__savings_status    0.604368\n",
            "Name: geo__Cluster 2 similarity, dtype: float64\n",
            "\n",
            "=====geo__Cluster 4 similarity=====\n",
            "geo__Cluster 4 similarity      1.000000\n",
            "cat_ordinal__savings_status   -0.504253\n",
            "Name: geo__Cluster 4 similarity, dtype: float64\n",
            "\n",
            "=====geo__Cluster 6 similarity=====\n",
            "geo__Cluster 6 similarity       1.000000\n",
            "cat_ordinal__checking_status   -0.691783\n",
            "Name: geo__Cluster 6 similarity, dtype: float64\n",
            "\n",
            "=====geo__Cluster 8 similarity=====\n",
            "geo__Cluster 8 similarity       1.000000\n",
            "geo__Cluster 1 similarity       0.622303\n",
            "cat_ordinal__checking_status    0.614203\n",
            "Name: geo__Cluster 8 similarity, dtype: float64\n",
            "\n",
            "=====geo__Cluster 9 similarity=====\n",
            "geo__Cluster 9 similarity       1.000000\n",
            "cat_ordinal__checking_status    0.708038\n",
            "geo__Cluster 1 similarity       0.586287\n",
            "Name: geo__Cluster 9 similarity, dtype: float64\n",
            "\n"
          ]
        }
      ],
      "source": [
        "checkForCorrelation (0.5, corr_train_prep_matrix, corr_train_prep_matrix.select_dtypes(include=['float64']), True)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "collapsed": true,
        "id": "Z3BDWk6EJhYV",
        "outputId": "0171b435-b174-44b0-fff2-faa69825be7c"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "class                                                        1.000000\n",
              "cat_ordinal__checking_status                                 0.308926\n",
              "geo__Cluster 9 similarity                                    0.264498\n",
              "geo__Cluster 1 similarity                                    0.198765\n",
              "geo__Cluster 8 similarity                                    0.196115\n",
              "cat_onehot__credit_history_critical/other existing credit    0.187973\n",
              "geo__Cluster 2 similarity                                    0.178445\n",
              "geo__Cluster 5 similarity                                    0.124555\n",
              "cat_onehot__other_payment_plans_none                         0.123915\n",
              "log__age                                                     0.123420\n",
              "cat_onehot__housing_own                                      0.120177\n",
              "cat_onehot__property_magnitude_real estate                   0.101463\n",
              "dependents_per_credit__ratio                                 0.096405\n",
              "cat_ordinal__savings_status                                  0.088750\n",
              "monthly_credit_burden__ratio                                 0.080374\n",
              "cat_onehot__other_parties_guarantor                          0.077932\n",
              "cat_onehot__personal_status_male single                      0.070114\n",
              "log__existing_credits                                        0.060861\n",
              "cat_ordinal__own_telephone                                   0.046190\n",
              "installment_per_credit__ratio                                0.036617\n",
              "cat_onehot__personal_status_male mar/wid                     0.034313\n",
              "log__num_dependents                                          0.031696\n",
              "cat_onehot__job_skilled                                      0.023213\n",
              "cat_onehot__property_magnitude_car                           0.011033\n",
              "cat_onehot__other_parties_none                              -0.000925\n",
              "cat_onehot__job_unskilled resident                          -0.002051\n",
              "geo__Cluster 3 similarity                                   -0.004305\n",
              "cat_onehot__property_magnitude_life insurance               -0.008348\n",
              "log__residence_since                                        -0.013798\n",
              "cat_onehot__job_high qualif/self emp/mgmt                   -0.014667\n",
              "cat_onehot__credit_history_delayed previously               -0.021267\n",
              "cat_ordinal__employment                                     -0.026155\n",
              "geo__Cluster 7 similarity                                   -0.027033\n",
              "cat_onehot__personal_status_male div/sep                    -0.029352\n",
              "cat_onehot__job_unemp/unskilled non res                     -0.035937\n",
              "cat_onehot__housing_for free                                -0.057557\n",
              "cat_onehot__other_payment_plans_stores                      -0.058140\n",
              "cat_onehot__credit_history_existing paid                    -0.062861\n",
              "log__installment_commitment                                 -0.068767\n",
              "cat_onehot__other_parties_co applicant                      -0.081580\n",
              "cat_onehot__personal_status_female div/dep/mar              -0.082389\n",
              "log__credit_amount                                          -0.089910\n",
              "cat_onehot__housing_rent                                    -0.095749\n",
              "cat_onehot__other_payment_plans_bank                        -0.103820\n",
              "geo__Cluster 4 similarity                                   -0.104826\n",
              "cat_onehot__credit_history_all paid                         -0.116927\n",
              "cat_onehot__credit_history_no credits/all paid              -0.132514\n",
              "cat_onehot__property_magnitude_no known property            -0.136061\n",
              "log__duration                                               -0.201980\n",
              "geo__Cluster 6 similarity                                   -0.242524\n",
              "geo__Cluster 0 similarity                                   -0.301927\n",
              "Name: class, dtype: float64"
            ],
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>class</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>class</th>\n",
              "      <td>1.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>cat_ordinal__checking_status</th>\n",
              "      <td>0.308926</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>geo__Cluster 9 similarity</th>\n",
              "      <td>0.264498</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>geo__Cluster 1 similarity</th>\n",
              "      <td>0.198765</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>geo__Cluster 8 similarity</th>\n",
              "      <td>0.196115</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>cat_onehot__credit_history_critical/other existing credit</th>\n",
              "      <td>0.187973</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>geo__Cluster 2 similarity</th>\n",
              "      <td>0.178445</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>geo__Cluster 5 similarity</th>\n",
              "      <td>0.124555</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>cat_onehot__other_payment_plans_none</th>\n",
              "      <td>0.123915</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>log__age</th>\n",
              "      <td>0.123420</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>cat_onehot__housing_own</th>\n",
              "      <td>0.120177</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>cat_onehot__property_magnitude_real estate</th>\n",
              "      <td>0.101463</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>dependents_per_credit__ratio</th>\n",
              "      <td>0.096405</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>cat_ordinal__savings_status</th>\n",
              "      <td>0.088750</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>monthly_credit_burden__ratio</th>\n",
              "      <td>0.080374</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>cat_onehot__other_parties_guarantor</th>\n",
              "      <td>0.077932</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>cat_onehot__personal_status_male single</th>\n",
              "      <td>0.070114</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>log__existing_credits</th>\n",
              "      <td>0.060861</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>cat_ordinal__own_telephone</th>\n",
              "      <td>0.046190</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>installment_per_credit__ratio</th>\n",
              "      <td>0.036617</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>cat_onehot__personal_status_male mar/wid</th>\n",
              "      <td>0.034313</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>log__num_dependents</th>\n",
              "      <td>0.031696</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>cat_onehot__job_skilled</th>\n",
              "      <td>0.023213</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>cat_onehot__property_magnitude_car</th>\n",
              "      <td>0.011033</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>cat_onehot__other_parties_none</th>\n",
              "      <td>-0.000925</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>cat_onehot__job_unskilled resident</th>\n",
              "      <td>-0.002051</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>geo__Cluster 3 similarity</th>\n",
              "      <td>-0.004305</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>cat_onehot__property_magnitude_life insurance</th>\n",
              "      <td>-0.008348</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>log__residence_since</th>\n",
              "      <td>-0.013798</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>cat_onehot__job_high qualif/self emp/mgmt</th>\n",
              "      <td>-0.014667</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>cat_onehot__credit_history_delayed previously</th>\n",
              "      <td>-0.021267</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>cat_ordinal__employment</th>\n",
              "      <td>-0.026155</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>geo__Cluster 7 similarity</th>\n",
              "      <td>-0.027033</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>cat_onehot__personal_status_male div/sep</th>\n",
              "      <td>-0.029352</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>cat_onehot__job_unemp/unskilled non res</th>\n",
              "      <td>-0.035937</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>cat_onehot__housing_for free</th>\n",
              "      <td>-0.057557</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>cat_onehot__other_payment_plans_stores</th>\n",
              "      <td>-0.058140</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>cat_onehot__credit_history_existing paid</th>\n",
              "      <td>-0.062861</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>log__installment_commitment</th>\n",
              "      <td>-0.068767</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>cat_onehot__other_parties_co applicant</th>\n",
              "      <td>-0.081580</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>cat_onehot__personal_status_female div/dep/mar</th>\n",
              "      <td>-0.082389</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>log__credit_amount</th>\n",
              "      <td>-0.089910</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>cat_onehot__housing_rent</th>\n",
              "      <td>-0.095749</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>cat_onehot__other_payment_plans_bank</th>\n",
              "      <td>-0.103820</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>geo__Cluster 4 similarity</th>\n",
              "      <td>-0.104826</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>cat_onehot__credit_history_all paid</th>\n",
              "      <td>-0.116927</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>cat_onehot__credit_history_no credits/all paid</th>\n",
              "      <td>-0.132514</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>cat_onehot__property_magnitude_no known property</th>\n",
              "      <td>-0.136061</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>log__duration</th>\n",
              "      <td>-0.201980</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>geo__Cluster 6 similarity</th>\n",
              "      <td>-0.242524</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>geo__Cluster 0 similarity</th>\n",
              "      <td>-0.301927</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div><br><label><b>dtype:</b> float64</label>"
            ]
          },
          "metadata": {},
          "execution_count": 167
        }
      ],
      "source": [
        "corr_train_prep_matrix[\"class\"].sort_values(ascending=False)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Wt45INQvoth8"
      },
      "source": [
        "# Model Selection"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "VGHl4roNkm9E"
      },
      "source": [
        "For model selection we are going to be experimenting with the following models:\n",
        "\n",
        "\n",
        "1.   Logistic Regression\n",
        "2.   Random Forest\n",
        "3. Gradient Boosting\n",
        "4. Support Vector Machine\n",
        "5. Extreme Gradient Boosting\n",
        "6. Stochastic Gradient Descent\n",
        "7. Ensemble Method\n",
        "\n",
        "Based on the following metrics we will be selecting a model for further evaluation and finetuining.\n",
        "\n",
        "\n",
        "\n",
        "1.   Accuracy\n",
        "2.   Precision\n",
        "3.   Recall\n",
        "4.   f1\n",
        "5.   roc_auc\n",
        "    \n"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Based on current literature about the German Credit Risk dataset, we expect the Random Forest and Extreme GRadient Boosting models to perform the best."
      ],
      "metadata": {
        "id": "Ie6XHbdgvxBL"
      }
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "hmR2l_Kus6ib"
      },
      "source": [
        "# Model Training\n",
        "\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "7nSEkt5rom24"
      },
      "source": [
        "### Logistic Regression"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "f3b3sdSJuaZn"
      },
      "source": [
        "First we will be training a Logistic Regression model on the preprocessed data."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 78,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 357
        },
        "id": "30XRb2KspCJ7",
        "outputId": "846a30cc-c13c-46f3-8da0-6c6bc79a9286"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Pipeline(steps=[('columntransformer',\n",
              "                 ColumnTransformer(remainder=Pipeline(steps=[('standardscaler',\n",
              "                                                              StandardScaler())]),\n",
              "                                   transformers=[('monthly_credit_burden',\n",
              "                                                  Pipeline(steps=[('functiontransformer-1',\n",
              "                                                                   FunctionTransformer(feature_names_out=<function ratio_name at 0x78a662f00dc0>,\n",
              "                                                                                       func=<function ratio_pipeline.<locals>.<lambda> at 0x78a6624f3250>)),\n",
              "                                                                  ('function...\n",
              "                                                   'personal_status',\n",
              "                                                   'credit_history']),\n",
              "                                                 ('geo',\n",
              "                                                  Pipeline(steps=[('ordered_ordinal_encoder',\n",
              "                                                                   OrderedOrdinalEncoder()),\n",
              "                                                                  ('cluster_simil',\n",
              "                                                                   ClusterSimilarity(random_state=42))]),\n",
              "                                                  ['savings_status',\n",
              "                                                   'checking_status']),\n",
              "                                                 ('drop_out', 'drop',\n",
              "                                                  ['num_dependents',\n",
              "                                                   'foreign_worker',\n",
              "                                                   'purpose'])])),\n",
              "                ('logisticregression',\n",
              "                 LogisticRegression(class_weight='balanced', random_state=42))])"
            ],
            "text/html": [
              "<style>#sk-container-id-10 {\n",
              "  /* Definition of color scheme common for light and dark mode */\n",
              "  --sklearn-color-text: black;\n",
              "  --sklearn-color-line: gray;\n",
              "  /* Definition of color scheme for unfitted estimators */\n",
              "  --sklearn-color-unfitted-level-0: #fff5e6;\n",
              "  --sklearn-color-unfitted-level-1: #f6e4d2;\n",
              "  --sklearn-color-unfitted-level-2: #ffe0b3;\n",
              "  --sklearn-color-unfitted-level-3: chocolate;\n",
              "  /* Definition of color scheme for fitted estimators */\n",
              "  --sklearn-color-fitted-level-0: #f0f8ff;\n",
              "  --sklearn-color-fitted-level-1: #d4ebff;\n",
              "  --sklearn-color-fitted-level-2: #b3dbfd;\n",
              "  --sklearn-color-fitted-level-3: cornflowerblue;\n",
              "\n",
              "  /* Specific color for light theme */\n",
              "  --sklearn-color-text-on-default-background: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, black)));\n",
              "  --sklearn-color-background: var(--sg-background-color, var(--theme-background, var(--jp-layout-color0, white)));\n",
              "  --sklearn-color-border-box: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, black)));\n",
              "  --sklearn-color-icon: #696969;\n",
              "\n",
              "  @media (prefers-color-scheme: dark) {\n",
              "    /* Redefinition of color scheme for dark theme */\n",
              "    --sklearn-color-text-on-default-background: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, white)));\n",
              "    --sklearn-color-background: var(--sg-background-color, var(--theme-background, var(--jp-layout-color0, #111)));\n",
              "    --sklearn-color-border-box: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, white)));\n",
              "    --sklearn-color-icon: #878787;\n",
              "  }\n",
              "}\n",
              "\n",
              "#sk-container-id-10 {\n",
              "  color: var(--sklearn-color-text);\n",
              "}\n",
              "\n",
              "#sk-container-id-10 pre {\n",
              "  padding: 0;\n",
              "}\n",
              "\n",
              "#sk-container-id-10 input.sk-hidden--visually {\n",
              "  border: 0;\n",
              "  clip: rect(1px 1px 1px 1px);\n",
              "  clip: rect(1px, 1px, 1px, 1px);\n",
              "  height: 1px;\n",
              "  margin: -1px;\n",
              "  overflow: hidden;\n",
              "  padding: 0;\n",
              "  position: absolute;\n",
              "  width: 1px;\n",
              "}\n",
              "\n",
              "#sk-container-id-10 div.sk-dashed-wrapped {\n",
              "  border: 1px dashed var(--sklearn-color-line);\n",
              "  margin: 0 0.4em 0.5em 0.4em;\n",
              "  box-sizing: border-box;\n",
              "  padding-bottom: 0.4em;\n",
              "  background-color: var(--sklearn-color-background);\n",
              "}\n",
              "\n",
              "#sk-container-id-10 div.sk-container {\n",
              "  /* jupyter's `normalize.less` sets `[hidden] { display: none; }`\n",
              "     but bootstrap.min.css set `[hidden] { display: none !important; }`\n",
              "     so we also need the `!important` here to be able to override the\n",
              "     default hidden behavior on the sphinx rendered scikit-learn.org.\n",
              "     See: https://github.com/scikit-learn/scikit-learn/issues/21755 */\n",
              "  display: inline-block !important;\n",
              "  position: relative;\n",
              "}\n",
              "\n",
              "#sk-container-id-10 div.sk-text-repr-fallback {\n",
              "  display: none;\n",
              "}\n",
              "\n",
              "div.sk-parallel-item,\n",
              "div.sk-serial,\n",
              "div.sk-item {\n",
              "  /* draw centered vertical line to link estimators */\n",
              "  background-image: linear-gradient(var(--sklearn-color-text-on-default-background), var(--sklearn-color-text-on-default-background));\n",
              "  background-size: 2px 100%;\n",
              "  background-repeat: no-repeat;\n",
              "  background-position: center center;\n",
              "}\n",
              "\n",
              "/* Parallel-specific style estimator block */\n",
              "\n",
              "#sk-container-id-10 div.sk-parallel-item::after {\n",
              "  content: \"\";\n",
              "  width: 100%;\n",
              "  border-bottom: 2px solid var(--sklearn-color-text-on-default-background);\n",
              "  flex-grow: 1;\n",
              "}\n",
              "\n",
              "#sk-container-id-10 div.sk-parallel {\n",
              "  display: flex;\n",
              "  align-items: stretch;\n",
              "  justify-content: center;\n",
              "  background-color: var(--sklearn-color-background);\n",
              "  position: relative;\n",
              "}\n",
              "\n",
              "#sk-container-id-10 div.sk-parallel-item {\n",
              "  display: flex;\n",
              "  flex-direction: column;\n",
              "}\n",
              "\n",
              "#sk-container-id-10 div.sk-parallel-item:first-child::after {\n",
              "  align-self: flex-end;\n",
              "  width: 50%;\n",
              "}\n",
              "\n",
              "#sk-container-id-10 div.sk-parallel-item:last-child::after {\n",
              "  align-self: flex-start;\n",
              "  width: 50%;\n",
              "}\n",
              "\n",
              "#sk-container-id-10 div.sk-parallel-item:only-child::after {\n",
              "  width: 0;\n",
              "}\n",
              "\n",
              "/* Serial-specific style estimator block */\n",
              "\n",
              "#sk-container-id-10 div.sk-serial {\n",
              "  display: flex;\n",
              "  flex-direction: column;\n",
              "  align-items: center;\n",
              "  background-color: var(--sklearn-color-background);\n",
              "  padding-right: 1em;\n",
              "  padding-left: 1em;\n",
              "}\n",
              "\n",
              "\n",
              "/* Toggleable style: style used for estimator/Pipeline/ColumnTransformer box that is\n",
              "clickable and can be expanded/collapsed.\n",
              "- Pipeline and ColumnTransformer use this feature and define the default style\n",
              "- Estimators will overwrite some part of the style using the `sk-estimator` class\n",
              "*/\n",
              "\n",
              "/* Pipeline and ColumnTransformer style (default) */\n",
              "\n",
              "#sk-container-id-10 div.sk-toggleable {\n",
              "  /* Default theme specific background. It is overwritten whether we have a\n",
              "  specific estimator or a Pipeline/ColumnTransformer */\n",
              "  background-color: var(--sklearn-color-background);\n",
              "}\n",
              "\n",
              "/* Toggleable label */\n",
              "#sk-container-id-10 label.sk-toggleable__label {\n",
              "  cursor: pointer;\n",
              "  display: block;\n",
              "  width: 100%;\n",
              "  margin-bottom: 0;\n",
              "  padding: 0.5em;\n",
              "  box-sizing: border-box;\n",
              "  text-align: center;\n",
              "}\n",
              "\n",
              "#sk-container-id-10 label.sk-toggleable__label-arrow:before {\n",
              "  /* Arrow on the left of the label */\n",
              "  content: \"▸\";\n",
              "  float: left;\n",
              "  margin-right: 0.25em;\n",
              "  color: var(--sklearn-color-icon);\n",
              "}\n",
              "\n",
              "#sk-container-id-10 label.sk-toggleable__label-arrow:hover:before {\n",
              "  color: var(--sklearn-color-text);\n",
              "}\n",
              "\n",
              "/* Toggleable content - dropdown */\n",
              "\n",
              "#sk-container-id-10 div.sk-toggleable__content {\n",
              "  max-height: 0;\n",
              "  max-width: 0;\n",
              "  overflow: hidden;\n",
              "  text-align: left;\n",
              "  /* unfitted */\n",
              "  background-color: var(--sklearn-color-unfitted-level-0);\n",
              "}\n",
              "\n",
              "#sk-container-id-10 div.sk-toggleable__content.fitted {\n",
              "  /* fitted */\n",
              "  background-color: var(--sklearn-color-fitted-level-0);\n",
              "}\n",
              "\n",
              "#sk-container-id-10 div.sk-toggleable__content pre {\n",
              "  margin: 0.2em;\n",
              "  border-radius: 0.25em;\n",
              "  color: var(--sklearn-color-text);\n",
              "  /* unfitted */\n",
              "  background-color: var(--sklearn-color-unfitted-level-0);\n",
              "}\n",
              "\n",
              "#sk-container-id-10 div.sk-toggleable__content.fitted pre {\n",
              "  /* unfitted */\n",
              "  background-color: var(--sklearn-color-fitted-level-0);\n",
              "}\n",
              "\n",
              "#sk-container-id-10 input.sk-toggleable__control:checked~div.sk-toggleable__content {\n",
              "  /* Expand drop-down */\n",
              "  max-height: 200px;\n",
              "  max-width: 100%;\n",
              "  overflow: auto;\n",
              "}\n",
              "\n",
              "#sk-container-id-10 input.sk-toggleable__control:checked~label.sk-toggleable__label-arrow:before {\n",
              "  content: \"▾\";\n",
              "}\n",
              "\n",
              "/* Pipeline/ColumnTransformer-specific style */\n",
              "\n",
              "#sk-container-id-10 div.sk-label input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
              "  color: var(--sklearn-color-text);\n",
              "  background-color: var(--sklearn-color-unfitted-level-2);\n",
              "}\n",
              "\n",
              "#sk-container-id-10 div.sk-label.fitted input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
              "  background-color: var(--sklearn-color-fitted-level-2);\n",
              "}\n",
              "\n",
              "/* Estimator-specific style */\n",
              "\n",
              "/* Colorize estimator box */\n",
              "#sk-container-id-10 div.sk-estimator input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
              "  /* unfitted */\n",
              "  background-color: var(--sklearn-color-unfitted-level-2);\n",
              "}\n",
              "\n",
              "#sk-container-id-10 div.sk-estimator.fitted input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
              "  /* fitted */\n",
              "  background-color: var(--sklearn-color-fitted-level-2);\n",
              "}\n",
              "\n",
              "#sk-container-id-10 div.sk-label label.sk-toggleable__label,\n",
              "#sk-container-id-10 div.sk-label label {\n",
              "  /* The background is the default theme color */\n",
              "  color: var(--sklearn-color-text-on-default-background);\n",
              "}\n",
              "\n",
              "/* On hover, darken the color of the background */\n",
              "#sk-container-id-10 div.sk-label:hover label.sk-toggleable__label {\n",
              "  color: var(--sklearn-color-text);\n",
              "  background-color: var(--sklearn-color-unfitted-level-2);\n",
              "}\n",
              "\n",
              "/* Label box, darken color on hover, fitted */\n",
              "#sk-container-id-10 div.sk-label.fitted:hover label.sk-toggleable__label.fitted {\n",
              "  color: var(--sklearn-color-text);\n",
              "  background-color: var(--sklearn-color-fitted-level-2);\n",
              "}\n",
              "\n",
              "/* Estimator label */\n",
              "\n",
              "#sk-container-id-10 div.sk-label label {\n",
              "  font-family: monospace;\n",
              "  font-weight: bold;\n",
              "  display: inline-block;\n",
              "  line-height: 1.2em;\n",
              "}\n",
              "\n",
              "#sk-container-id-10 div.sk-label-container {\n",
              "  text-align: center;\n",
              "}\n",
              "\n",
              "/* Estimator-specific */\n",
              "#sk-container-id-10 div.sk-estimator {\n",
              "  font-family: monospace;\n",
              "  border: 1px dotted var(--sklearn-color-border-box);\n",
              "  border-radius: 0.25em;\n",
              "  box-sizing: border-box;\n",
              "  margin-bottom: 0.5em;\n",
              "  /* unfitted */\n",
              "  background-color: var(--sklearn-color-unfitted-level-0);\n",
              "}\n",
              "\n",
              "#sk-container-id-10 div.sk-estimator.fitted {\n",
              "  /* fitted */\n",
              "  background-color: var(--sklearn-color-fitted-level-0);\n",
              "}\n",
              "\n",
              "/* on hover */\n",
              "#sk-container-id-10 div.sk-estimator:hover {\n",
              "  /* unfitted */\n",
              "  background-color: var(--sklearn-color-unfitted-level-2);\n",
              "}\n",
              "\n",
              "#sk-container-id-10 div.sk-estimator.fitted:hover {\n",
              "  /* fitted */\n",
              "  background-color: var(--sklearn-color-fitted-level-2);\n",
              "}\n",
              "\n",
              "/* Specification for estimator info (e.g. \"i\" and \"?\") */\n",
              "\n",
              "/* Common style for \"i\" and \"?\" */\n",
              "\n",
              ".sk-estimator-doc-link,\n",
              "a:link.sk-estimator-doc-link,\n",
              "a:visited.sk-estimator-doc-link {\n",
              "  float: right;\n",
              "  font-size: smaller;\n",
              "  line-height: 1em;\n",
              "  font-family: monospace;\n",
              "  background-color: var(--sklearn-color-background);\n",
              "  border-radius: 1em;\n",
              "  height: 1em;\n",
              "  width: 1em;\n",
              "  text-decoration: none !important;\n",
              "  margin-left: 1ex;\n",
              "  /* unfitted */\n",
              "  border: var(--sklearn-color-unfitted-level-1) 1pt solid;\n",
              "  color: var(--sklearn-color-unfitted-level-1);\n",
              "}\n",
              "\n",
              ".sk-estimator-doc-link.fitted,\n",
              "a:link.sk-estimator-doc-link.fitted,\n",
              "a:visited.sk-estimator-doc-link.fitted {\n",
              "  /* fitted */\n",
              "  border: var(--sklearn-color-fitted-level-1) 1pt solid;\n",
              "  color: var(--sklearn-color-fitted-level-1);\n",
              "}\n",
              "\n",
              "/* On hover */\n",
              "div.sk-estimator:hover .sk-estimator-doc-link:hover,\n",
              ".sk-estimator-doc-link:hover,\n",
              "div.sk-label-container:hover .sk-estimator-doc-link:hover,\n",
              ".sk-estimator-doc-link:hover {\n",
              "  /* unfitted */\n",
              "  background-color: var(--sklearn-color-unfitted-level-3);\n",
              "  color: var(--sklearn-color-background);\n",
              "  text-decoration: none;\n",
              "}\n",
              "\n",
              "div.sk-estimator.fitted:hover .sk-estimator-doc-link.fitted:hover,\n",
              ".sk-estimator-doc-link.fitted:hover,\n",
              "div.sk-label-container:hover .sk-estimator-doc-link.fitted:hover,\n",
              ".sk-estimator-doc-link.fitted:hover {\n",
              "  /* fitted */\n",
              "  background-color: var(--sklearn-color-fitted-level-3);\n",
              "  color: var(--sklearn-color-background);\n",
              "  text-decoration: none;\n",
              "}\n",
              "\n",
              "/* Span, style for the box shown on hovering the info icon */\n",
              ".sk-estimator-doc-link span {\n",
              "  display: none;\n",
              "  z-index: 9999;\n",
              "  position: relative;\n",
              "  font-weight: normal;\n",
              "  right: .2ex;\n",
              "  padding: .5ex;\n",
              "  margin: .5ex;\n",
              "  width: min-content;\n",
              "  min-width: 20ex;\n",
              "  max-width: 50ex;\n",
              "  color: var(--sklearn-color-text);\n",
              "  box-shadow: 2pt 2pt 4pt #999;\n",
              "  /* unfitted */\n",
              "  background: var(--sklearn-color-unfitted-level-0);\n",
              "  border: .5pt solid var(--sklearn-color-unfitted-level-3);\n",
              "}\n",
              "\n",
              ".sk-estimator-doc-link.fitted span {\n",
              "  /* fitted */\n",
              "  background: var(--sklearn-color-fitted-level-0);\n",
              "  border: var(--sklearn-color-fitted-level-3);\n",
              "}\n",
              "\n",
              ".sk-estimator-doc-link:hover span {\n",
              "  display: block;\n",
              "}\n",
              "\n",
              "/* \"?\"-specific style due to the `<a>` HTML tag */\n",
              "\n",
              "#sk-container-id-10 a.estimator_doc_link {\n",
              "  float: right;\n",
              "  font-size: 1rem;\n",
              "  line-height: 1em;\n",
              "  font-family: monospace;\n",
              "  background-color: var(--sklearn-color-background);\n",
              "  border-radius: 1rem;\n",
              "  height: 1rem;\n",
              "  width: 1rem;\n",
              "  text-decoration: none;\n",
              "  /* unfitted */\n",
              "  color: var(--sklearn-color-unfitted-level-1);\n",
              "  border: var(--sklearn-color-unfitted-level-1) 1pt solid;\n",
              "}\n",
              "\n",
              "#sk-container-id-10 a.estimator_doc_link.fitted {\n",
              "  /* fitted */\n",
              "  border: var(--sklearn-color-fitted-level-1) 1pt solid;\n",
              "  color: var(--sklearn-color-fitted-level-1);\n",
              "}\n",
              "\n",
              "/* On hover */\n",
              "#sk-container-id-10 a.estimator_doc_link:hover {\n",
              "  /* unfitted */\n",
              "  background-color: var(--sklearn-color-unfitted-level-3);\n",
              "  color: var(--sklearn-color-background);\n",
              "  text-decoration: none;\n",
              "}\n",
              "\n",
              "#sk-container-id-10 a.estimator_doc_link.fitted:hover {\n",
              "  /* fitted */\n",
              "  background-color: var(--sklearn-color-fitted-level-3);\n",
              "}\n",
              "</style><div id=\"sk-container-id-10\" class=\"sk-top-container\"><div class=\"sk-text-repr-fallback\"><pre>Pipeline(steps=[(&#x27;columntransformer&#x27;,\n",
              "                 ColumnTransformer(remainder=Pipeline(steps=[(&#x27;standardscaler&#x27;,\n",
              "                                                              StandardScaler())]),\n",
              "                                   transformers=[(&#x27;monthly_credit_burden&#x27;,\n",
              "                                                  Pipeline(steps=[(&#x27;functiontransformer-1&#x27;,\n",
              "                                                                   FunctionTransformer(feature_names_out=&lt;function ratio_name at 0x78a662f00dc0&gt;,\n",
              "                                                                                       func=&lt;function ratio_pipeline.&lt;locals&gt;.&lt;lambda&gt; at 0x78a6624f3250&gt;)),\n",
              "                                                                  (&#x27;function...\n",
              "                                                   &#x27;personal_status&#x27;,\n",
              "                                                   &#x27;credit_history&#x27;]),\n",
              "                                                 (&#x27;geo&#x27;,\n",
              "                                                  Pipeline(steps=[(&#x27;ordered_ordinal_encoder&#x27;,\n",
              "                                                                   OrderedOrdinalEncoder()),\n",
              "                                                                  (&#x27;cluster_simil&#x27;,\n",
              "                                                                   ClusterSimilarity(random_state=42))]),\n",
              "                                                  [&#x27;savings_status&#x27;,\n",
              "                                                   &#x27;checking_status&#x27;]),\n",
              "                                                 (&#x27;drop_out&#x27;, &#x27;drop&#x27;,\n",
              "                                                  [&#x27;num_dependents&#x27;,\n",
              "                                                   &#x27;foreign_worker&#x27;,\n",
              "                                                   &#x27;purpose&#x27;])])),\n",
              "                (&#x27;logisticregression&#x27;,\n",
              "                 LogisticRegression(class_weight=&#x27;balanced&#x27;, random_state=42))])</pre><b>In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook. <br />On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.</b></div><div class=\"sk-container\" hidden><div class=\"sk-item sk-dashed-wrapped\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-262\" type=\"checkbox\" ><label for=\"sk-estimator-id-262\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;&nbsp;Pipeline<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.pipeline.Pipeline.html\">?<span>Documentation for Pipeline</span></a><span class=\"sk-estimator-doc-link fitted\">i<span>Fitted</span></span></label><div class=\"sk-toggleable__content fitted\"><pre>Pipeline(steps=[(&#x27;columntransformer&#x27;,\n",
              "                 ColumnTransformer(remainder=Pipeline(steps=[(&#x27;standardscaler&#x27;,\n",
              "                                                              StandardScaler())]),\n",
              "                                   transformers=[(&#x27;monthly_credit_burden&#x27;,\n",
              "                                                  Pipeline(steps=[(&#x27;functiontransformer-1&#x27;,\n",
              "                                                                   FunctionTransformer(feature_names_out=&lt;function ratio_name at 0x78a662f00dc0&gt;,\n",
              "                                                                                       func=&lt;function ratio_pipeline.&lt;locals&gt;.&lt;lambda&gt; at 0x78a6624f3250&gt;)),\n",
              "                                                                  (&#x27;function...\n",
              "                                                   &#x27;personal_status&#x27;,\n",
              "                                                   &#x27;credit_history&#x27;]),\n",
              "                                                 (&#x27;geo&#x27;,\n",
              "                                                  Pipeline(steps=[(&#x27;ordered_ordinal_encoder&#x27;,\n",
              "                                                                   OrderedOrdinalEncoder()),\n",
              "                                                                  (&#x27;cluster_simil&#x27;,\n",
              "                                                                   ClusterSimilarity(random_state=42))]),\n",
              "                                                  [&#x27;savings_status&#x27;,\n",
              "                                                   &#x27;checking_status&#x27;]),\n",
              "                                                 (&#x27;drop_out&#x27;, &#x27;drop&#x27;,\n",
              "                                                  [&#x27;num_dependents&#x27;,\n",
              "                                                   &#x27;foreign_worker&#x27;,\n",
              "                                                   &#x27;purpose&#x27;])])),\n",
              "                (&#x27;logisticregression&#x27;,\n",
              "                 LogisticRegression(class_weight=&#x27;balanced&#x27;, random_state=42))])</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item sk-dashed-wrapped\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-263\" type=\"checkbox\" ><label for=\"sk-estimator-id-263\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;columntransformer: ColumnTransformer<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.compose.ColumnTransformer.html\">?<span>Documentation for columntransformer: ColumnTransformer</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>ColumnTransformer(remainder=Pipeline(steps=[(&#x27;standardscaler&#x27;,\n",
              "                                             StandardScaler())]),\n",
              "                  transformers=[(&#x27;monthly_credit_burden&#x27;,\n",
              "                                 Pipeline(steps=[(&#x27;functiontransformer-1&#x27;,\n",
              "                                                  FunctionTransformer(feature_names_out=&lt;function ratio_name at 0x78a662f00dc0&gt;,\n",
              "                                                                      func=&lt;function ratio_pipeline.&lt;locals&gt;.&lt;lambda&gt; at 0x78a6624f3250&gt;)),\n",
              "                                                 (&#x27;functiontransformer-2&#x27;,\n",
              "                                                  FunctionTransformer(fe...\n",
              "                                                  OneHotEncoder(handle_unknown=&#x27;ignore&#x27;))]),\n",
              "                                 [&#x27;job&#x27;, &#x27;housing&#x27;, &#x27;other_payment_plans&#x27;,\n",
              "                                  &#x27;property_magnitude&#x27;, &#x27;other_parties&#x27;,\n",
              "                                  &#x27;personal_status&#x27;, &#x27;credit_history&#x27;]),\n",
              "                                (&#x27;geo&#x27;,\n",
              "                                 Pipeline(steps=[(&#x27;ordered_ordinal_encoder&#x27;,\n",
              "                                                  OrderedOrdinalEncoder()),\n",
              "                                                 (&#x27;cluster_simil&#x27;,\n",
              "                                                  ClusterSimilarity(random_state=42))]),\n",
              "                                 [&#x27;savings_status&#x27;, &#x27;checking_status&#x27;]),\n",
              "                                (&#x27;drop_out&#x27;, &#x27;drop&#x27;,\n",
              "                                 [&#x27;num_dependents&#x27;, &#x27;foreign_worker&#x27;,\n",
              "                                  &#x27;purpose&#x27;])])</pre></div> </div></div><div class=\"sk-parallel\"><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-264\" type=\"checkbox\" ><label for=\"sk-estimator-id-264\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">monthly_credit_burden</label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;credit_amount&#x27;, &#x27;duration&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-265\" type=\"checkbox\" ><label for=\"sk-estimator-id-265\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;FunctionTransformer<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.FunctionTransformer.html\">?<span>Documentation for FunctionTransformer</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>FunctionTransformer(feature_names_out=&lt;function ratio_name at 0x78a662f00dc0&gt;,\n",
              "                    func=&lt;function ratio_pipeline.&lt;locals&gt;.&lt;lambda&gt; at 0x78a6624f3250&gt;)</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-266\" type=\"checkbox\" ><label for=\"sk-estimator-id-266\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;FunctionTransformer<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.FunctionTransformer.html\">?<span>Documentation for FunctionTransformer</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>FunctionTransformer(feature_names_out=&#x27;one-to-one&#x27;, func=&lt;ufunc &#x27;log&#x27;&gt;,\n",
              "                    inverse_func=&lt;ufunc &#x27;exp&#x27;&gt;)</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-267\" type=\"checkbox\" ><label for=\"sk-estimator-id-267\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;StandardScaler<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.StandardScaler.html\">?<span>Documentation for StandardScaler</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>StandardScaler()</pre></div> </div></div></div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-268\" type=\"checkbox\" ><label for=\"sk-estimator-id-268\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">installment_per_credit</label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;installment_commitment&#x27;, &#x27;credit_amount&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-269\" type=\"checkbox\" ><label for=\"sk-estimator-id-269\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;FunctionTransformer<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.FunctionTransformer.html\">?<span>Documentation for FunctionTransformer</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>FunctionTransformer(feature_names_out=&lt;function ratio_name at 0x78a662f00dc0&gt;,\n",
              "                    func=&lt;function ratio_pipeline.&lt;locals&gt;.&lt;lambda&gt; at 0x78a6624f3400&gt;)</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-270\" type=\"checkbox\" ><label for=\"sk-estimator-id-270\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;FunctionTransformer<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.FunctionTransformer.html\">?<span>Documentation for FunctionTransformer</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>FunctionTransformer(feature_names_out=&#x27;one-to-one&#x27;, func=&lt;ufunc &#x27;log&#x27;&gt;,\n",
              "                    inverse_func=&lt;ufunc &#x27;exp&#x27;&gt;)</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-271\" type=\"checkbox\" ><label for=\"sk-estimator-id-271\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;StandardScaler<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.StandardScaler.html\">?<span>Documentation for StandardScaler</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>StandardScaler()</pre></div> </div></div></div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-272\" type=\"checkbox\" ><label for=\"sk-estimator-id-272\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">dependents_per_credit</label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;num_dependents&#x27;, &#x27;credit_amount&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-273\" type=\"checkbox\" ><label for=\"sk-estimator-id-273\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;FunctionTransformer<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.FunctionTransformer.html\">?<span>Documentation for FunctionTransformer</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>FunctionTransformer(feature_names_out=&lt;function ratio_name at 0x78a662f00dc0&gt;,\n",
              "                    func=&lt;function ratio_pipeline.&lt;locals&gt;.&lt;lambda&gt; at 0x78a6624f35b0&gt;)</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-274\" type=\"checkbox\" ><label for=\"sk-estimator-id-274\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;FunctionTransformer<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.FunctionTransformer.html\">?<span>Documentation for FunctionTransformer</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>FunctionTransformer(feature_names_out=&#x27;one-to-one&#x27;, func=&lt;ufunc &#x27;log&#x27;&gt;,\n",
              "                    inverse_func=&lt;ufunc &#x27;exp&#x27;&gt;)</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-275\" type=\"checkbox\" ><label for=\"sk-estimator-id-275\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;StandardScaler<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.StandardScaler.html\">?<span>Documentation for StandardScaler</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>StandardScaler()</pre></div> </div></div></div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-276\" type=\"checkbox\" ><label for=\"sk-estimator-id-276\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">log</label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;duration&#x27;, &#x27;credit_amount&#x27;, &#x27;installment_commitment&#x27;, &#x27;residence_since&#x27;, &#x27;age&#x27;, &#x27;existing_credits&#x27;, &#x27;num_dependents&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-277\" type=\"checkbox\" ><label for=\"sk-estimator-id-277\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;FunctionTransformer<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.FunctionTransformer.html\">?<span>Documentation for FunctionTransformer</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>FunctionTransformer(feature_names_out=&#x27;one-to-one&#x27;, func=&lt;ufunc &#x27;log&#x27;&gt;,\n",
              "                    inverse_func=&lt;ufunc &#x27;exp&#x27;&gt;)</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-278\" type=\"checkbox\" ><label for=\"sk-estimator-id-278\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;StandardScaler<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.StandardScaler.html\">?<span>Documentation for StandardScaler</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>StandardScaler()</pre></div> </div></div></div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-279\" type=\"checkbox\" ><label for=\"sk-estimator-id-279\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">cat_ordinal</label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;checking_status&#x27;, &#x27;savings_status&#x27;, &#x27;employment&#x27;, &#x27;own_telephone&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-280\" type=\"checkbox\" ><label for=\"sk-estimator-id-280\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">OrderedOrdinalEncoder</label><div class=\"sk-toggleable__content fitted\"><pre>OrderedOrdinalEncoder()</pre></div> </div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-281\" type=\"checkbox\" ><label for=\"sk-estimator-id-281\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">cat_onehot</label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;job&#x27;, &#x27;housing&#x27;, &#x27;other_payment_plans&#x27;, &#x27;property_magnitude&#x27;, &#x27;other_parties&#x27;, &#x27;personal_status&#x27;, &#x27;credit_history&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-282\" type=\"checkbox\" ><label for=\"sk-estimator-id-282\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;OneHotEncoder<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.OneHotEncoder.html\">?<span>Documentation for OneHotEncoder</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>OneHotEncoder(handle_unknown=&#x27;ignore&#x27;)</pre></div> </div></div></div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-283\" type=\"checkbox\" ><label for=\"sk-estimator-id-283\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">geo</label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;savings_status&#x27;, &#x27;checking_status&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-284\" type=\"checkbox\" ><label for=\"sk-estimator-id-284\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">OrderedOrdinalEncoder</label><div class=\"sk-toggleable__content fitted\"><pre>OrderedOrdinalEncoder()</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-285\" type=\"checkbox\" ><label for=\"sk-estimator-id-285\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">ClusterSimilarity</label><div class=\"sk-toggleable__content fitted\"><pre>ClusterSimilarity(random_state=42)</pre></div> </div></div></div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-286\" type=\"checkbox\" ><label for=\"sk-estimator-id-286\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">drop_out</label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;num_dependents&#x27;, &#x27;foreign_worker&#x27;, &#x27;purpose&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-287\" type=\"checkbox\" ><label for=\"sk-estimator-id-287\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">drop</label><div class=\"sk-toggleable__content fitted\"><pre>drop</pre></div> </div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-288\" type=\"checkbox\" ><label for=\"sk-estimator-id-288\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">remainder</label><div class=\"sk-toggleable__content fitted\"><pre>[]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-289\" type=\"checkbox\" ><label for=\"sk-estimator-id-289\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;StandardScaler<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.StandardScaler.html\">?<span>Documentation for StandardScaler</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>StandardScaler()</pre></div> </div></div></div></div></div></div></div></div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-290\" type=\"checkbox\" ><label for=\"sk-estimator-id-290\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;LogisticRegression<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.linear_model.LogisticRegression.html\">?<span>Documentation for LogisticRegression</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>LogisticRegression(class_weight=&#x27;balanced&#x27;, random_state=42)</pre></div> </div></div></div></div></div></div>"
            ]
          },
          "metadata": {},
          "execution_count": 78
        }
      ],
      "source": [
        "from sklearn.linear_model import LogisticRegression\n",
        "from sklearn.pipeline import make_pipeline\n",
        "\n",
        "log_reg = make_pipeline(preprocessing, LogisticRegression(random_state=42, class_weight='balanced'))\n",
        "log_reg.fit(X_train, y_train)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "2WGYPacEuhXt"
      },
      "source": [
        "Testing the accuracy of the model:"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 77,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "0_vZY4Y9uvwx",
        "outputId": "bdd08c2a-ac55-46ab-c597-7330e2e1059c"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Training Set Performance:\n",
            "Accuracy: 0.7425\n",
            "Precision: 0.8766\n",
            "Recall: 0.7357\n",
            "F1-Score: 0.8000\n",
            "ROC-AUC: 0.8123\n",
            "------------------------------\n"
          ]
        }
      ],
      "source": [
        "from sklearn.metrics import accuracy_score, precision_score, recall_score, f1_score, roc_auc_score\n",
        "\n",
        "# Make predictions\n",
        "train_predictions = log_reg.predict(X_train)\n",
        "\n",
        "# Evaluation metrics\n",
        "train_accuracy = accuracy_score(y_train, train_predictions)\n",
        "train_precision = precision_score(y_train, train_predictions, pos_label='good')\n",
        "train_recall = recall_score(y_train, train_predictions, pos_label='good')\n",
        "train_f1 = f1_score(y_train, train_predictions, pos_label='good')\n",
        "train_roc_auc = roc_auc_score(y_train, log_reg.predict_proba(X_train)[:, 1])\n",
        "\n",
        "# Print training set results\n",
        "print(\"Training Set Performance:\")\n",
        "print(f\"Accuracy: {train_accuracy:.4f}\")\n",
        "print(f\"Precision: {train_precision:.4f}\")\n",
        "print(f\"Recall: {train_recall:.4f}\")\n",
        "print(f\"F1-Score: {train_f1:.4f}\")\n",
        "print(f\"ROC-AUC: {train_roc_auc:.4f}\")\n",
        "print(\"-\" * 30)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "DpF4K2GFu5Pk"
      },
      "source": [
        "### Random Forest"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ijnCCoNt2ir5"
      },
      "source": [
        "We are now going to be training a random forest model."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 72,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 357
        },
        "id": "vWh4H5dHvDGC",
        "outputId": "1ee1a3a1-78c5-4ece-b60b-035a315d38a2"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Pipeline(steps=[('columntransformer',\n",
              "                 ColumnTransformer(remainder=Pipeline(steps=[('standardscaler',\n",
              "                                                              StandardScaler())]),\n",
              "                                   transformers=[('monthly_credit_burden',\n",
              "                                                  Pipeline(steps=[('functiontransformer-1',\n",
              "                                                                   FunctionTransformer(feature_names_out=<function ratio_name at 0x78a662f00dc0>,\n",
              "                                                                                       func=<function ratio_pipeline.<locals>.<lambda> at 0x78a6624f3250>)),\n",
              "                                                                  ('function...\n",
              "                                                   'personal_status',\n",
              "                                                   'credit_history']),\n",
              "                                                 ('geo',\n",
              "                                                  Pipeline(steps=[('ordered_ordinal_encoder',\n",
              "                                                                   OrderedOrdinalEncoder()),\n",
              "                                                                  ('cluster_simil',\n",
              "                                                                   ClusterSimilarity(random_state=42))]),\n",
              "                                                  ['savings_status',\n",
              "                                                   'checking_status']),\n",
              "                                                 ('drop_out', 'drop',\n",
              "                                                  ['num_dependents',\n",
              "                                                   'foreign_worker',\n",
              "                                                   'purpose'])])),\n",
              "                ('randomforestclassifier',\n",
              "                 RandomForestClassifier(class_weight='balanced',\n",
              "                                        random_state=42))])"
            ],
            "text/html": [
              "<style>#sk-container-id-8 {\n",
              "  /* Definition of color scheme common for light and dark mode */\n",
              "  --sklearn-color-text: black;\n",
              "  --sklearn-color-line: gray;\n",
              "  /* Definition of color scheme for unfitted estimators */\n",
              "  --sklearn-color-unfitted-level-0: #fff5e6;\n",
              "  --sklearn-color-unfitted-level-1: #f6e4d2;\n",
              "  --sklearn-color-unfitted-level-2: #ffe0b3;\n",
              "  --sklearn-color-unfitted-level-3: chocolate;\n",
              "  /* Definition of color scheme for fitted estimators */\n",
              "  --sklearn-color-fitted-level-0: #f0f8ff;\n",
              "  --sklearn-color-fitted-level-1: #d4ebff;\n",
              "  --sklearn-color-fitted-level-2: #b3dbfd;\n",
              "  --sklearn-color-fitted-level-3: cornflowerblue;\n",
              "\n",
              "  /* Specific color for light theme */\n",
              "  --sklearn-color-text-on-default-background: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, black)));\n",
              "  --sklearn-color-background: var(--sg-background-color, var(--theme-background, var(--jp-layout-color0, white)));\n",
              "  --sklearn-color-border-box: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, black)));\n",
              "  --sklearn-color-icon: #696969;\n",
              "\n",
              "  @media (prefers-color-scheme: dark) {\n",
              "    /* Redefinition of color scheme for dark theme */\n",
              "    --sklearn-color-text-on-default-background: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, white)));\n",
              "    --sklearn-color-background: var(--sg-background-color, var(--theme-background, var(--jp-layout-color0, #111)));\n",
              "    --sklearn-color-border-box: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, white)));\n",
              "    --sklearn-color-icon: #878787;\n",
              "  }\n",
              "}\n",
              "\n",
              "#sk-container-id-8 {\n",
              "  color: var(--sklearn-color-text);\n",
              "}\n",
              "\n",
              "#sk-container-id-8 pre {\n",
              "  padding: 0;\n",
              "}\n",
              "\n",
              "#sk-container-id-8 input.sk-hidden--visually {\n",
              "  border: 0;\n",
              "  clip: rect(1px 1px 1px 1px);\n",
              "  clip: rect(1px, 1px, 1px, 1px);\n",
              "  height: 1px;\n",
              "  margin: -1px;\n",
              "  overflow: hidden;\n",
              "  padding: 0;\n",
              "  position: absolute;\n",
              "  width: 1px;\n",
              "}\n",
              "\n",
              "#sk-container-id-8 div.sk-dashed-wrapped {\n",
              "  border: 1px dashed var(--sklearn-color-line);\n",
              "  margin: 0 0.4em 0.5em 0.4em;\n",
              "  box-sizing: border-box;\n",
              "  padding-bottom: 0.4em;\n",
              "  background-color: var(--sklearn-color-background);\n",
              "}\n",
              "\n",
              "#sk-container-id-8 div.sk-container {\n",
              "  /* jupyter's `normalize.less` sets `[hidden] { display: none; }`\n",
              "     but bootstrap.min.css set `[hidden] { display: none !important; }`\n",
              "     so we also need the `!important` here to be able to override the\n",
              "     default hidden behavior on the sphinx rendered scikit-learn.org.\n",
              "     See: https://github.com/scikit-learn/scikit-learn/issues/21755 */\n",
              "  display: inline-block !important;\n",
              "  position: relative;\n",
              "}\n",
              "\n",
              "#sk-container-id-8 div.sk-text-repr-fallback {\n",
              "  display: none;\n",
              "}\n",
              "\n",
              "div.sk-parallel-item,\n",
              "div.sk-serial,\n",
              "div.sk-item {\n",
              "  /* draw centered vertical line to link estimators */\n",
              "  background-image: linear-gradient(var(--sklearn-color-text-on-default-background), var(--sklearn-color-text-on-default-background));\n",
              "  background-size: 2px 100%;\n",
              "  background-repeat: no-repeat;\n",
              "  background-position: center center;\n",
              "}\n",
              "\n",
              "/* Parallel-specific style estimator block */\n",
              "\n",
              "#sk-container-id-8 div.sk-parallel-item::after {\n",
              "  content: \"\";\n",
              "  width: 100%;\n",
              "  border-bottom: 2px solid var(--sklearn-color-text-on-default-background);\n",
              "  flex-grow: 1;\n",
              "}\n",
              "\n",
              "#sk-container-id-8 div.sk-parallel {\n",
              "  display: flex;\n",
              "  align-items: stretch;\n",
              "  justify-content: center;\n",
              "  background-color: var(--sklearn-color-background);\n",
              "  position: relative;\n",
              "}\n",
              "\n",
              "#sk-container-id-8 div.sk-parallel-item {\n",
              "  display: flex;\n",
              "  flex-direction: column;\n",
              "}\n",
              "\n",
              "#sk-container-id-8 div.sk-parallel-item:first-child::after {\n",
              "  align-self: flex-end;\n",
              "  width: 50%;\n",
              "}\n",
              "\n",
              "#sk-container-id-8 div.sk-parallel-item:last-child::after {\n",
              "  align-self: flex-start;\n",
              "  width: 50%;\n",
              "}\n",
              "\n",
              "#sk-container-id-8 div.sk-parallel-item:only-child::after {\n",
              "  width: 0;\n",
              "}\n",
              "\n",
              "/* Serial-specific style estimator block */\n",
              "\n",
              "#sk-container-id-8 div.sk-serial {\n",
              "  display: flex;\n",
              "  flex-direction: column;\n",
              "  align-items: center;\n",
              "  background-color: var(--sklearn-color-background);\n",
              "  padding-right: 1em;\n",
              "  padding-left: 1em;\n",
              "}\n",
              "\n",
              "\n",
              "/* Toggleable style: style used for estimator/Pipeline/ColumnTransformer box that is\n",
              "clickable and can be expanded/collapsed.\n",
              "- Pipeline and ColumnTransformer use this feature and define the default style\n",
              "- Estimators will overwrite some part of the style using the `sk-estimator` class\n",
              "*/\n",
              "\n",
              "/* Pipeline and ColumnTransformer style (default) */\n",
              "\n",
              "#sk-container-id-8 div.sk-toggleable {\n",
              "  /* Default theme specific background. It is overwritten whether we have a\n",
              "  specific estimator or a Pipeline/ColumnTransformer */\n",
              "  background-color: var(--sklearn-color-background);\n",
              "}\n",
              "\n",
              "/* Toggleable label */\n",
              "#sk-container-id-8 label.sk-toggleable__label {\n",
              "  cursor: pointer;\n",
              "  display: block;\n",
              "  width: 100%;\n",
              "  margin-bottom: 0;\n",
              "  padding: 0.5em;\n",
              "  box-sizing: border-box;\n",
              "  text-align: center;\n",
              "}\n",
              "\n",
              "#sk-container-id-8 label.sk-toggleable__label-arrow:before {\n",
              "  /* Arrow on the left of the label */\n",
              "  content: \"▸\";\n",
              "  float: left;\n",
              "  margin-right: 0.25em;\n",
              "  color: var(--sklearn-color-icon);\n",
              "}\n",
              "\n",
              "#sk-container-id-8 label.sk-toggleable__label-arrow:hover:before {\n",
              "  color: var(--sklearn-color-text);\n",
              "}\n",
              "\n",
              "/* Toggleable content - dropdown */\n",
              "\n",
              "#sk-container-id-8 div.sk-toggleable__content {\n",
              "  max-height: 0;\n",
              "  max-width: 0;\n",
              "  overflow: hidden;\n",
              "  text-align: left;\n",
              "  /* unfitted */\n",
              "  background-color: var(--sklearn-color-unfitted-level-0);\n",
              "}\n",
              "\n",
              "#sk-container-id-8 div.sk-toggleable__content.fitted {\n",
              "  /* fitted */\n",
              "  background-color: var(--sklearn-color-fitted-level-0);\n",
              "}\n",
              "\n",
              "#sk-container-id-8 div.sk-toggleable__content pre {\n",
              "  margin: 0.2em;\n",
              "  border-radius: 0.25em;\n",
              "  color: var(--sklearn-color-text);\n",
              "  /* unfitted */\n",
              "  background-color: var(--sklearn-color-unfitted-level-0);\n",
              "}\n",
              "\n",
              "#sk-container-id-8 div.sk-toggleable__content.fitted pre {\n",
              "  /* unfitted */\n",
              "  background-color: var(--sklearn-color-fitted-level-0);\n",
              "}\n",
              "\n",
              "#sk-container-id-8 input.sk-toggleable__control:checked~div.sk-toggleable__content {\n",
              "  /* Expand drop-down */\n",
              "  max-height: 200px;\n",
              "  max-width: 100%;\n",
              "  overflow: auto;\n",
              "}\n",
              "\n",
              "#sk-container-id-8 input.sk-toggleable__control:checked~label.sk-toggleable__label-arrow:before {\n",
              "  content: \"▾\";\n",
              "}\n",
              "\n",
              "/* Pipeline/ColumnTransformer-specific style */\n",
              "\n",
              "#sk-container-id-8 div.sk-label input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
              "  color: var(--sklearn-color-text);\n",
              "  background-color: var(--sklearn-color-unfitted-level-2);\n",
              "}\n",
              "\n",
              "#sk-container-id-8 div.sk-label.fitted input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
              "  background-color: var(--sklearn-color-fitted-level-2);\n",
              "}\n",
              "\n",
              "/* Estimator-specific style */\n",
              "\n",
              "/* Colorize estimator box */\n",
              "#sk-container-id-8 div.sk-estimator input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
              "  /* unfitted */\n",
              "  background-color: var(--sklearn-color-unfitted-level-2);\n",
              "}\n",
              "\n",
              "#sk-container-id-8 div.sk-estimator.fitted input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
              "  /* fitted */\n",
              "  background-color: var(--sklearn-color-fitted-level-2);\n",
              "}\n",
              "\n",
              "#sk-container-id-8 div.sk-label label.sk-toggleable__label,\n",
              "#sk-container-id-8 div.sk-label label {\n",
              "  /* The background is the default theme color */\n",
              "  color: var(--sklearn-color-text-on-default-background);\n",
              "}\n",
              "\n",
              "/* On hover, darken the color of the background */\n",
              "#sk-container-id-8 div.sk-label:hover label.sk-toggleable__label {\n",
              "  color: var(--sklearn-color-text);\n",
              "  background-color: var(--sklearn-color-unfitted-level-2);\n",
              "}\n",
              "\n",
              "/* Label box, darken color on hover, fitted */\n",
              "#sk-container-id-8 div.sk-label.fitted:hover label.sk-toggleable__label.fitted {\n",
              "  color: var(--sklearn-color-text);\n",
              "  background-color: var(--sklearn-color-fitted-level-2);\n",
              "}\n",
              "\n",
              "/* Estimator label */\n",
              "\n",
              "#sk-container-id-8 div.sk-label label {\n",
              "  font-family: monospace;\n",
              "  font-weight: bold;\n",
              "  display: inline-block;\n",
              "  line-height: 1.2em;\n",
              "}\n",
              "\n",
              "#sk-container-id-8 div.sk-label-container {\n",
              "  text-align: center;\n",
              "}\n",
              "\n",
              "/* Estimator-specific */\n",
              "#sk-container-id-8 div.sk-estimator {\n",
              "  font-family: monospace;\n",
              "  border: 1px dotted var(--sklearn-color-border-box);\n",
              "  border-radius: 0.25em;\n",
              "  box-sizing: border-box;\n",
              "  margin-bottom: 0.5em;\n",
              "  /* unfitted */\n",
              "  background-color: var(--sklearn-color-unfitted-level-0);\n",
              "}\n",
              "\n",
              "#sk-container-id-8 div.sk-estimator.fitted {\n",
              "  /* fitted */\n",
              "  background-color: var(--sklearn-color-fitted-level-0);\n",
              "}\n",
              "\n",
              "/* on hover */\n",
              "#sk-container-id-8 div.sk-estimator:hover {\n",
              "  /* unfitted */\n",
              "  background-color: var(--sklearn-color-unfitted-level-2);\n",
              "}\n",
              "\n",
              "#sk-container-id-8 div.sk-estimator.fitted:hover {\n",
              "  /* fitted */\n",
              "  background-color: var(--sklearn-color-fitted-level-2);\n",
              "}\n",
              "\n",
              "/* Specification for estimator info (e.g. \"i\" and \"?\") */\n",
              "\n",
              "/* Common style for \"i\" and \"?\" */\n",
              "\n",
              ".sk-estimator-doc-link,\n",
              "a:link.sk-estimator-doc-link,\n",
              "a:visited.sk-estimator-doc-link {\n",
              "  float: right;\n",
              "  font-size: smaller;\n",
              "  line-height: 1em;\n",
              "  font-family: monospace;\n",
              "  background-color: var(--sklearn-color-background);\n",
              "  border-radius: 1em;\n",
              "  height: 1em;\n",
              "  width: 1em;\n",
              "  text-decoration: none !important;\n",
              "  margin-left: 1ex;\n",
              "  /* unfitted */\n",
              "  border: var(--sklearn-color-unfitted-level-1) 1pt solid;\n",
              "  color: var(--sklearn-color-unfitted-level-1);\n",
              "}\n",
              "\n",
              ".sk-estimator-doc-link.fitted,\n",
              "a:link.sk-estimator-doc-link.fitted,\n",
              "a:visited.sk-estimator-doc-link.fitted {\n",
              "  /* fitted */\n",
              "  border: var(--sklearn-color-fitted-level-1) 1pt solid;\n",
              "  color: var(--sklearn-color-fitted-level-1);\n",
              "}\n",
              "\n",
              "/* On hover */\n",
              "div.sk-estimator:hover .sk-estimator-doc-link:hover,\n",
              ".sk-estimator-doc-link:hover,\n",
              "div.sk-label-container:hover .sk-estimator-doc-link:hover,\n",
              ".sk-estimator-doc-link:hover {\n",
              "  /* unfitted */\n",
              "  background-color: var(--sklearn-color-unfitted-level-3);\n",
              "  color: var(--sklearn-color-background);\n",
              "  text-decoration: none;\n",
              "}\n",
              "\n",
              "div.sk-estimator.fitted:hover .sk-estimator-doc-link.fitted:hover,\n",
              ".sk-estimator-doc-link.fitted:hover,\n",
              "div.sk-label-container:hover .sk-estimator-doc-link.fitted:hover,\n",
              ".sk-estimator-doc-link.fitted:hover {\n",
              "  /* fitted */\n",
              "  background-color: var(--sklearn-color-fitted-level-3);\n",
              "  color: var(--sklearn-color-background);\n",
              "  text-decoration: none;\n",
              "}\n",
              "\n",
              "/* Span, style for the box shown on hovering the info icon */\n",
              ".sk-estimator-doc-link span {\n",
              "  display: none;\n",
              "  z-index: 9999;\n",
              "  position: relative;\n",
              "  font-weight: normal;\n",
              "  right: .2ex;\n",
              "  padding: .5ex;\n",
              "  margin: .5ex;\n",
              "  width: min-content;\n",
              "  min-width: 20ex;\n",
              "  max-width: 50ex;\n",
              "  color: var(--sklearn-color-text);\n",
              "  box-shadow: 2pt 2pt 4pt #999;\n",
              "  /* unfitted */\n",
              "  background: var(--sklearn-color-unfitted-level-0);\n",
              "  border: .5pt solid var(--sklearn-color-unfitted-level-3);\n",
              "}\n",
              "\n",
              ".sk-estimator-doc-link.fitted span {\n",
              "  /* fitted */\n",
              "  background: var(--sklearn-color-fitted-level-0);\n",
              "  border: var(--sklearn-color-fitted-level-3);\n",
              "}\n",
              "\n",
              ".sk-estimator-doc-link:hover span {\n",
              "  display: block;\n",
              "}\n",
              "\n",
              "/* \"?\"-specific style due to the `<a>` HTML tag */\n",
              "\n",
              "#sk-container-id-8 a.estimator_doc_link {\n",
              "  float: right;\n",
              "  font-size: 1rem;\n",
              "  line-height: 1em;\n",
              "  font-family: monospace;\n",
              "  background-color: var(--sklearn-color-background);\n",
              "  border-radius: 1rem;\n",
              "  height: 1rem;\n",
              "  width: 1rem;\n",
              "  text-decoration: none;\n",
              "  /* unfitted */\n",
              "  color: var(--sklearn-color-unfitted-level-1);\n",
              "  border: var(--sklearn-color-unfitted-level-1) 1pt solid;\n",
              "}\n",
              "\n",
              "#sk-container-id-8 a.estimator_doc_link.fitted {\n",
              "  /* fitted */\n",
              "  border: var(--sklearn-color-fitted-level-1) 1pt solid;\n",
              "  color: var(--sklearn-color-fitted-level-1);\n",
              "}\n",
              "\n",
              "/* On hover */\n",
              "#sk-container-id-8 a.estimator_doc_link:hover {\n",
              "  /* unfitted */\n",
              "  background-color: var(--sklearn-color-unfitted-level-3);\n",
              "  color: var(--sklearn-color-background);\n",
              "  text-decoration: none;\n",
              "}\n",
              "\n",
              "#sk-container-id-8 a.estimator_doc_link.fitted:hover {\n",
              "  /* fitted */\n",
              "  background-color: var(--sklearn-color-fitted-level-3);\n",
              "}\n",
              "</style><div id=\"sk-container-id-8\" class=\"sk-top-container\"><div class=\"sk-text-repr-fallback\"><pre>Pipeline(steps=[(&#x27;columntransformer&#x27;,\n",
              "                 ColumnTransformer(remainder=Pipeline(steps=[(&#x27;standardscaler&#x27;,\n",
              "                                                              StandardScaler())]),\n",
              "                                   transformers=[(&#x27;monthly_credit_burden&#x27;,\n",
              "                                                  Pipeline(steps=[(&#x27;functiontransformer-1&#x27;,\n",
              "                                                                   FunctionTransformer(feature_names_out=&lt;function ratio_name at 0x78a662f00dc0&gt;,\n",
              "                                                                                       func=&lt;function ratio_pipeline.&lt;locals&gt;.&lt;lambda&gt; at 0x78a6624f3250&gt;)),\n",
              "                                                                  (&#x27;function...\n",
              "                                                   &#x27;personal_status&#x27;,\n",
              "                                                   &#x27;credit_history&#x27;]),\n",
              "                                                 (&#x27;geo&#x27;,\n",
              "                                                  Pipeline(steps=[(&#x27;ordered_ordinal_encoder&#x27;,\n",
              "                                                                   OrderedOrdinalEncoder()),\n",
              "                                                                  (&#x27;cluster_simil&#x27;,\n",
              "                                                                   ClusterSimilarity(random_state=42))]),\n",
              "                                                  [&#x27;savings_status&#x27;,\n",
              "                                                   &#x27;checking_status&#x27;]),\n",
              "                                                 (&#x27;drop_out&#x27;, &#x27;drop&#x27;,\n",
              "                                                  [&#x27;num_dependents&#x27;,\n",
              "                                                   &#x27;foreign_worker&#x27;,\n",
              "                                                   &#x27;purpose&#x27;])])),\n",
              "                (&#x27;randomforestclassifier&#x27;,\n",
              "                 RandomForestClassifier(class_weight=&#x27;balanced&#x27;,\n",
              "                                        random_state=42))])</pre><b>In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook. <br />On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.</b></div><div class=\"sk-container\" hidden><div class=\"sk-item sk-dashed-wrapped\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-204\" type=\"checkbox\" ><label for=\"sk-estimator-id-204\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;&nbsp;Pipeline<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.pipeline.Pipeline.html\">?<span>Documentation for Pipeline</span></a><span class=\"sk-estimator-doc-link fitted\">i<span>Fitted</span></span></label><div class=\"sk-toggleable__content fitted\"><pre>Pipeline(steps=[(&#x27;columntransformer&#x27;,\n",
              "                 ColumnTransformer(remainder=Pipeline(steps=[(&#x27;standardscaler&#x27;,\n",
              "                                                              StandardScaler())]),\n",
              "                                   transformers=[(&#x27;monthly_credit_burden&#x27;,\n",
              "                                                  Pipeline(steps=[(&#x27;functiontransformer-1&#x27;,\n",
              "                                                                   FunctionTransformer(feature_names_out=&lt;function ratio_name at 0x78a662f00dc0&gt;,\n",
              "                                                                                       func=&lt;function ratio_pipeline.&lt;locals&gt;.&lt;lambda&gt; at 0x78a6624f3250&gt;)),\n",
              "                                                                  (&#x27;function...\n",
              "                                                   &#x27;personal_status&#x27;,\n",
              "                                                   &#x27;credit_history&#x27;]),\n",
              "                                                 (&#x27;geo&#x27;,\n",
              "                                                  Pipeline(steps=[(&#x27;ordered_ordinal_encoder&#x27;,\n",
              "                                                                   OrderedOrdinalEncoder()),\n",
              "                                                                  (&#x27;cluster_simil&#x27;,\n",
              "                                                                   ClusterSimilarity(random_state=42))]),\n",
              "                                                  [&#x27;savings_status&#x27;,\n",
              "                                                   &#x27;checking_status&#x27;]),\n",
              "                                                 (&#x27;drop_out&#x27;, &#x27;drop&#x27;,\n",
              "                                                  [&#x27;num_dependents&#x27;,\n",
              "                                                   &#x27;foreign_worker&#x27;,\n",
              "                                                   &#x27;purpose&#x27;])])),\n",
              "                (&#x27;randomforestclassifier&#x27;,\n",
              "                 RandomForestClassifier(class_weight=&#x27;balanced&#x27;,\n",
              "                                        random_state=42))])</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item sk-dashed-wrapped\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-205\" type=\"checkbox\" ><label for=\"sk-estimator-id-205\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;columntransformer: ColumnTransformer<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.compose.ColumnTransformer.html\">?<span>Documentation for columntransformer: ColumnTransformer</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>ColumnTransformer(remainder=Pipeline(steps=[(&#x27;standardscaler&#x27;,\n",
              "                                             StandardScaler())]),\n",
              "                  transformers=[(&#x27;monthly_credit_burden&#x27;,\n",
              "                                 Pipeline(steps=[(&#x27;functiontransformer-1&#x27;,\n",
              "                                                  FunctionTransformer(feature_names_out=&lt;function ratio_name at 0x78a662f00dc0&gt;,\n",
              "                                                                      func=&lt;function ratio_pipeline.&lt;locals&gt;.&lt;lambda&gt; at 0x78a6624f3250&gt;)),\n",
              "                                                 (&#x27;functiontransformer-2&#x27;,\n",
              "                                                  FunctionTransformer(fe...\n",
              "                                                  OneHotEncoder(handle_unknown=&#x27;ignore&#x27;))]),\n",
              "                                 [&#x27;job&#x27;, &#x27;housing&#x27;, &#x27;other_payment_plans&#x27;,\n",
              "                                  &#x27;property_magnitude&#x27;, &#x27;other_parties&#x27;,\n",
              "                                  &#x27;personal_status&#x27;, &#x27;credit_history&#x27;]),\n",
              "                                (&#x27;geo&#x27;,\n",
              "                                 Pipeline(steps=[(&#x27;ordered_ordinal_encoder&#x27;,\n",
              "                                                  OrderedOrdinalEncoder()),\n",
              "                                                 (&#x27;cluster_simil&#x27;,\n",
              "                                                  ClusterSimilarity(random_state=42))]),\n",
              "                                 [&#x27;savings_status&#x27;, &#x27;checking_status&#x27;]),\n",
              "                                (&#x27;drop_out&#x27;, &#x27;drop&#x27;,\n",
              "                                 [&#x27;num_dependents&#x27;, &#x27;foreign_worker&#x27;,\n",
              "                                  &#x27;purpose&#x27;])])</pre></div> </div></div><div class=\"sk-parallel\"><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-206\" type=\"checkbox\" ><label for=\"sk-estimator-id-206\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">monthly_credit_burden</label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;credit_amount&#x27;, &#x27;duration&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-207\" type=\"checkbox\" ><label for=\"sk-estimator-id-207\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;FunctionTransformer<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.FunctionTransformer.html\">?<span>Documentation for FunctionTransformer</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>FunctionTransformer(feature_names_out=&lt;function ratio_name at 0x78a662f00dc0&gt;,\n",
              "                    func=&lt;function ratio_pipeline.&lt;locals&gt;.&lt;lambda&gt; at 0x78a6624f3250&gt;)</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-208\" type=\"checkbox\" ><label for=\"sk-estimator-id-208\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;FunctionTransformer<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.FunctionTransformer.html\">?<span>Documentation for FunctionTransformer</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>FunctionTransformer(feature_names_out=&#x27;one-to-one&#x27;, func=&lt;ufunc &#x27;log&#x27;&gt;,\n",
              "                    inverse_func=&lt;ufunc &#x27;exp&#x27;&gt;)</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-209\" type=\"checkbox\" ><label for=\"sk-estimator-id-209\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;StandardScaler<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.StandardScaler.html\">?<span>Documentation for StandardScaler</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>StandardScaler()</pre></div> </div></div></div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-210\" type=\"checkbox\" ><label for=\"sk-estimator-id-210\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">installment_per_credit</label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;installment_commitment&#x27;, &#x27;credit_amount&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-211\" type=\"checkbox\" ><label for=\"sk-estimator-id-211\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;FunctionTransformer<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.FunctionTransformer.html\">?<span>Documentation for FunctionTransformer</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>FunctionTransformer(feature_names_out=&lt;function ratio_name at 0x78a662f00dc0&gt;,\n",
              "                    func=&lt;function ratio_pipeline.&lt;locals&gt;.&lt;lambda&gt; at 0x78a6624f3400&gt;)</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-212\" type=\"checkbox\" ><label for=\"sk-estimator-id-212\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;FunctionTransformer<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.FunctionTransformer.html\">?<span>Documentation for FunctionTransformer</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>FunctionTransformer(feature_names_out=&#x27;one-to-one&#x27;, func=&lt;ufunc &#x27;log&#x27;&gt;,\n",
              "                    inverse_func=&lt;ufunc &#x27;exp&#x27;&gt;)</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-213\" type=\"checkbox\" ><label for=\"sk-estimator-id-213\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;StandardScaler<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.StandardScaler.html\">?<span>Documentation for StandardScaler</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>StandardScaler()</pre></div> </div></div></div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-214\" type=\"checkbox\" ><label for=\"sk-estimator-id-214\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">dependents_per_credit</label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;num_dependents&#x27;, &#x27;credit_amount&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-215\" type=\"checkbox\" ><label for=\"sk-estimator-id-215\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;FunctionTransformer<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.FunctionTransformer.html\">?<span>Documentation for FunctionTransformer</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>FunctionTransformer(feature_names_out=&lt;function ratio_name at 0x78a662f00dc0&gt;,\n",
              "                    func=&lt;function ratio_pipeline.&lt;locals&gt;.&lt;lambda&gt; at 0x78a6624f35b0&gt;)</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-216\" type=\"checkbox\" ><label for=\"sk-estimator-id-216\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;FunctionTransformer<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.FunctionTransformer.html\">?<span>Documentation for FunctionTransformer</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>FunctionTransformer(feature_names_out=&#x27;one-to-one&#x27;, func=&lt;ufunc &#x27;log&#x27;&gt;,\n",
              "                    inverse_func=&lt;ufunc &#x27;exp&#x27;&gt;)</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-217\" type=\"checkbox\" ><label for=\"sk-estimator-id-217\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;StandardScaler<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.StandardScaler.html\">?<span>Documentation for StandardScaler</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>StandardScaler()</pre></div> </div></div></div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-218\" type=\"checkbox\" ><label for=\"sk-estimator-id-218\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">log</label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;duration&#x27;, &#x27;credit_amount&#x27;, &#x27;installment_commitment&#x27;, &#x27;residence_since&#x27;, &#x27;age&#x27;, &#x27;existing_credits&#x27;, &#x27;num_dependents&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-219\" type=\"checkbox\" ><label for=\"sk-estimator-id-219\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;FunctionTransformer<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.FunctionTransformer.html\">?<span>Documentation for FunctionTransformer</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>FunctionTransformer(feature_names_out=&#x27;one-to-one&#x27;, func=&lt;ufunc &#x27;log&#x27;&gt;,\n",
              "                    inverse_func=&lt;ufunc &#x27;exp&#x27;&gt;)</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-220\" type=\"checkbox\" ><label for=\"sk-estimator-id-220\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;StandardScaler<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.StandardScaler.html\">?<span>Documentation for StandardScaler</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>StandardScaler()</pre></div> </div></div></div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-221\" type=\"checkbox\" ><label for=\"sk-estimator-id-221\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">cat_ordinal</label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;checking_status&#x27;, &#x27;savings_status&#x27;, &#x27;employment&#x27;, &#x27;own_telephone&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-222\" type=\"checkbox\" ><label for=\"sk-estimator-id-222\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">OrderedOrdinalEncoder</label><div class=\"sk-toggleable__content fitted\"><pre>OrderedOrdinalEncoder()</pre></div> </div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-223\" type=\"checkbox\" ><label for=\"sk-estimator-id-223\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">cat_onehot</label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;job&#x27;, &#x27;housing&#x27;, &#x27;other_payment_plans&#x27;, &#x27;property_magnitude&#x27;, &#x27;other_parties&#x27;, &#x27;personal_status&#x27;, &#x27;credit_history&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-224\" type=\"checkbox\" ><label for=\"sk-estimator-id-224\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;OneHotEncoder<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.OneHotEncoder.html\">?<span>Documentation for OneHotEncoder</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>OneHotEncoder(handle_unknown=&#x27;ignore&#x27;)</pre></div> </div></div></div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-225\" type=\"checkbox\" ><label for=\"sk-estimator-id-225\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">geo</label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;savings_status&#x27;, &#x27;checking_status&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-226\" type=\"checkbox\" ><label for=\"sk-estimator-id-226\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">OrderedOrdinalEncoder</label><div class=\"sk-toggleable__content fitted\"><pre>OrderedOrdinalEncoder()</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-227\" type=\"checkbox\" ><label for=\"sk-estimator-id-227\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">ClusterSimilarity</label><div class=\"sk-toggleable__content fitted\"><pre>ClusterSimilarity(random_state=42)</pre></div> </div></div></div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-228\" type=\"checkbox\" ><label for=\"sk-estimator-id-228\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">drop_out</label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;num_dependents&#x27;, &#x27;foreign_worker&#x27;, &#x27;purpose&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-229\" type=\"checkbox\" ><label for=\"sk-estimator-id-229\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">drop</label><div class=\"sk-toggleable__content fitted\"><pre>drop</pre></div> </div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-230\" type=\"checkbox\" ><label for=\"sk-estimator-id-230\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">remainder</label><div class=\"sk-toggleable__content fitted\"><pre>[]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-231\" type=\"checkbox\" ><label for=\"sk-estimator-id-231\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;StandardScaler<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.StandardScaler.html\">?<span>Documentation for StandardScaler</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>StandardScaler()</pre></div> </div></div></div></div></div></div></div></div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-232\" type=\"checkbox\" ><label for=\"sk-estimator-id-232\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;RandomForestClassifier<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.ensemble.RandomForestClassifier.html\">?<span>Documentation for RandomForestClassifier</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>RandomForestClassifier(class_weight=&#x27;balanced&#x27;, random_state=42)</pre></div> </div></div></div></div></div></div>"
            ]
          },
          "metadata": {},
          "execution_count": 72
        }
      ],
      "source": [
        "from sklearn.ensemble import RandomForestClassifier\n",
        "from sklearn.pipeline import make_pipeline\n",
        "\n",
        "random_forest = make_pipeline(preprocessing, RandomForestClassifier(random_state=42, class_weight='balanced'))\n",
        "random_forest.fit(X_train, y_train)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "XNN4g5Vuo0av"
      },
      "source": [
        "Testing the accuracy of the model:"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 76,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "FK1kx17Mo5Ok",
        "outputId": "3323c606-0adf-4e7d-ad1f-1e3a736676b6"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Training Set Performance:\n",
            "Accuracy: 1.0000\n",
            "Precision: 1.0000\n",
            "Recall: 1.0000\n",
            "F1-Score: 1.0000\n",
            "ROC-AUC: 1.0000\n",
            "------------------------------\n"
          ]
        }
      ],
      "source": [
        "from sklearn.metrics import accuracy_score, precision_score, recall_score, f1_score, roc_auc_score\n",
        "\n",
        "# Make predictions\n",
        "train_predictions = random_forest.predict(X_train)\n",
        "\n",
        "# Evaluation metrics\n",
        "train_accuracy = accuracy_score(y_train, train_predictions)\n",
        "train_precision = precision_score(y_train, train_predictions, pos_label='good')\n",
        "train_recall = recall_score(y_train, train_predictions, pos_label='good')\n",
        "train_f1 = f1_score(y_train, train_predictions, pos_label='good')\n",
        "train_roc_auc = roc_auc_score(y_train, random_forest.predict_proba(X_train)[:, 1])\n",
        "\n",
        "# Print training set results\n",
        "print(\"Training Set Performance:\")\n",
        "print(f\"Accuracy: {train_accuracy:.4f}\")\n",
        "print(f\"Precision: {train_precision:.4f}\")\n",
        "print(f\"Recall: {train_recall:.4f}\")\n",
        "print(f\"F1-Score: {train_f1:.4f}\")\n",
        "print(f\"ROC-AUC: {train_roc_auc:.4f}\")\n",
        "print(\"-\" * 30)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "pnwlaWfHvDyI"
      },
      "source": [
        "### Gradient Boosting"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Ro-dfWD5sFl4"
      },
      "source": [
        "We are now going to be training a random forest model."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 82,
      "metadata": {
        "id": "BDc7pxTKvGiH",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 357
        },
        "outputId": "76aa7fd9-5084-47b7-8eff-07df9f7ce12a"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Pipeline(steps=[('columntransformer',\n",
              "                 ColumnTransformer(remainder=Pipeline(steps=[('standardscaler',\n",
              "                                                              StandardScaler())]),\n",
              "                                   transformers=[('monthly_credit_burden',\n",
              "                                                  Pipeline(steps=[('functiontransformer-1',\n",
              "                                                                   FunctionTransformer(feature_names_out=<function ratio_name at 0x78a662f00dc0>,\n",
              "                                                                                       func=<function ratio_pipeline.<locals>.<lambda> at 0x78a6624f3250>)),\n",
              "                                                                  ('function...\n",
              "                                                   'personal_status',\n",
              "                                                   'credit_history']),\n",
              "                                                 ('geo',\n",
              "                                                  Pipeline(steps=[('ordered_ordinal_encoder',\n",
              "                                                                   OrderedOrdinalEncoder()),\n",
              "                                                                  ('cluster_simil',\n",
              "                                                                   ClusterSimilarity(random_state=42))]),\n",
              "                                                  ['savings_status',\n",
              "                                                   'checking_status']),\n",
              "                                                 ('drop_out', 'drop',\n",
              "                                                  ['num_dependents',\n",
              "                                                   'foreign_worker',\n",
              "                                                   'purpose'])])),\n",
              "                ('gradientboostingclassifier',\n",
              "                 GradientBoostingClassifier(random_state=42))])"
            ],
            "text/html": [
              "<style>#sk-container-id-12 {\n",
              "  /* Definition of color scheme common for light and dark mode */\n",
              "  --sklearn-color-text: black;\n",
              "  --sklearn-color-line: gray;\n",
              "  /* Definition of color scheme for unfitted estimators */\n",
              "  --sklearn-color-unfitted-level-0: #fff5e6;\n",
              "  --sklearn-color-unfitted-level-1: #f6e4d2;\n",
              "  --sklearn-color-unfitted-level-2: #ffe0b3;\n",
              "  --sklearn-color-unfitted-level-3: chocolate;\n",
              "  /* Definition of color scheme for fitted estimators */\n",
              "  --sklearn-color-fitted-level-0: #f0f8ff;\n",
              "  --sklearn-color-fitted-level-1: #d4ebff;\n",
              "  --sklearn-color-fitted-level-2: #b3dbfd;\n",
              "  --sklearn-color-fitted-level-3: cornflowerblue;\n",
              "\n",
              "  /* Specific color for light theme */\n",
              "  --sklearn-color-text-on-default-background: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, black)));\n",
              "  --sklearn-color-background: var(--sg-background-color, var(--theme-background, var(--jp-layout-color0, white)));\n",
              "  --sklearn-color-border-box: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, black)));\n",
              "  --sklearn-color-icon: #696969;\n",
              "\n",
              "  @media (prefers-color-scheme: dark) {\n",
              "    /* Redefinition of color scheme for dark theme */\n",
              "    --sklearn-color-text-on-default-background: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, white)));\n",
              "    --sklearn-color-background: var(--sg-background-color, var(--theme-background, var(--jp-layout-color0, #111)));\n",
              "    --sklearn-color-border-box: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, white)));\n",
              "    --sklearn-color-icon: #878787;\n",
              "  }\n",
              "}\n",
              "\n",
              "#sk-container-id-12 {\n",
              "  color: var(--sklearn-color-text);\n",
              "}\n",
              "\n",
              "#sk-container-id-12 pre {\n",
              "  padding: 0;\n",
              "}\n",
              "\n",
              "#sk-container-id-12 input.sk-hidden--visually {\n",
              "  border: 0;\n",
              "  clip: rect(1px 1px 1px 1px);\n",
              "  clip: rect(1px, 1px, 1px, 1px);\n",
              "  height: 1px;\n",
              "  margin: -1px;\n",
              "  overflow: hidden;\n",
              "  padding: 0;\n",
              "  position: absolute;\n",
              "  width: 1px;\n",
              "}\n",
              "\n",
              "#sk-container-id-12 div.sk-dashed-wrapped {\n",
              "  border: 1px dashed var(--sklearn-color-line);\n",
              "  margin: 0 0.4em 0.5em 0.4em;\n",
              "  box-sizing: border-box;\n",
              "  padding-bottom: 0.4em;\n",
              "  background-color: var(--sklearn-color-background);\n",
              "}\n",
              "\n",
              "#sk-container-id-12 div.sk-container {\n",
              "  /* jupyter's `normalize.less` sets `[hidden] { display: none; }`\n",
              "     but bootstrap.min.css set `[hidden] { display: none !important; }`\n",
              "     so we also need the `!important` here to be able to override the\n",
              "     default hidden behavior on the sphinx rendered scikit-learn.org.\n",
              "     See: https://github.com/scikit-learn/scikit-learn/issues/21755 */\n",
              "  display: inline-block !important;\n",
              "  position: relative;\n",
              "}\n",
              "\n",
              "#sk-container-id-12 div.sk-text-repr-fallback {\n",
              "  display: none;\n",
              "}\n",
              "\n",
              "div.sk-parallel-item,\n",
              "div.sk-serial,\n",
              "div.sk-item {\n",
              "  /* draw centered vertical line to link estimators */\n",
              "  background-image: linear-gradient(var(--sklearn-color-text-on-default-background), var(--sklearn-color-text-on-default-background));\n",
              "  background-size: 2px 100%;\n",
              "  background-repeat: no-repeat;\n",
              "  background-position: center center;\n",
              "}\n",
              "\n",
              "/* Parallel-specific style estimator block */\n",
              "\n",
              "#sk-container-id-12 div.sk-parallel-item::after {\n",
              "  content: \"\";\n",
              "  width: 100%;\n",
              "  border-bottom: 2px solid var(--sklearn-color-text-on-default-background);\n",
              "  flex-grow: 1;\n",
              "}\n",
              "\n",
              "#sk-container-id-12 div.sk-parallel {\n",
              "  display: flex;\n",
              "  align-items: stretch;\n",
              "  justify-content: center;\n",
              "  background-color: var(--sklearn-color-background);\n",
              "  position: relative;\n",
              "}\n",
              "\n",
              "#sk-container-id-12 div.sk-parallel-item {\n",
              "  display: flex;\n",
              "  flex-direction: column;\n",
              "}\n",
              "\n",
              "#sk-container-id-12 div.sk-parallel-item:first-child::after {\n",
              "  align-self: flex-end;\n",
              "  width: 50%;\n",
              "}\n",
              "\n",
              "#sk-container-id-12 div.sk-parallel-item:last-child::after {\n",
              "  align-self: flex-start;\n",
              "  width: 50%;\n",
              "}\n",
              "\n",
              "#sk-container-id-12 div.sk-parallel-item:only-child::after {\n",
              "  width: 0;\n",
              "}\n",
              "\n",
              "/* Serial-specific style estimator block */\n",
              "\n",
              "#sk-container-id-12 div.sk-serial {\n",
              "  display: flex;\n",
              "  flex-direction: column;\n",
              "  align-items: center;\n",
              "  background-color: var(--sklearn-color-background);\n",
              "  padding-right: 1em;\n",
              "  padding-left: 1em;\n",
              "}\n",
              "\n",
              "\n",
              "/* Toggleable style: style used for estimator/Pipeline/ColumnTransformer box that is\n",
              "clickable and can be expanded/collapsed.\n",
              "- Pipeline and ColumnTransformer use this feature and define the default style\n",
              "- Estimators will overwrite some part of the style using the `sk-estimator` class\n",
              "*/\n",
              "\n",
              "/* Pipeline and ColumnTransformer style (default) */\n",
              "\n",
              "#sk-container-id-12 div.sk-toggleable {\n",
              "  /* Default theme specific background. It is overwritten whether we have a\n",
              "  specific estimator or a Pipeline/ColumnTransformer */\n",
              "  background-color: var(--sklearn-color-background);\n",
              "}\n",
              "\n",
              "/* Toggleable label */\n",
              "#sk-container-id-12 label.sk-toggleable__label {\n",
              "  cursor: pointer;\n",
              "  display: block;\n",
              "  width: 100%;\n",
              "  margin-bottom: 0;\n",
              "  padding: 0.5em;\n",
              "  box-sizing: border-box;\n",
              "  text-align: center;\n",
              "}\n",
              "\n",
              "#sk-container-id-12 label.sk-toggleable__label-arrow:before {\n",
              "  /* Arrow on the left of the label */\n",
              "  content: \"▸\";\n",
              "  float: left;\n",
              "  margin-right: 0.25em;\n",
              "  color: var(--sklearn-color-icon);\n",
              "}\n",
              "\n",
              "#sk-container-id-12 label.sk-toggleable__label-arrow:hover:before {\n",
              "  color: var(--sklearn-color-text);\n",
              "}\n",
              "\n",
              "/* Toggleable content - dropdown */\n",
              "\n",
              "#sk-container-id-12 div.sk-toggleable__content {\n",
              "  max-height: 0;\n",
              "  max-width: 0;\n",
              "  overflow: hidden;\n",
              "  text-align: left;\n",
              "  /* unfitted */\n",
              "  background-color: var(--sklearn-color-unfitted-level-0);\n",
              "}\n",
              "\n",
              "#sk-container-id-12 div.sk-toggleable__content.fitted {\n",
              "  /* fitted */\n",
              "  background-color: var(--sklearn-color-fitted-level-0);\n",
              "}\n",
              "\n",
              "#sk-container-id-12 div.sk-toggleable__content pre {\n",
              "  margin: 0.2em;\n",
              "  border-radius: 0.25em;\n",
              "  color: var(--sklearn-color-text);\n",
              "  /* unfitted */\n",
              "  background-color: var(--sklearn-color-unfitted-level-0);\n",
              "}\n",
              "\n",
              "#sk-container-id-12 div.sk-toggleable__content.fitted pre {\n",
              "  /* unfitted */\n",
              "  background-color: var(--sklearn-color-fitted-level-0);\n",
              "}\n",
              "\n",
              "#sk-container-id-12 input.sk-toggleable__control:checked~div.sk-toggleable__content {\n",
              "  /* Expand drop-down */\n",
              "  max-height: 200px;\n",
              "  max-width: 100%;\n",
              "  overflow: auto;\n",
              "}\n",
              "\n",
              "#sk-container-id-12 input.sk-toggleable__control:checked~label.sk-toggleable__label-arrow:before {\n",
              "  content: \"▾\";\n",
              "}\n",
              "\n",
              "/* Pipeline/ColumnTransformer-specific style */\n",
              "\n",
              "#sk-container-id-12 div.sk-label input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
              "  color: var(--sklearn-color-text);\n",
              "  background-color: var(--sklearn-color-unfitted-level-2);\n",
              "}\n",
              "\n",
              "#sk-container-id-12 div.sk-label.fitted input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
              "  background-color: var(--sklearn-color-fitted-level-2);\n",
              "}\n",
              "\n",
              "/* Estimator-specific style */\n",
              "\n",
              "/* Colorize estimator box */\n",
              "#sk-container-id-12 div.sk-estimator input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
              "  /* unfitted */\n",
              "  background-color: var(--sklearn-color-unfitted-level-2);\n",
              "}\n",
              "\n",
              "#sk-container-id-12 div.sk-estimator.fitted input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
              "  /* fitted */\n",
              "  background-color: var(--sklearn-color-fitted-level-2);\n",
              "}\n",
              "\n",
              "#sk-container-id-12 div.sk-label label.sk-toggleable__label,\n",
              "#sk-container-id-12 div.sk-label label {\n",
              "  /* The background is the default theme color */\n",
              "  color: var(--sklearn-color-text-on-default-background);\n",
              "}\n",
              "\n",
              "/* On hover, darken the color of the background */\n",
              "#sk-container-id-12 div.sk-label:hover label.sk-toggleable__label {\n",
              "  color: var(--sklearn-color-text);\n",
              "  background-color: var(--sklearn-color-unfitted-level-2);\n",
              "}\n",
              "\n",
              "/* Label box, darken color on hover, fitted */\n",
              "#sk-container-id-12 div.sk-label.fitted:hover label.sk-toggleable__label.fitted {\n",
              "  color: var(--sklearn-color-text);\n",
              "  background-color: var(--sklearn-color-fitted-level-2);\n",
              "}\n",
              "\n",
              "/* Estimator label */\n",
              "\n",
              "#sk-container-id-12 div.sk-label label {\n",
              "  font-family: monospace;\n",
              "  font-weight: bold;\n",
              "  display: inline-block;\n",
              "  line-height: 1.2em;\n",
              "}\n",
              "\n",
              "#sk-container-id-12 div.sk-label-container {\n",
              "  text-align: center;\n",
              "}\n",
              "\n",
              "/* Estimator-specific */\n",
              "#sk-container-id-12 div.sk-estimator {\n",
              "  font-family: monospace;\n",
              "  border: 1px dotted var(--sklearn-color-border-box);\n",
              "  border-radius: 0.25em;\n",
              "  box-sizing: border-box;\n",
              "  margin-bottom: 0.5em;\n",
              "  /* unfitted */\n",
              "  background-color: var(--sklearn-color-unfitted-level-0);\n",
              "}\n",
              "\n",
              "#sk-container-id-12 div.sk-estimator.fitted {\n",
              "  /* fitted */\n",
              "  background-color: var(--sklearn-color-fitted-level-0);\n",
              "}\n",
              "\n",
              "/* on hover */\n",
              "#sk-container-id-12 div.sk-estimator:hover {\n",
              "  /* unfitted */\n",
              "  background-color: var(--sklearn-color-unfitted-level-2);\n",
              "}\n",
              "\n",
              "#sk-container-id-12 div.sk-estimator.fitted:hover {\n",
              "  /* fitted */\n",
              "  background-color: var(--sklearn-color-fitted-level-2);\n",
              "}\n",
              "\n",
              "/* Specification for estimator info (e.g. \"i\" and \"?\") */\n",
              "\n",
              "/* Common style for \"i\" and \"?\" */\n",
              "\n",
              ".sk-estimator-doc-link,\n",
              "a:link.sk-estimator-doc-link,\n",
              "a:visited.sk-estimator-doc-link {\n",
              "  float: right;\n",
              "  font-size: smaller;\n",
              "  line-height: 1em;\n",
              "  font-family: monospace;\n",
              "  background-color: var(--sklearn-color-background);\n",
              "  border-radius: 1em;\n",
              "  height: 1em;\n",
              "  width: 1em;\n",
              "  text-decoration: none !important;\n",
              "  margin-left: 1ex;\n",
              "  /* unfitted */\n",
              "  border: var(--sklearn-color-unfitted-level-1) 1pt solid;\n",
              "  color: var(--sklearn-color-unfitted-level-1);\n",
              "}\n",
              "\n",
              ".sk-estimator-doc-link.fitted,\n",
              "a:link.sk-estimator-doc-link.fitted,\n",
              "a:visited.sk-estimator-doc-link.fitted {\n",
              "  /* fitted */\n",
              "  border: var(--sklearn-color-fitted-level-1) 1pt solid;\n",
              "  color: var(--sklearn-color-fitted-level-1);\n",
              "}\n",
              "\n",
              "/* On hover */\n",
              "div.sk-estimator:hover .sk-estimator-doc-link:hover,\n",
              ".sk-estimator-doc-link:hover,\n",
              "div.sk-label-container:hover .sk-estimator-doc-link:hover,\n",
              ".sk-estimator-doc-link:hover {\n",
              "  /* unfitted */\n",
              "  background-color: var(--sklearn-color-unfitted-level-3);\n",
              "  color: var(--sklearn-color-background);\n",
              "  text-decoration: none;\n",
              "}\n",
              "\n",
              "div.sk-estimator.fitted:hover .sk-estimator-doc-link.fitted:hover,\n",
              ".sk-estimator-doc-link.fitted:hover,\n",
              "div.sk-label-container:hover .sk-estimator-doc-link.fitted:hover,\n",
              ".sk-estimator-doc-link.fitted:hover {\n",
              "  /* fitted */\n",
              "  background-color: var(--sklearn-color-fitted-level-3);\n",
              "  color: var(--sklearn-color-background);\n",
              "  text-decoration: none;\n",
              "}\n",
              "\n",
              "/* Span, style for the box shown on hovering the info icon */\n",
              ".sk-estimator-doc-link span {\n",
              "  display: none;\n",
              "  z-index: 9999;\n",
              "  position: relative;\n",
              "  font-weight: normal;\n",
              "  right: .2ex;\n",
              "  padding: .5ex;\n",
              "  margin: .5ex;\n",
              "  width: min-content;\n",
              "  min-width: 20ex;\n",
              "  max-width: 50ex;\n",
              "  color: var(--sklearn-color-text);\n",
              "  box-shadow: 2pt 2pt 4pt #999;\n",
              "  /* unfitted */\n",
              "  background: var(--sklearn-color-unfitted-level-0);\n",
              "  border: .5pt solid var(--sklearn-color-unfitted-level-3);\n",
              "}\n",
              "\n",
              ".sk-estimator-doc-link.fitted span {\n",
              "  /* fitted */\n",
              "  background: var(--sklearn-color-fitted-level-0);\n",
              "  border: var(--sklearn-color-fitted-level-3);\n",
              "}\n",
              "\n",
              ".sk-estimator-doc-link:hover span {\n",
              "  display: block;\n",
              "}\n",
              "\n",
              "/* \"?\"-specific style due to the `<a>` HTML tag */\n",
              "\n",
              "#sk-container-id-12 a.estimator_doc_link {\n",
              "  float: right;\n",
              "  font-size: 1rem;\n",
              "  line-height: 1em;\n",
              "  font-family: monospace;\n",
              "  background-color: var(--sklearn-color-background);\n",
              "  border-radius: 1rem;\n",
              "  height: 1rem;\n",
              "  width: 1rem;\n",
              "  text-decoration: none;\n",
              "  /* unfitted */\n",
              "  color: var(--sklearn-color-unfitted-level-1);\n",
              "  border: var(--sklearn-color-unfitted-level-1) 1pt solid;\n",
              "}\n",
              "\n",
              "#sk-container-id-12 a.estimator_doc_link.fitted {\n",
              "  /* fitted */\n",
              "  border: var(--sklearn-color-fitted-level-1) 1pt solid;\n",
              "  color: var(--sklearn-color-fitted-level-1);\n",
              "}\n",
              "\n",
              "/* On hover */\n",
              "#sk-container-id-12 a.estimator_doc_link:hover {\n",
              "  /* unfitted */\n",
              "  background-color: var(--sklearn-color-unfitted-level-3);\n",
              "  color: var(--sklearn-color-background);\n",
              "  text-decoration: none;\n",
              "}\n",
              "\n",
              "#sk-container-id-12 a.estimator_doc_link.fitted:hover {\n",
              "  /* fitted */\n",
              "  background-color: var(--sklearn-color-fitted-level-3);\n",
              "}\n",
              "</style><div id=\"sk-container-id-12\" class=\"sk-top-container\"><div class=\"sk-text-repr-fallback\"><pre>Pipeline(steps=[(&#x27;columntransformer&#x27;,\n",
              "                 ColumnTransformer(remainder=Pipeline(steps=[(&#x27;standardscaler&#x27;,\n",
              "                                                              StandardScaler())]),\n",
              "                                   transformers=[(&#x27;monthly_credit_burden&#x27;,\n",
              "                                                  Pipeline(steps=[(&#x27;functiontransformer-1&#x27;,\n",
              "                                                                   FunctionTransformer(feature_names_out=&lt;function ratio_name at 0x78a662f00dc0&gt;,\n",
              "                                                                                       func=&lt;function ratio_pipeline.&lt;locals&gt;.&lt;lambda&gt; at 0x78a6624f3250&gt;)),\n",
              "                                                                  (&#x27;function...\n",
              "                                                   &#x27;personal_status&#x27;,\n",
              "                                                   &#x27;credit_history&#x27;]),\n",
              "                                                 (&#x27;geo&#x27;,\n",
              "                                                  Pipeline(steps=[(&#x27;ordered_ordinal_encoder&#x27;,\n",
              "                                                                   OrderedOrdinalEncoder()),\n",
              "                                                                  (&#x27;cluster_simil&#x27;,\n",
              "                                                                   ClusterSimilarity(random_state=42))]),\n",
              "                                                  [&#x27;savings_status&#x27;,\n",
              "                                                   &#x27;checking_status&#x27;]),\n",
              "                                                 (&#x27;drop_out&#x27;, &#x27;drop&#x27;,\n",
              "                                                  [&#x27;num_dependents&#x27;,\n",
              "                                                   &#x27;foreign_worker&#x27;,\n",
              "                                                   &#x27;purpose&#x27;])])),\n",
              "                (&#x27;gradientboostingclassifier&#x27;,\n",
              "                 GradientBoostingClassifier(random_state=42))])</pre><b>In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook. <br />On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.</b></div><div class=\"sk-container\" hidden><div class=\"sk-item sk-dashed-wrapped\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-320\" type=\"checkbox\" ><label for=\"sk-estimator-id-320\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;&nbsp;Pipeline<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.pipeline.Pipeline.html\">?<span>Documentation for Pipeline</span></a><span class=\"sk-estimator-doc-link fitted\">i<span>Fitted</span></span></label><div class=\"sk-toggleable__content fitted\"><pre>Pipeline(steps=[(&#x27;columntransformer&#x27;,\n",
              "                 ColumnTransformer(remainder=Pipeline(steps=[(&#x27;standardscaler&#x27;,\n",
              "                                                              StandardScaler())]),\n",
              "                                   transformers=[(&#x27;monthly_credit_burden&#x27;,\n",
              "                                                  Pipeline(steps=[(&#x27;functiontransformer-1&#x27;,\n",
              "                                                                   FunctionTransformer(feature_names_out=&lt;function ratio_name at 0x78a662f00dc0&gt;,\n",
              "                                                                                       func=&lt;function ratio_pipeline.&lt;locals&gt;.&lt;lambda&gt; at 0x78a6624f3250&gt;)),\n",
              "                                                                  (&#x27;function...\n",
              "                                                   &#x27;personal_status&#x27;,\n",
              "                                                   &#x27;credit_history&#x27;]),\n",
              "                                                 (&#x27;geo&#x27;,\n",
              "                                                  Pipeline(steps=[(&#x27;ordered_ordinal_encoder&#x27;,\n",
              "                                                                   OrderedOrdinalEncoder()),\n",
              "                                                                  (&#x27;cluster_simil&#x27;,\n",
              "                                                                   ClusterSimilarity(random_state=42))]),\n",
              "                                                  [&#x27;savings_status&#x27;,\n",
              "                                                   &#x27;checking_status&#x27;]),\n",
              "                                                 (&#x27;drop_out&#x27;, &#x27;drop&#x27;,\n",
              "                                                  [&#x27;num_dependents&#x27;,\n",
              "                                                   &#x27;foreign_worker&#x27;,\n",
              "                                                   &#x27;purpose&#x27;])])),\n",
              "                (&#x27;gradientboostingclassifier&#x27;,\n",
              "                 GradientBoostingClassifier(random_state=42))])</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item sk-dashed-wrapped\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-321\" type=\"checkbox\" ><label for=\"sk-estimator-id-321\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;columntransformer: ColumnTransformer<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.compose.ColumnTransformer.html\">?<span>Documentation for columntransformer: ColumnTransformer</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>ColumnTransformer(remainder=Pipeline(steps=[(&#x27;standardscaler&#x27;,\n",
              "                                             StandardScaler())]),\n",
              "                  transformers=[(&#x27;monthly_credit_burden&#x27;,\n",
              "                                 Pipeline(steps=[(&#x27;functiontransformer-1&#x27;,\n",
              "                                                  FunctionTransformer(feature_names_out=&lt;function ratio_name at 0x78a662f00dc0&gt;,\n",
              "                                                                      func=&lt;function ratio_pipeline.&lt;locals&gt;.&lt;lambda&gt; at 0x78a6624f3250&gt;)),\n",
              "                                                 (&#x27;functiontransformer-2&#x27;,\n",
              "                                                  FunctionTransformer(fe...\n",
              "                                                  OneHotEncoder(handle_unknown=&#x27;ignore&#x27;))]),\n",
              "                                 [&#x27;job&#x27;, &#x27;housing&#x27;, &#x27;other_payment_plans&#x27;,\n",
              "                                  &#x27;property_magnitude&#x27;, &#x27;other_parties&#x27;,\n",
              "                                  &#x27;personal_status&#x27;, &#x27;credit_history&#x27;]),\n",
              "                                (&#x27;geo&#x27;,\n",
              "                                 Pipeline(steps=[(&#x27;ordered_ordinal_encoder&#x27;,\n",
              "                                                  OrderedOrdinalEncoder()),\n",
              "                                                 (&#x27;cluster_simil&#x27;,\n",
              "                                                  ClusterSimilarity(random_state=42))]),\n",
              "                                 [&#x27;savings_status&#x27;, &#x27;checking_status&#x27;]),\n",
              "                                (&#x27;drop_out&#x27;, &#x27;drop&#x27;,\n",
              "                                 [&#x27;num_dependents&#x27;, &#x27;foreign_worker&#x27;,\n",
              "                                  &#x27;purpose&#x27;])])</pre></div> </div></div><div class=\"sk-parallel\"><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-322\" type=\"checkbox\" ><label for=\"sk-estimator-id-322\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">monthly_credit_burden</label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;credit_amount&#x27;, &#x27;duration&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-323\" type=\"checkbox\" ><label for=\"sk-estimator-id-323\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;FunctionTransformer<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.FunctionTransformer.html\">?<span>Documentation for FunctionTransformer</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>FunctionTransformer(feature_names_out=&lt;function ratio_name at 0x78a662f00dc0&gt;,\n",
              "                    func=&lt;function ratio_pipeline.&lt;locals&gt;.&lt;lambda&gt; at 0x78a6624f3250&gt;)</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-324\" type=\"checkbox\" ><label for=\"sk-estimator-id-324\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;FunctionTransformer<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.FunctionTransformer.html\">?<span>Documentation for FunctionTransformer</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>FunctionTransformer(feature_names_out=&#x27;one-to-one&#x27;, func=&lt;ufunc &#x27;log&#x27;&gt;,\n",
              "                    inverse_func=&lt;ufunc &#x27;exp&#x27;&gt;)</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-325\" type=\"checkbox\" ><label for=\"sk-estimator-id-325\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;StandardScaler<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.StandardScaler.html\">?<span>Documentation for StandardScaler</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>StandardScaler()</pre></div> </div></div></div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-326\" type=\"checkbox\" ><label for=\"sk-estimator-id-326\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">installment_per_credit</label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;installment_commitment&#x27;, &#x27;credit_amount&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-327\" type=\"checkbox\" ><label for=\"sk-estimator-id-327\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;FunctionTransformer<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.FunctionTransformer.html\">?<span>Documentation for FunctionTransformer</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>FunctionTransformer(feature_names_out=&lt;function ratio_name at 0x78a662f00dc0&gt;,\n",
              "                    func=&lt;function ratio_pipeline.&lt;locals&gt;.&lt;lambda&gt; at 0x78a6624f3400&gt;)</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-328\" type=\"checkbox\" ><label for=\"sk-estimator-id-328\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;FunctionTransformer<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.FunctionTransformer.html\">?<span>Documentation for FunctionTransformer</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>FunctionTransformer(feature_names_out=&#x27;one-to-one&#x27;, func=&lt;ufunc &#x27;log&#x27;&gt;,\n",
              "                    inverse_func=&lt;ufunc &#x27;exp&#x27;&gt;)</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-329\" type=\"checkbox\" ><label for=\"sk-estimator-id-329\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;StandardScaler<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.StandardScaler.html\">?<span>Documentation for StandardScaler</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>StandardScaler()</pre></div> </div></div></div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-330\" type=\"checkbox\" ><label for=\"sk-estimator-id-330\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">dependents_per_credit</label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;num_dependents&#x27;, &#x27;credit_amount&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-331\" type=\"checkbox\" ><label for=\"sk-estimator-id-331\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;FunctionTransformer<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.FunctionTransformer.html\">?<span>Documentation for FunctionTransformer</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>FunctionTransformer(feature_names_out=&lt;function ratio_name at 0x78a662f00dc0&gt;,\n",
              "                    func=&lt;function ratio_pipeline.&lt;locals&gt;.&lt;lambda&gt; at 0x78a6624f35b0&gt;)</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-332\" type=\"checkbox\" ><label for=\"sk-estimator-id-332\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;FunctionTransformer<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.FunctionTransformer.html\">?<span>Documentation for FunctionTransformer</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>FunctionTransformer(feature_names_out=&#x27;one-to-one&#x27;, func=&lt;ufunc &#x27;log&#x27;&gt;,\n",
              "                    inverse_func=&lt;ufunc &#x27;exp&#x27;&gt;)</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-333\" type=\"checkbox\" ><label for=\"sk-estimator-id-333\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;StandardScaler<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.StandardScaler.html\">?<span>Documentation for StandardScaler</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>StandardScaler()</pre></div> </div></div></div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-334\" type=\"checkbox\" ><label for=\"sk-estimator-id-334\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">log</label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;duration&#x27;, &#x27;credit_amount&#x27;, &#x27;installment_commitment&#x27;, &#x27;residence_since&#x27;, &#x27;age&#x27;, &#x27;existing_credits&#x27;, &#x27;num_dependents&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-335\" type=\"checkbox\" ><label for=\"sk-estimator-id-335\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;FunctionTransformer<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.FunctionTransformer.html\">?<span>Documentation for FunctionTransformer</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>FunctionTransformer(feature_names_out=&#x27;one-to-one&#x27;, func=&lt;ufunc &#x27;log&#x27;&gt;,\n",
              "                    inverse_func=&lt;ufunc &#x27;exp&#x27;&gt;)</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-336\" type=\"checkbox\" ><label for=\"sk-estimator-id-336\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;StandardScaler<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.StandardScaler.html\">?<span>Documentation for StandardScaler</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>StandardScaler()</pre></div> </div></div></div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-337\" type=\"checkbox\" ><label for=\"sk-estimator-id-337\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">cat_ordinal</label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;checking_status&#x27;, &#x27;savings_status&#x27;, &#x27;employment&#x27;, &#x27;own_telephone&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-338\" type=\"checkbox\" ><label for=\"sk-estimator-id-338\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">OrderedOrdinalEncoder</label><div class=\"sk-toggleable__content fitted\"><pre>OrderedOrdinalEncoder()</pre></div> </div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-339\" type=\"checkbox\" ><label for=\"sk-estimator-id-339\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">cat_onehot</label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;job&#x27;, &#x27;housing&#x27;, &#x27;other_payment_plans&#x27;, &#x27;property_magnitude&#x27;, &#x27;other_parties&#x27;, &#x27;personal_status&#x27;, &#x27;credit_history&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-340\" type=\"checkbox\" ><label for=\"sk-estimator-id-340\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;OneHotEncoder<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.OneHotEncoder.html\">?<span>Documentation for OneHotEncoder</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>OneHotEncoder(handle_unknown=&#x27;ignore&#x27;)</pre></div> </div></div></div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-341\" type=\"checkbox\" ><label for=\"sk-estimator-id-341\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">geo</label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;savings_status&#x27;, &#x27;checking_status&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-342\" type=\"checkbox\" ><label for=\"sk-estimator-id-342\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">OrderedOrdinalEncoder</label><div class=\"sk-toggleable__content fitted\"><pre>OrderedOrdinalEncoder()</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-343\" type=\"checkbox\" ><label for=\"sk-estimator-id-343\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">ClusterSimilarity</label><div class=\"sk-toggleable__content fitted\"><pre>ClusterSimilarity(random_state=42)</pre></div> </div></div></div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-344\" type=\"checkbox\" ><label for=\"sk-estimator-id-344\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">drop_out</label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;num_dependents&#x27;, &#x27;foreign_worker&#x27;, &#x27;purpose&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-345\" type=\"checkbox\" ><label for=\"sk-estimator-id-345\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">drop</label><div class=\"sk-toggleable__content fitted\"><pre>drop</pre></div> </div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-346\" type=\"checkbox\" ><label for=\"sk-estimator-id-346\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">remainder</label><div class=\"sk-toggleable__content fitted\"><pre>[]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-347\" type=\"checkbox\" ><label for=\"sk-estimator-id-347\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;StandardScaler<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.StandardScaler.html\">?<span>Documentation for StandardScaler</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>StandardScaler()</pre></div> </div></div></div></div></div></div></div></div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-348\" type=\"checkbox\" ><label for=\"sk-estimator-id-348\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;GradientBoostingClassifier<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.ensemble.GradientBoostingClassifier.html\">?<span>Documentation for GradientBoostingClassifier</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>GradientBoostingClassifier(random_state=42)</pre></div> </div></div></div></div></div></div>"
            ]
          },
          "metadata": {},
          "execution_count": 82
        }
      ],
      "source": [
        "from sklearn.ensemble import GradientBoostingClassifier\n",
        "from sklearn.pipeline import make_pipeline\n",
        "\n",
        "grad_boost = make_pipeline(preprocessing, GradientBoostingClassifier(random_state=42))\n",
        "grad_boost.fit(X_train, y_train)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "u0XoQJpKsIam"
      },
      "source": [
        "Testing the accuracy of the model:"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 81,
      "metadata": {
        "id": "GaOydETWsCyM",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "0dc5974b-d7e8-4a63-c53b-a38aa8e4f1c3"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Training Set Performance:\n",
            "Accuracy: 0.9250\n",
            "Precision: 0.9153\n",
            "Recall: 0.9839\n",
            "F1-Score: 0.9484\n",
            "ROC-AUC: 0.9743\n",
            "------------------------------\n"
          ]
        }
      ],
      "source": [
        "from sklearn.metrics import accuracy_score, precision_score, recall_score, f1_score, roc_auc_score\n",
        "\n",
        "# Make predictions\n",
        "train_predictions = grad_boost.predict(X_train)\n",
        "\n",
        "# Evaluation metrics\n",
        "train_accuracy = accuracy_score(y_train, train_predictions)\n",
        "train_precision = precision_score(y_train, train_predictions, pos_label='good')\n",
        "train_recall = recall_score(y_train, train_predictions, pos_label='good')\n",
        "train_f1 = f1_score(y_train, train_predictions, pos_label='good')\n",
        "train_roc_auc = roc_auc_score(y_train, grad_boost.predict_proba(X_train)[:, 1])\n",
        "\n",
        "# Print training set results\n",
        "print(\"Training Set Performance:\")\n",
        "print(f\"Accuracy: {train_accuracy:.4f}\")\n",
        "print(f\"Precision: {train_precision:.4f}\")\n",
        "print(f\"Recall: {train_recall:.4f}\")\n",
        "print(f\"F1-Score: {train_f1:.4f}\")\n",
        "print(f\"ROC-AUC: {train_roc_auc:.4f}\")\n",
        "print(\"-\" * 30)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Ju7_pRGvEmw5"
      },
      "source": [
        "### Support Vector Machine"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "BTHZW-cFEmw6"
      },
      "source": [
        "We are now going to be training a random forest model."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 84,
      "metadata": {
        "id": "Xw04f0_GEmw6",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 357
        },
        "outputId": "3bddda17-b96c-4ebf-ea6c-54dfbb54468e"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Pipeline(steps=[('columntransformer',\n",
              "                 ColumnTransformer(remainder=Pipeline(steps=[('standardscaler',\n",
              "                                                              StandardScaler())]),\n",
              "                                   transformers=[('monthly_credit_burden',\n",
              "                                                  Pipeline(steps=[('functiontransformer-1',\n",
              "                                                                   FunctionTransformer(feature_names_out=<function ratio_name at 0x78a662f00dc0>,\n",
              "                                                                                       func=<function ratio_pipeline.<locals>.<lambda> at 0x78a6624f3250>)),\n",
              "                                                                  ('function...\n",
              "                                                   'other_parties',\n",
              "                                                   'personal_status',\n",
              "                                                   'credit_history']),\n",
              "                                                 ('geo',\n",
              "                                                  Pipeline(steps=[('ordered_ordinal_encoder',\n",
              "                                                                   OrderedOrdinalEncoder()),\n",
              "                                                                  ('cluster_simil',\n",
              "                                                                   ClusterSimilarity(random_state=42))]),\n",
              "                                                  ['savings_status',\n",
              "                                                   'checking_status']),\n",
              "                                                 ('drop_out', 'drop',\n",
              "                                                  ['num_dependents',\n",
              "                                                   'foreign_worker',\n",
              "                                                   'purpose'])])),\n",
              "                ('svc',\n",
              "                 SVC(class_weight='balanced', probability=True,\n",
              "                     random_state=42))])"
            ],
            "text/html": [
              "<style>#sk-container-id-14 {\n",
              "  /* Definition of color scheme common for light and dark mode */\n",
              "  --sklearn-color-text: black;\n",
              "  --sklearn-color-line: gray;\n",
              "  /* Definition of color scheme for unfitted estimators */\n",
              "  --sklearn-color-unfitted-level-0: #fff5e6;\n",
              "  --sklearn-color-unfitted-level-1: #f6e4d2;\n",
              "  --sklearn-color-unfitted-level-2: #ffe0b3;\n",
              "  --sklearn-color-unfitted-level-3: chocolate;\n",
              "  /* Definition of color scheme for fitted estimators */\n",
              "  --sklearn-color-fitted-level-0: #f0f8ff;\n",
              "  --sklearn-color-fitted-level-1: #d4ebff;\n",
              "  --sklearn-color-fitted-level-2: #b3dbfd;\n",
              "  --sklearn-color-fitted-level-3: cornflowerblue;\n",
              "\n",
              "  /* Specific color for light theme */\n",
              "  --sklearn-color-text-on-default-background: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, black)));\n",
              "  --sklearn-color-background: var(--sg-background-color, var(--theme-background, var(--jp-layout-color0, white)));\n",
              "  --sklearn-color-border-box: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, black)));\n",
              "  --sklearn-color-icon: #696969;\n",
              "\n",
              "  @media (prefers-color-scheme: dark) {\n",
              "    /* Redefinition of color scheme for dark theme */\n",
              "    --sklearn-color-text-on-default-background: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, white)));\n",
              "    --sklearn-color-background: var(--sg-background-color, var(--theme-background, var(--jp-layout-color0, #111)));\n",
              "    --sklearn-color-border-box: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, white)));\n",
              "    --sklearn-color-icon: #878787;\n",
              "  }\n",
              "}\n",
              "\n",
              "#sk-container-id-14 {\n",
              "  color: var(--sklearn-color-text);\n",
              "}\n",
              "\n",
              "#sk-container-id-14 pre {\n",
              "  padding: 0;\n",
              "}\n",
              "\n",
              "#sk-container-id-14 input.sk-hidden--visually {\n",
              "  border: 0;\n",
              "  clip: rect(1px 1px 1px 1px);\n",
              "  clip: rect(1px, 1px, 1px, 1px);\n",
              "  height: 1px;\n",
              "  margin: -1px;\n",
              "  overflow: hidden;\n",
              "  padding: 0;\n",
              "  position: absolute;\n",
              "  width: 1px;\n",
              "}\n",
              "\n",
              "#sk-container-id-14 div.sk-dashed-wrapped {\n",
              "  border: 1px dashed var(--sklearn-color-line);\n",
              "  margin: 0 0.4em 0.5em 0.4em;\n",
              "  box-sizing: border-box;\n",
              "  padding-bottom: 0.4em;\n",
              "  background-color: var(--sklearn-color-background);\n",
              "}\n",
              "\n",
              "#sk-container-id-14 div.sk-container {\n",
              "  /* jupyter's `normalize.less` sets `[hidden] { display: none; }`\n",
              "     but bootstrap.min.css set `[hidden] { display: none !important; }`\n",
              "     so we also need the `!important` here to be able to override the\n",
              "     default hidden behavior on the sphinx rendered scikit-learn.org.\n",
              "     See: https://github.com/scikit-learn/scikit-learn/issues/21755 */\n",
              "  display: inline-block !important;\n",
              "  position: relative;\n",
              "}\n",
              "\n",
              "#sk-container-id-14 div.sk-text-repr-fallback {\n",
              "  display: none;\n",
              "}\n",
              "\n",
              "div.sk-parallel-item,\n",
              "div.sk-serial,\n",
              "div.sk-item {\n",
              "  /* draw centered vertical line to link estimators */\n",
              "  background-image: linear-gradient(var(--sklearn-color-text-on-default-background), var(--sklearn-color-text-on-default-background));\n",
              "  background-size: 2px 100%;\n",
              "  background-repeat: no-repeat;\n",
              "  background-position: center center;\n",
              "}\n",
              "\n",
              "/* Parallel-specific style estimator block */\n",
              "\n",
              "#sk-container-id-14 div.sk-parallel-item::after {\n",
              "  content: \"\";\n",
              "  width: 100%;\n",
              "  border-bottom: 2px solid var(--sklearn-color-text-on-default-background);\n",
              "  flex-grow: 1;\n",
              "}\n",
              "\n",
              "#sk-container-id-14 div.sk-parallel {\n",
              "  display: flex;\n",
              "  align-items: stretch;\n",
              "  justify-content: center;\n",
              "  background-color: var(--sklearn-color-background);\n",
              "  position: relative;\n",
              "}\n",
              "\n",
              "#sk-container-id-14 div.sk-parallel-item {\n",
              "  display: flex;\n",
              "  flex-direction: column;\n",
              "}\n",
              "\n",
              "#sk-container-id-14 div.sk-parallel-item:first-child::after {\n",
              "  align-self: flex-end;\n",
              "  width: 50%;\n",
              "}\n",
              "\n",
              "#sk-container-id-14 div.sk-parallel-item:last-child::after {\n",
              "  align-self: flex-start;\n",
              "  width: 50%;\n",
              "}\n",
              "\n",
              "#sk-container-id-14 div.sk-parallel-item:only-child::after {\n",
              "  width: 0;\n",
              "}\n",
              "\n",
              "/* Serial-specific style estimator block */\n",
              "\n",
              "#sk-container-id-14 div.sk-serial {\n",
              "  display: flex;\n",
              "  flex-direction: column;\n",
              "  align-items: center;\n",
              "  background-color: var(--sklearn-color-background);\n",
              "  padding-right: 1em;\n",
              "  padding-left: 1em;\n",
              "}\n",
              "\n",
              "\n",
              "/* Toggleable style: style used for estimator/Pipeline/ColumnTransformer box that is\n",
              "clickable and can be expanded/collapsed.\n",
              "- Pipeline and ColumnTransformer use this feature and define the default style\n",
              "- Estimators will overwrite some part of the style using the `sk-estimator` class\n",
              "*/\n",
              "\n",
              "/* Pipeline and ColumnTransformer style (default) */\n",
              "\n",
              "#sk-container-id-14 div.sk-toggleable {\n",
              "  /* Default theme specific background. It is overwritten whether we have a\n",
              "  specific estimator or a Pipeline/ColumnTransformer */\n",
              "  background-color: var(--sklearn-color-background);\n",
              "}\n",
              "\n",
              "/* Toggleable label */\n",
              "#sk-container-id-14 label.sk-toggleable__label {\n",
              "  cursor: pointer;\n",
              "  display: block;\n",
              "  width: 100%;\n",
              "  margin-bottom: 0;\n",
              "  padding: 0.5em;\n",
              "  box-sizing: border-box;\n",
              "  text-align: center;\n",
              "}\n",
              "\n",
              "#sk-container-id-14 label.sk-toggleable__label-arrow:before {\n",
              "  /* Arrow on the left of the label */\n",
              "  content: \"▸\";\n",
              "  float: left;\n",
              "  margin-right: 0.25em;\n",
              "  color: var(--sklearn-color-icon);\n",
              "}\n",
              "\n",
              "#sk-container-id-14 label.sk-toggleable__label-arrow:hover:before {\n",
              "  color: var(--sklearn-color-text);\n",
              "}\n",
              "\n",
              "/* Toggleable content - dropdown */\n",
              "\n",
              "#sk-container-id-14 div.sk-toggleable__content {\n",
              "  max-height: 0;\n",
              "  max-width: 0;\n",
              "  overflow: hidden;\n",
              "  text-align: left;\n",
              "  /* unfitted */\n",
              "  background-color: var(--sklearn-color-unfitted-level-0);\n",
              "}\n",
              "\n",
              "#sk-container-id-14 div.sk-toggleable__content.fitted {\n",
              "  /* fitted */\n",
              "  background-color: var(--sklearn-color-fitted-level-0);\n",
              "}\n",
              "\n",
              "#sk-container-id-14 div.sk-toggleable__content pre {\n",
              "  margin: 0.2em;\n",
              "  border-radius: 0.25em;\n",
              "  color: var(--sklearn-color-text);\n",
              "  /* unfitted */\n",
              "  background-color: var(--sklearn-color-unfitted-level-0);\n",
              "}\n",
              "\n",
              "#sk-container-id-14 div.sk-toggleable__content.fitted pre {\n",
              "  /* unfitted */\n",
              "  background-color: var(--sklearn-color-fitted-level-0);\n",
              "}\n",
              "\n",
              "#sk-container-id-14 input.sk-toggleable__control:checked~div.sk-toggleable__content {\n",
              "  /* Expand drop-down */\n",
              "  max-height: 200px;\n",
              "  max-width: 100%;\n",
              "  overflow: auto;\n",
              "}\n",
              "\n",
              "#sk-container-id-14 input.sk-toggleable__control:checked~label.sk-toggleable__label-arrow:before {\n",
              "  content: \"▾\";\n",
              "}\n",
              "\n",
              "/* Pipeline/ColumnTransformer-specific style */\n",
              "\n",
              "#sk-container-id-14 div.sk-label input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
              "  color: var(--sklearn-color-text);\n",
              "  background-color: var(--sklearn-color-unfitted-level-2);\n",
              "}\n",
              "\n",
              "#sk-container-id-14 div.sk-label.fitted input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
              "  background-color: var(--sklearn-color-fitted-level-2);\n",
              "}\n",
              "\n",
              "/* Estimator-specific style */\n",
              "\n",
              "/* Colorize estimator box */\n",
              "#sk-container-id-14 div.sk-estimator input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
              "  /* unfitted */\n",
              "  background-color: var(--sklearn-color-unfitted-level-2);\n",
              "}\n",
              "\n",
              "#sk-container-id-14 div.sk-estimator.fitted input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
              "  /* fitted */\n",
              "  background-color: var(--sklearn-color-fitted-level-2);\n",
              "}\n",
              "\n",
              "#sk-container-id-14 div.sk-label label.sk-toggleable__label,\n",
              "#sk-container-id-14 div.sk-label label {\n",
              "  /* The background is the default theme color */\n",
              "  color: var(--sklearn-color-text-on-default-background);\n",
              "}\n",
              "\n",
              "/* On hover, darken the color of the background */\n",
              "#sk-container-id-14 div.sk-label:hover label.sk-toggleable__label {\n",
              "  color: var(--sklearn-color-text);\n",
              "  background-color: var(--sklearn-color-unfitted-level-2);\n",
              "}\n",
              "\n",
              "/* Label box, darken color on hover, fitted */\n",
              "#sk-container-id-14 div.sk-label.fitted:hover label.sk-toggleable__label.fitted {\n",
              "  color: var(--sklearn-color-text);\n",
              "  background-color: var(--sklearn-color-fitted-level-2);\n",
              "}\n",
              "\n",
              "/* Estimator label */\n",
              "\n",
              "#sk-container-id-14 div.sk-label label {\n",
              "  font-family: monospace;\n",
              "  font-weight: bold;\n",
              "  display: inline-block;\n",
              "  line-height: 1.2em;\n",
              "}\n",
              "\n",
              "#sk-container-id-14 div.sk-label-container {\n",
              "  text-align: center;\n",
              "}\n",
              "\n",
              "/* Estimator-specific */\n",
              "#sk-container-id-14 div.sk-estimator {\n",
              "  font-family: monospace;\n",
              "  border: 1px dotted var(--sklearn-color-border-box);\n",
              "  border-radius: 0.25em;\n",
              "  box-sizing: border-box;\n",
              "  margin-bottom: 0.5em;\n",
              "  /* unfitted */\n",
              "  background-color: var(--sklearn-color-unfitted-level-0);\n",
              "}\n",
              "\n",
              "#sk-container-id-14 div.sk-estimator.fitted {\n",
              "  /* fitted */\n",
              "  background-color: var(--sklearn-color-fitted-level-0);\n",
              "}\n",
              "\n",
              "/* on hover */\n",
              "#sk-container-id-14 div.sk-estimator:hover {\n",
              "  /* unfitted */\n",
              "  background-color: var(--sklearn-color-unfitted-level-2);\n",
              "}\n",
              "\n",
              "#sk-container-id-14 div.sk-estimator.fitted:hover {\n",
              "  /* fitted */\n",
              "  background-color: var(--sklearn-color-fitted-level-2);\n",
              "}\n",
              "\n",
              "/* Specification for estimator info (e.g. \"i\" and \"?\") */\n",
              "\n",
              "/* Common style for \"i\" and \"?\" */\n",
              "\n",
              ".sk-estimator-doc-link,\n",
              "a:link.sk-estimator-doc-link,\n",
              "a:visited.sk-estimator-doc-link {\n",
              "  float: right;\n",
              "  font-size: smaller;\n",
              "  line-height: 1em;\n",
              "  font-family: monospace;\n",
              "  background-color: var(--sklearn-color-background);\n",
              "  border-radius: 1em;\n",
              "  height: 1em;\n",
              "  width: 1em;\n",
              "  text-decoration: none !important;\n",
              "  margin-left: 1ex;\n",
              "  /* unfitted */\n",
              "  border: var(--sklearn-color-unfitted-level-1) 1pt solid;\n",
              "  color: var(--sklearn-color-unfitted-level-1);\n",
              "}\n",
              "\n",
              ".sk-estimator-doc-link.fitted,\n",
              "a:link.sk-estimator-doc-link.fitted,\n",
              "a:visited.sk-estimator-doc-link.fitted {\n",
              "  /* fitted */\n",
              "  border: var(--sklearn-color-fitted-level-1) 1pt solid;\n",
              "  color: var(--sklearn-color-fitted-level-1);\n",
              "}\n",
              "\n",
              "/* On hover */\n",
              "div.sk-estimator:hover .sk-estimator-doc-link:hover,\n",
              ".sk-estimator-doc-link:hover,\n",
              "div.sk-label-container:hover .sk-estimator-doc-link:hover,\n",
              ".sk-estimator-doc-link:hover {\n",
              "  /* unfitted */\n",
              "  background-color: var(--sklearn-color-unfitted-level-3);\n",
              "  color: var(--sklearn-color-background);\n",
              "  text-decoration: none;\n",
              "}\n",
              "\n",
              "div.sk-estimator.fitted:hover .sk-estimator-doc-link.fitted:hover,\n",
              ".sk-estimator-doc-link.fitted:hover,\n",
              "div.sk-label-container:hover .sk-estimator-doc-link.fitted:hover,\n",
              ".sk-estimator-doc-link.fitted:hover {\n",
              "  /* fitted */\n",
              "  background-color: var(--sklearn-color-fitted-level-3);\n",
              "  color: var(--sklearn-color-background);\n",
              "  text-decoration: none;\n",
              "}\n",
              "\n",
              "/* Span, style for the box shown on hovering the info icon */\n",
              ".sk-estimator-doc-link span {\n",
              "  display: none;\n",
              "  z-index: 9999;\n",
              "  position: relative;\n",
              "  font-weight: normal;\n",
              "  right: .2ex;\n",
              "  padding: .5ex;\n",
              "  margin: .5ex;\n",
              "  width: min-content;\n",
              "  min-width: 20ex;\n",
              "  max-width: 50ex;\n",
              "  color: var(--sklearn-color-text);\n",
              "  box-shadow: 2pt 2pt 4pt #999;\n",
              "  /* unfitted */\n",
              "  background: var(--sklearn-color-unfitted-level-0);\n",
              "  border: .5pt solid var(--sklearn-color-unfitted-level-3);\n",
              "}\n",
              "\n",
              ".sk-estimator-doc-link.fitted span {\n",
              "  /* fitted */\n",
              "  background: var(--sklearn-color-fitted-level-0);\n",
              "  border: var(--sklearn-color-fitted-level-3);\n",
              "}\n",
              "\n",
              ".sk-estimator-doc-link:hover span {\n",
              "  display: block;\n",
              "}\n",
              "\n",
              "/* \"?\"-specific style due to the `<a>` HTML tag */\n",
              "\n",
              "#sk-container-id-14 a.estimator_doc_link {\n",
              "  float: right;\n",
              "  font-size: 1rem;\n",
              "  line-height: 1em;\n",
              "  font-family: monospace;\n",
              "  background-color: var(--sklearn-color-background);\n",
              "  border-radius: 1rem;\n",
              "  height: 1rem;\n",
              "  width: 1rem;\n",
              "  text-decoration: none;\n",
              "  /* unfitted */\n",
              "  color: var(--sklearn-color-unfitted-level-1);\n",
              "  border: var(--sklearn-color-unfitted-level-1) 1pt solid;\n",
              "}\n",
              "\n",
              "#sk-container-id-14 a.estimator_doc_link.fitted {\n",
              "  /* fitted */\n",
              "  border: var(--sklearn-color-fitted-level-1) 1pt solid;\n",
              "  color: var(--sklearn-color-fitted-level-1);\n",
              "}\n",
              "\n",
              "/* On hover */\n",
              "#sk-container-id-14 a.estimator_doc_link:hover {\n",
              "  /* unfitted */\n",
              "  background-color: var(--sklearn-color-unfitted-level-3);\n",
              "  color: var(--sklearn-color-background);\n",
              "  text-decoration: none;\n",
              "}\n",
              "\n",
              "#sk-container-id-14 a.estimator_doc_link.fitted:hover {\n",
              "  /* fitted */\n",
              "  background-color: var(--sklearn-color-fitted-level-3);\n",
              "}\n",
              "</style><div id=\"sk-container-id-14\" class=\"sk-top-container\"><div class=\"sk-text-repr-fallback\"><pre>Pipeline(steps=[(&#x27;columntransformer&#x27;,\n",
              "                 ColumnTransformer(remainder=Pipeline(steps=[(&#x27;standardscaler&#x27;,\n",
              "                                                              StandardScaler())]),\n",
              "                                   transformers=[(&#x27;monthly_credit_burden&#x27;,\n",
              "                                                  Pipeline(steps=[(&#x27;functiontransformer-1&#x27;,\n",
              "                                                                   FunctionTransformer(feature_names_out=&lt;function ratio_name at 0x78a662f00dc0&gt;,\n",
              "                                                                                       func=&lt;function ratio_pipeline.&lt;locals&gt;.&lt;lambda&gt; at 0x78a6624f3250&gt;)),\n",
              "                                                                  (&#x27;function...\n",
              "                                                   &#x27;other_parties&#x27;,\n",
              "                                                   &#x27;personal_status&#x27;,\n",
              "                                                   &#x27;credit_history&#x27;]),\n",
              "                                                 (&#x27;geo&#x27;,\n",
              "                                                  Pipeline(steps=[(&#x27;ordered_ordinal_encoder&#x27;,\n",
              "                                                                   OrderedOrdinalEncoder()),\n",
              "                                                                  (&#x27;cluster_simil&#x27;,\n",
              "                                                                   ClusterSimilarity(random_state=42))]),\n",
              "                                                  [&#x27;savings_status&#x27;,\n",
              "                                                   &#x27;checking_status&#x27;]),\n",
              "                                                 (&#x27;drop_out&#x27;, &#x27;drop&#x27;,\n",
              "                                                  [&#x27;num_dependents&#x27;,\n",
              "                                                   &#x27;foreign_worker&#x27;,\n",
              "                                                   &#x27;purpose&#x27;])])),\n",
              "                (&#x27;svc&#x27;,\n",
              "                 SVC(class_weight=&#x27;balanced&#x27;, probability=True,\n",
              "                     random_state=42))])</pre><b>In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook. <br />On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.</b></div><div class=\"sk-container\" hidden><div class=\"sk-item sk-dashed-wrapped\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-378\" type=\"checkbox\" ><label for=\"sk-estimator-id-378\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;&nbsp;Pipeline<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.pipeline.Pipeline.html\">?<span>Documentation for Pipeline</span></a><span class=\"sk-estimator-doc-link fitted\">i<span>Fitted</span></span></label><div class=\"sk-toggleable__content fitted\"><pre>Pipeline(steps=[(&#x27;columntransformer&#x27;,\n",
              "                 ColumnTransformer(remainder=Pipeline(steps=[(&#x27;standardscaler&#x27;,\n",
              "                                                              StandardScaler())]),\n",
              "                                   transformers=[(&#x27;monthly_credit_burden&#x27;,\n",
              "                                                  Pipeline(steps=[(&#x27;functiontransformer-1&#x27;,\n",
              "                                                                   FunctionTransformer(feature_names_out=&lt;function ratio_name at 0x78a662f00dc0&gt;,\n",
              "                                                                                       func=&lt;function ratio_pipeline.&lt;locals&gt;.&lt;lambda&gt; at 0x78a6624f3250&gt;)),\n",
              "                                                                  (&#x27;function...\n",
              "                                                   &#x27;other_parties&#x27;,\n",
              "                                                   &#x27;personal_status&#x27;,\n",
              "                                                   &#x27;credit_history&#x27;]),\n",
              "                                                 (&#x27;geo&#x27;,\n",
              "                                                  Pipeline(steps=[(&#x27;ordered_ordinal_encoder&#x27;,\n",
              "                                                                   OrderedOrdinalEncoder()),\n",
              "                                                                  (&#x27;cluster_simil&#x27;,\n",
              "                                                                   ClusterSimilarity(random_state=42))]),\n",
              "                                                  [&#x27;savings_status&#x27;,\n",
              "                                                   &#x27;checking_status&#x27;]),\n",
              "                                                 (&#x27;drop_out&#x27;, &#x27;drop&#x27;,\n",
              "                                                  [&#x27;num_dependents&#x27;,\n",
              "                                                   &#x27;foreign_worker&#x27;,\n",
              "                                                   &#x27;purpose&#x27;])])),\n",
              "                (&#x27;svc&#x27;,\n",
              "                 SVC(class_weight=&#x27;balanced&#x27;, probability=True,\n",
              "                     random_state=42))])</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item sk-dashed-wrapped\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-379\" type=\"checkbox\" ><label for=\"sk-estimator-id-379\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;columntransformer: ColumnTransformer<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.compose.ColumnTransformer.html\">?<span>Documentation for columntransformer: ColumnTransformer</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>ColumnTransformer(remainder=Pipeline(steps=[(&#x27;standardscaler&#x27;,\n",
              "                                             StandardScaler())]),\n",
              "                  transformers=[(&#x27;monthly_credit_burden&#x27;,\n",
              "                                 Pipeline(steps=[(&#x27;functiontransformer-1&#x27;,\n",
              "                                                  FunctionTransformer(feature_names_out=&lt;function ratio_name at 0x78a662f00dc0&gt;,\n",
              "                                                                      func=&lt;function ratio_pipeline.&lt;locals&gt;.&lt;lambda&gt; at 0x78a6624f3250&gt;)),\n",
              "                                                 (&#x27;functiontransformer-2&#x27;,\n",
              "                                                  FunctionTransformer(fe...\n",
              "                                                  OneHotEncoder(handle_unknown=&#x27;ignore&#x27;))]),\n",
              "                                 [&#x27;job&#x27;, &#x27;housing&#x27;, &#x27;other_payment_plans&#x27;,\n",
              "                                  &#x27;property_magnitude&#x27;, &#x27;other_parties&#x27;,\n",
              "                                  &#x27;personal_status&#x27;, &#x27;credit_history&#x27;]),\n",
              "                                (&#x27;geo&#x27;,\n",
              "                                 Pipeline(steps=[(&#x27;ordered_ordinal_encoder&#x27;,\n",
              "                                                  OrderedOrdinalEncoder()),\n",
              "                                                 (&#x27;cluster_simil&#x27;,\n",
              "                                                  ClusterSimilarity(random_state=42))]),\n",
              "                                 [&#x27;savings_status&#x27;, &#x27;checking_status&#x27;]),\n",
              "                                (&#x27;drop_out&#x27;, &#x27;drop&#x27;,\n",
              "                                 [&#x27;num_dependents&#x27;, &#x27;foreign_worker&#x27;,\n",
              "                                  &#x27;purpose&#x27;])])</pre></div> </div></div><div class=\"sk-parallel\"><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-380\" type=\"checkbox\" ><label for=\"sk-estimator-id-380\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">monthly_credit_burden</label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;credit_amount&#x27;, &#x27;duration&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-381\" type=\"checkbox\" ><label for=\"sk-estimator-id-381\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;FunctionTransformer<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.FunctionTransformer.html\">?<span>Documentation for FunctionTransformer</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>FunctionTransformer(feature_names_out=&lt;function ratio_name at 0x78a662f00dc0&gt;,\n",
              "                    func=&lt;function ratio_pipeline.&lt;locals&gt;.&lt;lambda&gt; at 0x78a6624f3250&gt;)</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-382\" type=\"checkbox\" ><label for=\"sk-estimator-id-382\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;FunctionTransformer<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.FunctionTransformer.html\">?<span>Documentation for FunctionTransformer</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>FunctionTransformer(feature_names_out=&#x27;one-to-one&#x27;, func=&lt;ufunc &#x27;log&#x27;&gt;,\n",
              "                    inverse_func=&lt;ufunc &#x27;exp&#x27;&gt;)</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-383\" type=\"checkbox\" ><label for=\"sk-estimator-id-383\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;StandardScaler<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.StandardScaler.html\">?<span>Documentation for StandardScaler</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>StandardScaler()</pre></div> </div></div></div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-384\" type=\"checkbox\" ><label for=\"sk-estimator-id-384\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">installment_per_credit</label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;installment_commitment&#x27;, &#x27;credit_amount&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-385\" type=\"checkbox\" ><label for=\"sk-estimator-id-385\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;FunctionTransformer<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.FunctionTransformer.html\">?<span>Documentation for FunctionTransformer</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>FunctionTransformer(feature_names_out=&lt;function ratio_name at 0x78a662f00dc0&gt;,\n",
              "                    func=&lt;function ratio_pipeline.&lt;locals&gt;.&lt;lambda&gt; at 0x78a6624f3400&gt;)</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-386\" type=\"checkbox\" ><label for=\"sk-estimator-id-386\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;FunctionTransformer<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.FunctionTransformer.html\">?<span>Documentation for FunctionTransformer</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>FunctionTransformer(feature_names_out=&#x27;one-to-one&#x27;, func=&lt;ufunc &#x27;log&#x27;&gt;,\n",
              "                    inverse_func=&lt;ufunc &#x27;exp&#x27;&gt;)</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-387\" type=\"checkbox\" ><label for=\"sk-estimator-id-387\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;StandardScaler<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.StandardScaler.html\">?<span>Documentation for StandardScaler</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>StandardScaler()</pre></div> </div></div></div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-388\" type=\"checkbox\" ><label for=\"sk-estimator-id-388\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">dependents_per_credit</label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;num_dependents&#x27;, &#x27;credit_amount&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-389\" type=\"checkbox\" ><label for=\"sk-estimator-id-389\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;FunctionTransformer<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.FunctionTransformer.html\">?<span>Documentation for FunctionTransformer</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>FunctionTransformer(feature_names_out=&lt;function ratio_name at 0x78a662f00dc0&gt;,\n",
              "                    func=&lt;function ratio_pipeline.&lt;locals&gt;.&lt;lambda&gt; at 0x78a6624f35b0&gt;)</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-390\" type=\"checkbox\" ><label for=\"sk-estimator-id-390\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;FunctionTransformer<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.FunctionTransformer.html\">?<span>Documentation for FunctionTransformer</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>FunctionTransformer(feature_names_out=&#x27;one-to-one&#x27;, func=&lt;ufunc &#x27;log&#x27;&gt;,\n",
              "                    inverse_func=&lt;ufunc &#x27;exp&#x27;&gt;)</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-391\" type=\"checkbox\" ><label for=\"sk-estimator-id-391\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;StandardScaler<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.StandardScaler.html\">?<span>Documentation for StandardScaler</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>StandardScaler()</pre></div> </div></div></div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-392\" type=\"checkbox\" ><label for=\"sk-estimator-id-392\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">log</label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;duration&#x27;, &#x27;credit_amount&#x27;, &#x27;installment_commitment&#x27;, &#x27;residence_since&#x27;, &#x27;age&#x27;, &#x27;existing_credits&#x27;, &#x27;num_dependents&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-393\" type=\"checkbox\" ><label for=\"sk-estimator-id-393\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;FunctionTransformer<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.FunctionTransformer.html\">?<span>Documentation for FunctionTransformer</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>FunctionTransformer(feature_names_out=&#x27;one-to-one&#x27;, func=&lt;ufunc &#x27;log&#x27;&gt;,\n",
              "                    inverse_func=&lt;ufunc &#x27;exp&#x27;&gt;)</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-394\" type=\"checkbox\" ><label for=\"sk-estimator-id-394\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;StandardScaler<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.StandardScaler.html\">?<span>Documentation for StandardScaler</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>StandardScaler()</pre></div> </div></div></div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-395\" type=\"checkbox\" ><label for=\"sk-estimator-id-395\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">cat_ordinal</label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;checking_status&#x27;, &#x27;savings_status&#x27;, &#x27;employment&#x27;, &#x27;own_telephone&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-396\" type=\"checkbox\" ><label for=\"sk-estimator-id-396\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">OrderedOrdinalEncoder</label><div class=\"sk-toggleable__content fitted\"><pre>OrderedOrdinalEncoder()</pre></div> </div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-397\" type=\"checkbox\" ><label for=\"sk-estimator-id-397\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">cat_onehot</label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;job&#x27;, &#x27;housing&#x27;, &#x27;other_payment_plans&#x27;, &#x27;property_magnitude&#x27;, &#x27;other_parties&#x27;, &#x27;personal_status&#x27;, &#x27;credit_history&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-398\" type=\"checkbox\" ><label for=\"sk-estimator-id-398\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;OneHotEncoder<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.OneHotEncoder.html\">?<span>Documentation for OneHotEncoder</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>OneHotEncoder(handle_unknown=&#x27;ignore&#x27;)</pre></div> </div></div></div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-399\" type=\"checkbox\" ><label for=\"sk-estimator-id-399\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">geo</label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;savings_status&#x27;, &#x27;checking_status&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-400\" type=\"checkbox\" ><label for=\"sk-estimator-id-400\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">OrderedOrdinalEncoder</label><div class=\"sk-toggleable__content fitted\"><pre>OrderedOrdinalEncoder()</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-401\" type=\"checkbox\" ><label for=\"sk-estimator-id-401\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">ClusterSimilarity</label><div class=\"sk-toggleable__content fitted\"><pre>ClusterSimilarity(random_state=42)</pre></div> </div></div></div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-402\" type=\"checkbox\" ><label for=\"sk-estimator-id-402\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">drop_out</label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;num_dependents&#x27;, &#x27;foreign_worker&#x27;, &#x27;purpose&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-403\" type=\"checkbox\" ><label for=\"sk-estimator-id-403\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">drop</label><div class=\"sk-toggleable__content fitted\"><pre>drop</pre></div> </div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-404\" type=\"checkbox\" ><label for=\"sk-estimator-id-404\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">remainder</label><div class=\"sk-toggleable__content fitted\"><pre>[]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-405\" type=\"checkbox\" ><label for=\"sk-estimator-id-405\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;StandardScaler<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.StandardScaler.html\">?<span>Documentation for StandardScaler</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>StandardScaler()</pre></div> </div></div></div></div></div></div></div></div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-406\" type=\"checkbox\" ><label for=\"sk-estimator-id-406\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;SVC<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.svm.SVC.html\">?<span>Documentation for SVC</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>SVC(class_weight=&#x27;balanced&#x27;, probability=True, random_state=42)</pre></div> </div></div></div></div></div></div>"
            ]
          },
          "metadata": {},
          "execution_count": 84
        }
      ],
      "source": [
        "from sklearn.svm import SVC\n",
        "from sklearn.pipeline import make_pipeline\n",
        "\n",
        "svm_model = make_pipeline(preprocessing, SVC(random_state=42, probability=True, class_weight='balanced'))\n",
        "svm_model.fit(X_train, y_train)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "32dnYcpXEmw6"
      },
      "source": [
        "Testing the accuracy of the model:"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 85,
      "metadata": {
        "id": "bfq5CYMYEmw6",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "88fa88f8-d8da-437a-8d4a-405c23b55334"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Training Set Performance:\n",
            "Accuracy: 0.7913\n",
            "Precision: 0.9119\n",
            "Recall: 0.7768\n",
            "F1-Score: 0.8390\n",
            "ROC-AUC: 0.8964\n",
            "------------------------------\n"
          ]
        }
      ],
      "source": [
        "from sklearn.metrics import accuracy_score, precision_score, recall_score, f1_score, roc_auc_score\n",
        "\n",
        "# Make predictions\n",
        "train_predictions = svm_model.predict(X_train)\n",
        "\n",
        "# Evaluation metrics\n",
        "train_accuracy = accuracy_score(y_train, train_predictions)\n",
        "train_precision = precision_score(y_train, train_predictions, pos_label='good')\n",
        "train_recall = recall_score(y_train, train_predictions, pos_label='good')\n",
        "train_f1 = f1_score(y_train, train_predictions, pos_label='good')\n",
        "train_roc_auc = roc_auc_score(y_train, svm_model.predict_proba(X_train)[:, 1])\n",
        "\n",
        "# Print training set results\n",
        "print(\"Training Set Performance:\")\n",
        "print(f\"Accuracy: {train_accuracy:.4f}\")\n",
        "print(f\"Precision: {train_precision:.4f}\")\n",
        "print(f\"Recall: {train_recall:.4f}\")\n",
        "print(f\"F1-Score: {train_f1:.4f}\")\n",
        "print(f\"ROC-AUC: {train_roc_auc:.4f}\")\n",
        "print(\"-\" * 30)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "BG7AnDoSEqTr"
      },
      "source": [
        "### Extreme Gradient Boosting"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "VqFtUtt9EqTr"
      },
      "source": [
        "We are now going to be training a random forest model."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 89,
      "metadata": {
        "id": "L7IhMLu8EqTs",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 357
        },
        "outputId": "97694e61-bea0-4145-a387-6bebd0df07d6"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Pipeline(steps=[('columntransformer',\n",
              "                 ColumnTransformer(remainder=Pipeline(steps=[('standardscaler',\n",
              "                                                              StandardScaler())]),\n",
              "                                   transformers=[('monthly_credit_burden',\n",
              "                                                  Pipeline(steps=[('functiontransformer-1',\n",
              "                                                                   FunctionTransformer(feature_names_out=<function ratio_name at 0x78a662f00dc0>,\n",
              "                                                                                       func=<function ratio_pipeline.<locals>.<lambda> at 0x78a6624f3250>)),\n",
              "                                                                  ('function...\n",
              "                               feature_types=None, gamma=None, grow_policy=None,\n",
              "                               importance_type=None,\n",
              "                               interaction_constraints=None, learning_rate=None,\n",
              "                               max_bin=None, max_cat_threshold=None,\n",
              "                               max_cat_to_onehot=None, max_delta_step=None,\n",
              "                               max_depth=None, max_leaves=None,\n",
              "                               min_child_weight=None, missing=nan,\n",
              "                               monotone_constraints=None, multi_strategy=None,\n",
              "                               n_estimators=None, n_jobs=None,\n",
              "                               num_parallel_tree=None, random_state=42, ...))])"
            ],
            "text/html": [
              "<style>#sk-container-id-16 {\n",
              "  /* Definition of color scheme common for light and dark mode */\n",
              "  --sklearn-color-text: black;\n",
              "  --sklearn-color-line: gray;\n",
              "  /* Definition of color scheme for unfitted estimators */\n",
              "  --sklearn-color-unfitted-level-0: #fff5e6;\n",
              "  --sklearn-color-unfitted-level-1: #f6e4d2;\n",
              "  --sklearn-color-unfitted-level-2: #ffe0b3;\n",
              "  --sklearn-color-unfitted-level-3: chocolate;\n",
              "  /* Definition of color scheme for fitted estimators */\n",
              "  --sklearn-color-fitted-level-0: #f0f8ff;\n",
              "  --sklearn-color-fitted-level-1: #d4ebff;\n",
              "  --sklearn-color-fitted-level-2: #b3dbfd;\n",
              "  --sklearn-color-fitted-level-3: cornflowerblue;\n",
              "\n",
              "  /* Specific color for light theme */\n",
              "  --sklearn-color-text-on-default-background: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, black)));\n",
              "  --sklearn-color-background: var(--sg-background-color, var(--theme-background, var(--jp-layout-color0, white)));\n",
              "  --sklearn-color-border-box: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, black)));\n",
              "  --sklearn-color-icon: #696969;\n",
              "\n",
              "  @media (prefers-color-scheme: dark) {\n",
              "    /* Redefinition of color scheme for dark theme */\n",
              "    --sklearn-color-text-on-default-background: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, white)));\n",
              "    --sklearn-color-background: var(--sg-background-color, var(--theme-background, var(--jp-layout-color0, #111)));\n",
              "    --sklearn-color-border-box: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, white)));\n",
              "    --sklearn-color-icon: #878787;\n",
              "  }\n",
              "}\n",
              "\n",
              "#sk-container-id-16 {\n",
              "  color: var(--sklearn-color-text);\n",
              "}\n",
              "\n",
              "#sk-container-id-16 pre {\n",
              "  padding: 0;\n",
              "}\n",
              "\n",
              "#sk-container-id-16 input.sk-hidden--visually {\n",
              "  border: 0;\n",
              "  clip: rect(1px 1px 1px 1px);\n",
              "  clip: rect(1px, 1px, 1px, 1px);\n",
              "  height: 1px;\n",
              "  margin: -1px;\n",
              "  overflow: hidden;\n",
              "  padding: 0;\n",
              "  position: absolute;\n",
              "  width: 1px;\n",
              "}\n",
              "\n",
              "#sk-container-id-16 div.sk-dashed-wrapped {\n",
              "  border: 1px dashed var(--sklearn-color-line);\n",
              "  margin: 0 0.4em 0.5em 0.4em;\n",
              "  box-sizing: border-box;\n",
              "  padding-bottom: 0.4em;\n",
              "  background-color: var(--sklearn-color-background);\n",
              "}\n",
              "\n",
              "#sk-container-id-16 div.sk-container {\n",
              "  /* jupyter's `normalize.less` sets `[hidden] { display: none; }`\n",
              "     but bootstrap.min.css set `[hidden] { display: none !important; }`\n",
              "     so we also need the `!important` here to be able to override the\n",
              "     default hidden behavior on the sphinx rendered scikit-learn.org.\n",
              "     See: https://github.com/scikit-learn/scikit-learn/issues/21755 */\n",
              "  display: inline-block !important;\n",
              "  position: relative;\n",
              "}\n",
              "\n",
              "#sk-container-id-16 div.sk-text-repr-fallback {\n",
              "  display: none;\n",
              "}\n",
              "\n",
              "div.sk-parallel-item,\n",
              "div.sk-serial,\n",
              "div.sk-item {\n",
              "  /* draw centered vertical line to link estimators */\n",
              "  background-image: linear-gradient(var(--sklearn-color-text-on-default-background), var(--sklearn-color-text-on-default-background));\n",
              "  background-size: 2px 100%;\n",
              "  background-repeat: no-repeat;\n",
              "  background-position: center center;\n",
              "}\n",
              "\n",
              "/* Parallel-specific style estimator block */\n",
              "\n",
              "#sk-container-id-16 div.sk-parallel-item::after {\n",
              "  content: \"\";\n",
              "  width: 100%;\n",
              "  border-bottom: 2px solid var(--sklearn-color-text-on-default-background);\n",
              "  flex-grow: 1;\n",
              "}\n",
              "\n",
              "#sk-container-id-16 div.sk-parallel {\n",
              "  display: flex;\n",
              "  align-items: stretch;\n",
              "  justify-content: center;\n",
              "  background-color: var(--sklearn-color-background);\n",
              "  position: relative;\n",
              "}\n",
              "\n",
              "#sk-container-id-16 div.sk-parallel-item {\n",
              "  display: flex;\n",
              "  flex-direction: column;\n",
              "}\n",
              "\n",
              "#sk-container-id-16 div.sk-parallel-item:first-child::after {\n",
              "  align-self: flex-end;\n",
              "  width: 50%;\n",
              "}\n",
              "\n",
              "#sk-container-id-16 div.sk-parallel-item:last-child::after {\n",
              "  align-self: flex-start;\n",
              "  width: 50%;\n",
              "}\n",
              "\n",
              "#sk-container-id-16 div.sk-parallel-item:only-child::after {\n",
              "  width: 0;\n",
              "}\n",
              "\n",
              "/* Serial-specific style estimator block */\n",
              "\n",
              "#sk-container-id-16 div.sk-serial {\n",
              "  display: flex;\n",
              "  flex-direction: column;\n",
              "  align-items: center;\n",
              "  background-color: var(--sklearn-color-background);\n",
              "  padding-right: 1em;\n",
              "  padding-left: 1em;\n",
              "}\n",
              "\n",
              "\n",
              "/* Toggleable style: style used for estimator/Pipeline/ColumnTransformer box that is\n",
              "clickable and can be expanded/collapsed.\n",
              "- Pipeline and ColumnTransformer use this feature and define the default style\n",
              "- Estimators will overwrite some part of the style using the `sk-estimator` class\n",
              "*/\n",
              "\n",
              "/* Pipeline and ColumnTransformer style (default) */\n",
              "\n",
              "#sk-container-id-16 div.sk-toggleable {\n",
              "  /* Default theme specific background. It is overwritten whether we have a\n",
              "  specific estimator or a Pipeline/ColumnTransformer */\n",
              "  background-color: var(--sklearn-color-background);\n",
              "}\n",
              "\n",
              "/* Toggleable label */\n",
              "#sk-container-id-16 label.sk-toggleable__label {\n",
              "  cursor: pointer;\n",
              "  display: block;\n",
              "  width: 100%;\n",
              "  margin-bottom: 0;\n",
              "  padding: 0.5em;\n",
              "  box-sizing: border-box;\n",
              "  text-align: center;\n",
              "}\n",
              "\n",
              "#sk-container-id-16 label.sk-toggleable__label-arrow:before {\n",
              "  /* Arrow on the left of the label */\n",
              "  content: \"▸\";\n",
              "  float: left;\n",
              "  margin-right: 0.25em;\n",
              "  color: var(--sklearn-color-icon);\n",
              "}\n",
              "\n",
              "#sk-container-id-16 label.sk-toggleable__label-arrow:hover:before {\n",
              "  color: var(--sklearn-color-text);\n",
              "}\n",
              "\n",
              "/* Toggleable content - dropdown */\n",
              "\n",
              "#sk-container-id-16 div.sk-toggleable__content {\n",
              "  max-height: 0;\n",
              "  max-width: 0;\n",
              "  overflow: hidden;\n",
              "  text-align: left;\n",
              "  /* unfitted */\n",
              "  background-color: var(--sklearn-color-unfitted-level-0);\n",
              "}\n",
              "\n",
              "#sk-container-id-16 div.sk-toggleable__content.fitted {\n",
              "  /* fitted */\n",
              "  background-color: var(--sklearn-color-fitted-level-0);\n",
              "}\n",
              "\n",
              "#sk-container-id-16 div.sk-toggleable__content pre {\n",
              "  margin: 0.2em;\n",
              "  border-radius: 0.25em;\n",
              "  color: var(--sklearn-color-text);\n",
              "  /* unfitted */\n",
              "  background-color: var(--sklearn-color-unfitted-level-0);\n",
              "}\n",
              "\n",
              "#sk-container-id-16 div.sk-toggleable__content.fitted pre {\n",
              "  /* unfitted */\n",
              "  background-color: var(--sklearn-color-fitted-level-0);\n",
              "}\n",
              "\n",
              "#sk-container-id-16 input.sk-toggleable__control:checked~div.sk-toggleable__content {\n",
              "  /* Expand drop-down */\n",
              "  max-height: 200px;\n",
              "  max-width: 100%;\n",
              "  overflow: auto;\n",
              "}\n",
              "\n",
              "#sk-container-id-16 input.sk-toggleable__control:checked~label.sk-toggleable__label-arrow:before {\n",
              "  content: \"▾\";\n",
              "}\n",
              "\n",
              "/* Pipeline/ColumnTransformer-specific style */\n",
              "\n",
              "#sk-container-id-16 div.sk-label input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
              "  color: var(--sklearn-color-text);\n",
              "  background-color: var(--sklearn-color-unfitted-level-2);\n",
              "}\n",
              "\n",
              "#sk-container-id-16 div.sk-label.fitted input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
              "  background-color: var(--sklearn-color-fitted-level-2);\n",
              "}\n",
              "\n",
              "/* Estimator-specific style */\n",
              "\n",
              "/* Colorize estimator box */\n",
              "#sk-container-id-16 div.sk-estimator input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
              "  /* unfitted */\n",
              "  background-color: var(--sklearn-color-unfitted-level-2);\n",
              "}\n",
              "\n",
              "#sk-container-id-16 div.sk-estimator.fitted input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
              "  /* fitted */\n",
              "  background-color: var(--sklearn-color-fitted-level-2);\n",
              "}\n",
              "\n",
              "#sk-container-id-16 div.sk-label label.sk-toggleable__label,\n",
              "#sk-container-id-16 div.sk-label label {\n",
              "  /* The background is the default theme color */\n",
              "  color: var(--sklearn-color-text-on-default-background);\n",
              "}\n",
              "\n",
              "/* On hover, darken the color of the background */\n",
              "#sk-container-id-16 div.sk-label:hover label.sk-toggleable__label {\n",
              "  color: var(--sklearn-color-text);\n",
              "  background-color: var(--sklearn-color-unfitted-level-2);\n",
              "}\n",
              "\n",
              "/* Label box, darken color on hover, fitted */\n",
              "#sk-container-id-16 div.sk-label.fitted:hover label.sk-toggleable__label.fitted {\n",
              "  color: var(--sklearn-color-text);\n",
              "  background-color: var(--sklearn-color-fitted-level-2);\n",
              "}\n",
              "\n",
              "/* Estimator label */\n",
              "\n",
              "#sk-container-id-16 div.sk-label label {\n",
              "  font-family: monospace;\n",
              "  font-weight: bold;\n",
              "  display: inline-block;\n",
              "  line-height: 1.2em;\n",
              "}\n",
              "\n",
              "#sk-container-id-16 div.sk-label-container {\n",
              "  text-align: center;\n",
              "}\n",
              "\n",
              "/* Estimator-specific */\n",
              "#sk-container-id-16 div.sk-estimator {\n",
              "  font-family: monospace;\n",
              "  border: 1px dotted var(--sklearn-color-border-box);\n",
              "  border-radius: 0.25em;\n",
              "  box-sizing: border-box;\n",
              "  margin-bottom: 0.5em;\n",
              "  /* unfitted */\n",
              "  background-color: var(--sklearn-color-unfitted-level-0);\n",
              "}\n",
              "\n",
              "#sk-container-id-16 div.sk-estimator.fitted {\n",
              "  /* fitted */\n",
              "  background-color: var(--sklearn-color-fitted-level-0);\n",
              "}\n",
              "\n",
              "/* on hover */\n",
              "#sk-container-id-16 div.sk-estimator:hover {\n",
              "  /* unfitted */\n",
              "  background-color: var(--sklearn-color-unfitted-level-2);\n",
              "}\n",
              "\n",
              "#sk-container-id-16 div.sk-estimator.fitted:hover {\n",
              "  /* fitted */\n",
              "  background-color: var(--sklearn-color-fitted-level-2);\n",
              "}\n",
              "\n",
              "/* Specification for estimator info (e.g. \"i\" and \"?\") */\n",
              "\n",
              "/* Common style for \"i\" and \"?\" */\n",
              "\n",
              ".sk-estimator-doc-link,\n",
              "a:link.sk-estimator-doc-link,\n",
              "a:visited.sk-estimator-doc-link {\n",
              "  float: right;\n",
              "  font-size: smaller;\n",
              "  line-height: 1em;\n",
              "  font-family: monospace;\n",
              "  background-color: var(--sklearn-color-background);\n",
              "  border-radius: 1em;\n",
              "  height: 1em;\n",
              "  width: 1em;\n",
              "  text-decoration: none !important;\n",
              "  margin-left: 1ex;\n",
              "  /* unfitted */\n",
              "  border: var(--sklearn-color-unfitted-level-1) 1pt solid;\n",
              "  color: var(--sklearn-color-unfitted-level-1);\n",
              "}\n",
              "\n",
              ".sk-estimator-doc-link.fitted,\n",
              "a:link.sk-estimator-doc-link.fitted,\n",
              "a:visited.sk-estimator-doc-link.fitted {\n",
              "  /* fitted */\n",
              "  border: var(--sklearn-color-fitted-level-1) 1pt solid;\n",
              "  color: var(--sklearn-color-fitted-level-1);\n",
              "}\n",
              "\n",
              "/* On hover */\n",
              "div.sk-estimator:hover .sk-estimator-doc-link:hover,\n",
              ".sk-estimator-doc-link:hover,\n",
              "div.sk-label-container:hover .sk-estimator-doc-link:hover,\n",
              ".sk-estimator-doc-link:hover {\n",
              "  /* unfitted */\n",
              "  background-color: var(--sklearn-color-unfitted-level-3);\n",
              "  color: var(--sklearn-color-background);\n",
              "  text-decoration: none;\n",
              "}\n",
              "\n",
              "div.sk-estimator.fitted:hover .sk-estimator-doc-link.fitted:hover,\n",
              ".sk-estimator-doc-link.fitted:hover,\n",
              "div.sk-label-container:hover .sk-estimator-doc-link.fitted:hover,\n",
              ".sk-estimator-doc-link.fitted:hover {\n",
              "  /* fitted */\n",
              "  background-color: var(--sklearn-color-fitted-level-3);\n",
              "  color: var(--sklearn-color-background);\n",
              "  text-decoration: none;\n",
              "}\n",
              "\n",
              "/* Span, style for the box shown on hovering the info icon */\n",
              ".sk-estimator-doc-link span {\n",
              "  display: none;\n",
              "  z-index: 9999;\n",
              "  position: relative;\n",
              "  font-weight: normal;\n",
              "  right: .2ex;\n",
              "  padding: .5ex;\n",
              "  margin: .5ex;\n",
              "  width: min-content;\n",
              "  min-width: 20ex;\n",
              "  max-width: 50ex;\n",
              "  color: var(--sklearn-color-text);\n",
              "  box-shadow: 2pt 2pt 4pt #999;\n",
              "  /* unfitted */\n",
              "  background: var(--sklearn-color-unfitted-level-0);\n",
              "  border: .5pt solid var(--sklearn-color-unfitted-level-3);\n",
              "}\n",
              "\n",
              ".sk-estimator-doc-link.fitted span {\n",
              "  /* fitted */\n",
              "  background: var(--sklearn-color-fitted-level-0);\n",
              "  border: var(--sklearn-color-fitted-level-3);\n",
              "}\n",
              "\n",
              ".sk-estimator-doc-link:hover span {\n",
              "  display: block;\n",
              "}\n",
              "\n",
              "/* \"?\"-specific style due to the `<a>` HTML tag */\n",
              "\n",
              "#sk-container-id-16 a.estimator_doc_link {\n",
              "  float: right;\n",
              "  font-size: 1rem;\n",
              "  line-height: 1em;\n",
              "  font-family: monospace;\n",
              "  background-color: var(--sklearn-color-background);\n",
              "  border-radius: 1rem;\n",
              "  height: 1rem;\n",
              "  width: 1rem;\n",
              "  text-decoration: none;\n",
              "  /* unfitted */\n",
              "  color: var(--sklearn-color-unfitted-level-1);\n",
              "  border: var(--sklearn-color-unfitted-level-1) 1pt solid;\n",
              "}\n",
              "\n",
              "#sk-container-id-16 a.estimator_doc_link.fitted {\n",
              "  /* fitted */\n",
              "  border: var(--sklearn-color-fitted-level-1) 1pt solid;\n",
              "  color: var(--sklearn-color-fitted-level-1);\n",
              "}\n",
              "\n",
              "/* On hover */\n",
              "#sk-container-id-16 a.estimator_doc_link:hover {\n",
              "  /* unfitted */\n",
              "  background-color: var(--sklearn-color-unfitted-level-3);\n",
              "  color: var(--sklearn-color-background);\n",
              "  text-decoration: none;\n",
              "}\n",
              "\n",
              "#sk-container-id-16 a.estimator_doc_link.fitted:hover {\n",
              "  /* fitted */\n",
              "  background-color: var(--sklearn-color-fitted-level-3);\n",
              "}\n",
              "</style><div id=\"sk-container-id-16\" class=\"sk-top-container\"><div class=\"sk-text-repr-fallback\"><pre>Pipeline(steps=[(&#x27;columntransformer&#x27;,\n",
              "                 ColumnTransformer(remainder=Pipeline(steps=[(&#x27;standardscaler&#x27;,\n",
              "                                                              StandardScaler())]),\n",
              "                                   transformers=[(&#x27;monthly_credit_burden&#x27;,\n",
              "                                                  Pipeline(steps=[(&#x27;functiontransformer-1&#x27;,\n",
              "                                                                   FunctionTransformer(feature_names_out=&lt;function ratio_name at 0x78a662f00dc0&gt;,\n",
              "                                                                                       func=&lt;function ratio_pipeline.&lt;locals&gt;.&lt;lambda&gt; at 0x78a6624f3250&gt;)),\n",
              "                                                                  (&#x27;function...\n",
              "                               feature_types=None, gamma=None, grow_policy=None,\n",
              "                               importance_type=None,\n",
              "                               interaction_constraints=None, learning_rate=None,\n",
              "                               max_bin=None, max_cat_threshold=None,\n",
              "                               max_cat_to_onehot=None, max_delta_step=None,\n",
              "                               max_depth=None, max_leaves=None,\n",
              "                               min_child_weight=None, missing=nan,\n",
              "                               monotone_constraints=None, multi_strategy=None,\n",
              "                               n_estimators=None, n_jobs=None,\n",
              "                               num_parallel_tree=None, random_state=42, ...))])</pre><b>In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook. <br />On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.</b></div><div class=\"sk-container\" hidden><div class=\"sk-item sk-dashed-wrapped\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-436\" type=\"checkbox\" ><label for=\"sk-estimator-id-436\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;&nbsp;Pipeline<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.pipeline.Pipeline.html\">?<span>Documentation for Pipeline</span></a><span class=\"sk-estimator-doc-link fitted\">i<span>Fitted</span></span></label><div class=\"sk-toggleable__content fitted\"><pre>Pipeline(steps=[(&#x27;columntransformer&#x27;,\n",
              "                 ColumnTransformer(remainder=Pipeline(steps=[(&#x27;standardscaler&#x27;,\n",
              "                                                              StandardScaler())]),\n",
              "                                   transformers=[(&#x27;monthly_credit_burden&#x27;,\n",
              "                                                  Pipeline(steps=[(&#x27;functiontransformer-1&#x27;,\n",
              "                                                                   FunctionTransformer(feature_names_out=&lt;function ratio_name at 0x78a662f00dc0&gt;,\n",
              "                                                                                       func=&lt;function ratio_pipeline.&lt;locals&gt;.&lt;lambda&gt; at 0x78a6624f3250&gt;)),\n",
              "                                                                  (&#x27;function...\n",
              "                               feature_types=None, gamma=None, grow_policy=None,\n",
              "                               importance_type=None,\n",
              "                               interaction_constraints=None, learning_rate=None,\n",
              "                               max_bin=None, max_cat_threshold=None,\n",
              "                               max_cat_to_onehot=None, max_delta_step=None,\n",
              "                               max_depth=None, max_leaves=None,\n",
              "                               min_child_weight=None, missing=nan,\n",
              "                               monotone_constraints=None, multi_strategy=None,\n",
              "                               n_estimators=None, n_jobs=None,\n",
              "                               num_parallel_tree=None, random_state=42, ...))])</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item sk-dashed-wrapped\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-437\" type=\"checkbox\" ><label for=\"sk-estimator-id-437\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;columntransformer: ColumnTransformer<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.compose.ColumnTransformer.html\">?<span>Documentation for columntransformer: ColumnTransformer</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>ColumnTransformer(remainder=Pipeline(steps=[(&#x27;standardscaler&#x27;,\n",
              "                                             StandardScaler())]),\n",
              "                  transformers=[(&#x27;monthly_credit_burden&#x27;,\n",
              "                                 Pipeline(steps=[(&#x27;functiontransformer-1&#x27;,\n",
              "                                                  FunctionTransformer(feature_names_out=&lt;function ratio_name at 0x78a662f00dc0&gt;,\n",
              "                                                                      func=&lt;function ratio_pipeline.&lt;locals&gt;.&lt;lambda&gt; at 0x78a6624f3250&gt;)),\n",
              "                                                 (&#x27;functiontransformer-2&#x27;,\n",
              "                                                  FunctionTransformer(fe...\n",
              "                                                  OneHotEncoder(handle_unknown=&#x27;ignore&#x27;))]),\n",
              "                                 [&#x27;job&#x27;, &#x27;housing&#x27;, &#x27;other_payment_plans&#x27;,\n",
              "                                  &#x27;property_magnitude&#x27;, &#x27;other_parties&#x27;,\n",
              "                                  &#x27;personal_status&#x27;, &#x27;credit_history&#x27;]),\n",
              "                                (&#x27;geo&#x27;,\n",
              "                                 Pipeline(steps=[(&#x27;ordered_ordinal_encoder&#x27;,\n",
              "                                                  OrderedOrdinalEncoder()),\n",
              "                                                 (&#x27;cluster_simil&#x27;,\n",
              "                                                  ClusterSimilarity(random_state=42))]),\n",
              "                                 [&#x27;savings_status&#x27;, &#x27;checking_status&#x27;]),\n",
              "                                (&#x27;drop_out&#x27;, &#x27;drop&#x27;,\n",
              "                                 [&#x27;num_dependents&#x27;, &#x27;foreign_worker&#x27;,\n",
              "                                  &#x27;purpose&#x27;])])</pre></div> </div></div><div class=\"sk-parallel\"><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-438\" type=\"checkbox\" ><label for=\"sk-estimator-id-438\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">monthly_credit_burden</label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;credit_amount&#x27;, &#x27;duration&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-439\" type=\"checkbox\" ><label for=\"sk-estimator-id-439\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;FunctionTransformer<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.FunctionTransformer.html\">?<span>Documentation for FunctionTransformer</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>FunctionTransformer(feature_names_out=&lt;function ratio_name at 0x78a662f00dc0&gt;,\n",
              "                    func=&lt;function ratio_pipeline.&lt;locals&gt;.&lt;lambda&gt; at 0x78a6624f3250&gt;)</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-440\" type=\"checkbox\" ><label for=\"sk-estimator-id-440\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;FunctionTransformer<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.FunctionTransformer.html\">?<span>Documentation for FunctionTransformer</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>FunctionTransformer(feature_names_out=&#x27;one-to-one&#x27;, func=&lt;ufunc &#x27;log&#x27;&gt;,\n",
              "                    inverse_func=&lt;ufunc &#x27;exp&#x27;&gt;)</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-441\" type=\"checkbox\" ><label for=\"sk-estimator-id-441\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;StandardScaler<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.StandardScaler.html\">?<span>Documentation for StandardScaler</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>StandardScaler()</pre></div> </div></div></div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-442\" type=\"checkbox\" ><label for=\"sk-estimator-id-442\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">installment_per_credit</label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;installment_commitment&#x27;, &#x27;credit_amount&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-443\" type=\"checkbox\" ><label for=\"sk-estimator-id-443\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;FunctionTransformer<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.FunctionTransformer.html\">?<span>Documentation for FunctionTransformer</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>FunctionTransformer(feature_names_out=&lt;function ratio_name at 0x78a662f00dc0&gt;,\n",
              "                    func=&lt;function ratio_pipeline.&lt;locals&gt;.&lt;lambda&gt; at 0x78a6624f3400&gt;)</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-444\" type=\"checkbox\" ><label for=\"sk-estimator-id-444\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;FunctionTransformer<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.FunctionTransformer.html\">?<span>Documentation for FunctionTransformer</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>FunctionTransformer(feature_names_out=&#x27;one-to-one&#x27;, func=&lt;ufunc &#x27;log&#x27;&gt;,\n",
              "                    inverse_func=&lt;ufunc &#x27;exp&#x27;&gt;)</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-445\" type=\"checkbox\" ><label for=\"sk-estimator-id-445\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;StandardScaler<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.StandardScaler.html\">?<span>Documentation for StandardScaler</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>StandardScaler()</pre></div> </div></div></div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-446\" type=\"checkbox\" ><label for=\"sk-estimator-id-446\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">dependents_per_credit</label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;num_dependents&#x27;, &#x27;credit_amount&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-447\" type=\"checkbox\" ><label for=\"sk-estimator-id-447\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;FunctionTransformer<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.FunctionTransformer.html\">?<span>Documentation for FunctionTransformer</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>FunctionTransformer(feature_names_out=&lt;function ratio_name at 0x78a662f00dc0&gt;,\n",
              "                    func=&lt;function ratio_pipeline.&lt;locals&gt;.&lt;lambda&gt; at 0x78a6624f35b0&gt;)</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-448\" type=\"checkbox\" ><label for=\"sk-estimator-id-448\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;FunctionTransformer<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.FunctionTransformer.html\">?<span>Documentation for FunctionTransformer</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>FunctionTransformer(feature_names_out=&#x27;one-to-one&#x27;, func=&lt;ufunc &#x27;log&#x27;&gt;,\n",
              "                    inverse_func=&lt;ufunc &#x27;exp&#x27;&gt;)</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-449\" type=\"checkbox\" ><label for=\"sk-estimator-id-449\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;StandardScaler<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.StandardScaler.html\">?<span>Documentation for StandardScaler</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>StandardScaler()</pre></div> </div></div></div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-450\" type=\"checkbox\" ><label for=\"sk-estimator-id-450\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">log</label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;duration&#x27;, &#x27;credit_amount&#x27;, &#x27;installment_commitment&#x27;, &#x27;residence_since&#x27;, &#x27;age&#x27;, &#x27;existing_credits&#x27;, &#x27;num_dependents&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-451\" type=\"checkbox\" ><label for=\"sk-estimator-id-451\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;FunctionTransformer<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.FunctionTransformer.html\">?<span>Documentation for FunctionTransformer</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>FunctionTransformer(feature_names_out=&#x27;one-to-one&#x27;, func=&lt;ufunc &#x27;log&#x27;&gt;,\n",
              "                    inverse_func=&lt;ufunc &#x27;exp&#x27;&gt;)</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-452\" type=\"checkbox\" ><label for=\"sk-estimator-id-452\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;StandardScaler<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.StandardScaler.html\">?<span>Documentation for StandardScaler</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>StandardScaler()</pre></div> </div></div></div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-453\" type=\"checkbox\" ><label for=\"sk-estimator-id-453\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">cat_ordinal</label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;checking_status&#x27;, &#x27;savings_status&#x27;, &#x27;employment&#x27;, &#x27;own_telephone&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-454\" type=\"checkbox\" ><label for=\"sk-estimator-id-454\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">OrderedOrdinalEncoder</label><div class=\"sk-toggleable__content fitted\"><pre>OrderedOrdinalEncoder()</pre></div> </div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-455\" type=\"checkbox\" ><label for=\"sk-estimator-id-455\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">cat_onehot</label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;job&#x27;, &#x27;housing&#x27;, &#x27;other_payment_plans&#x27;, &#x27;property_magnitude&#x27;, &#x27;other_parties&#x27;, &#x27;personal_status&#x27;, &#x27;credit_history&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-456\" type=\"checkbox\" ><label for=\"sk-estimator-id-456\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;OneHotEncoder<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.OneHotEncoder.html\">?<span>Documentation for OneHotEncoder</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>OneHotEncoder(handle_unknown=&#x27;ignore&#x27;)</pre></div> </div></div></div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-457\" type=\"checkbox\" ><label for=\"sk-estimator-id-457\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">geo</label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;savings_status&#x27;, &#x27;checking_status&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-458\" type=\"checkbox\" ><label for=\"sk-estimator-id-458\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">OrderedOrdinalEncoder</label><div class=\"sk-toggleable__content fitted\"><pre>OrderedOrdinalEncoder()</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-459\" type=\"checkbox\" ><label for=\"sk-estimator-id-459\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">ClusterSimilarity</label><div class=\"sk-toggleable__content fitted\"><pre>ClusterSimilarity(random_state=42)</pre></div> </div></div></div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-460\" type=\"checkbox\" ><label for=\"sk-estimator-id-460\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">drop_out</label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;num_dependents&#x27;, &#x27;foreign_worker&#x27;, &#x27;purpose&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-461\" type=\"checkbox\" ><label for=\"sk-estimator-id-461\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">drop</label><div class=\"sk-toggleable__content fitted\"><pre>drop</pre></div> </div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-462\" type=\"checkbox\" ><label for=\"sk-estimator-id-462\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">remainder</label><div class=\"sk-toggleable__content fitted\"><pre>[]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-463\" type=\"checkbox\" ><label for=\"sk-estimator-id-463\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;StandardScaler<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.StandardScaler.html\">?<span>Documentation for StandardScaler</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>StandardScaler()</pre></div> </div></div></div></div></div></div></div></div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-464\" type=\"checkbox\" ><label for=\"sk-estimator-id-464\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">XGBClassifier</label><div class=\"sk-toggleable__content fitted\"><pre>XGBClassifier(base_score=None, booster=None, callbacks=None,\n",
              "              colsample_bylevel=None, colsample_bynode=None,\n",
              "              colsample_bytree=None, device=None, early_stopping_rounds=None,\n",
              "              enable_categorical=False, eval_metric=None, feature_types=None,\n",
              "              gamma=None, grow_policy=None, importance_type=None,\n",
              "              interaction_constraints=None, learning_rate=None, max_bin=None,\n",
              "              max_cat_threshold=None, max_cat_to_onehot=None,\n",
              "              max_delta_step=None, max_depth=None, max_leaves=None,\n",
              "              min_child_weight=None, missing=nan, monotone_constraints=None,\n",
              "              multi_strategy=None, n_estimators=None, n_jobs=None,\n",
              "              num_parallel_tree=None, random_state=42, ...)</pre></div> </div></div></div></div></div></div>"
            ]
          },
          "metadata": {},
          "execution_count": 89
        }
      ],
      "source": [
        "from xgboost import XGBClassifier\n",
        "from sklearn.pipeline import make_pipeline\n",
        "from sklearn.preprocessing import LabelEncoder\n",
        "\n",
        "le = LabelEncoder()\n",
        "y_train_encoded = le.fit_transform(y_train)\n",
        "y_test_encoded = le.transform(y_test)\n",
        "xgb_model = make_pipeline(preprocessing, XGBClassifier(random_state=42))\n",
        "xgb_model.fit(X_train, y_train_encoded)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 87,
      "metadata": {
        "id": "cYFZap-bH9KI",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "90ba7470-53fb-40e9-996e-81c116311620"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "{'bad': 0, 'good': 1}\n"
          ]
        }
      ],
      "source": [
        "print(dict(zip(le.classes_, le.transform(le.classes_))))"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ANN8BPAbEqTs"
      },
      "source": [
        "Testing the accuracy of the model:"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 90,
      "metadata": {
        "id": "HTHkr58LEqTs",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "2f6e51be-08d2-45dc-fdaf-c11911fce20f"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Training Set Performance:\n",
            "Accuracy: 1.0000\n",
            "Precision: 1.0000\n",
            "Recall: 1.0000\n",
            "F1-Score: 1.0000\n",
            "ROC-AUC: 1.0000\n",
            "------------------------------\n"
          ]
        }
      ],
      "source": [
        "from sklearn.metrics import accuracy_score, precision_score, recall_score, f1_score, roc_auc_score\n",
        "\n",
        "train_predictions = xgb_model.predict(X_train)\n",
        "\n",
        "train_accuracy = accuracy_score(y_train_encoded, train_predictions)\n",
        "train_precision = precision_score(y_train_encoded, train_predictions)\n",
        "train_recall = recall_score(y_train_encoded, train_predictions)\n",
        "train_f1 = f1_score(y_train_encoded, train_predictions)\n",
        "train_roc_auc = roc_auc_score(y_train_encoded, xgb_model.predict_proba(X_train)[:, 1])\n",
        "\n",
        "print(\"Training Set Performance:\")\n",
        "print(f\"Accuracy: {train_accuracy:.4f}\")\n",
        "print(f\"Precision: {train_precision:.4f}\")\n",
        "print(f\"Recall: {train_recall:.4f}\")\n",
        "print(f\"F1-Score: {train_f1:.4f}\")\n",
        "print(f\"ROC-AUC: {train_roc_auc:.4f}\")\n",
        "print(\"-\" * 30)\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "PXTRrwoH79w5"
      },
      "source": [
        "### Stochastic Gradient Descent"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "M_p-DOZv79w5"
      },
      "source": [
        "We are now going to be training a random forest model."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 91,
      "metadata": {
        "id": "LUkywjCP79w6",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 357
        },
        "outputId": "42873484-cf67-4b24-cd72-7a52b3a59c7e"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Pipeline(steps=[('columntransformer',\n",
              "                 ColumnTransformer(remainder=Pipeline(steps=[('standardscaler',\n",
              "                                                              StandardScaler())]),\n",
              "                                   transformers=[('monthly_credit_burden',\n",
              "                                                  Pipeline(steps=[('functiontransformer-1',\n",
              "                                                                   FunctionTransformer(feature_names_out=<function ratio_name at 0x78a662f00dc0>,\n",
              "                                                                                       func=<function ratio_pipeline.<locals>.<lambda> at 0x78a6624f3250>)),\n",
              "                                                                  ('function...\n",
              "                                                   'credit_history']),\n",
              "                                                 ('geo',\n",
              "                                                  Pipeline(steps=[('ordered_ordinal_encoder',\n",
              "                                                                   OrderedOrdinalEncoder()),\n",
              "                                                                  ('cluster_simil',\n",
              "                                                                   ClusterSimilarity(random_state=42))]),\n",
              "                                                  ['savings_status',\n",
              "                                                   'checking_status']),\n",
              "                                                 ('drop_out', 'drop',\n",
              "                                                  ['num_dependents',\n",
              "                                                   'foreign_worker',\n",
              "                                                   'purpose'])])),\n",
              "                ('sgdclassifier',\n",
              "                 SGDClassifier(class_weight='balanced', loss='log_loss',\n",
              "                               penalty='elasticnet', random_state=42))])"
            ],
            "text/html": [
              "<style>#sk-container-id-17 {\n",
              "  /* Definition of color scheme common for light and dark mode */\n",
              "  --sklearn-color-text: black;\n",
              "  --sklearn-color-line: gray;\n",
              "  /* Definition of color scheme for unfitted estimators */\n",
              "  --sklearn-color-unfitted-level-0: #fff5e6;\n",
              "  --sklearn-color-unfitted-level-1: #f6e4d2;\n",
              "  --sklearn-color-unfitted-level-2: #ffe0b3;\n",
              "  --sklearn-color-unfitted-level-3: chocolate;\n",
              "  /* Definition of color scheme for fitted estimators */\n",
              "  --sklearn-color-fitted-level-0: #f0f8ff;\n",
              "  --sklearn-color-fitted-level-1: #d4ebff;\n",
              "  --sklearn-color-fitted-level-2: #b3dbfd;\n",
              "  --sklearn-color-fitted-level-3: cornflowerblue;\n",
              "\n",
              "  /* Specific color for light theme */\n",
              "  --sklearn-color-text-on-default-background: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, black)));\n",
              "  --sklearn-color-background: var(--sg-background-color, var(--theme-background, var(--jp-layout-color0, white)));\n",
              "  --sklearn-color-border-box: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, black)));\n",
              "  --sklearn-color-icon: #696969;\n",
              "\n",
              "  @media (prefers-color-scheme: dark) {\n",
              "    /* Redefinition of color scheme for dark theme */\n",
              "    --sklearn-color-text-on-default-background: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, white)));\n",
              "    --sklearn-color-background: var(--sg-background-color, var(--theme-background, var(--jp-layout-color0, #111)));\n",
              "    --sklearn-color-border-box: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, white)));\n",
              "    --sklearn-color-icon: #878787;\n",
              "  }\n",
              "}\n",
              "\n",
              "#sk-container-id-17 {\n",
              "  color: var(--sklearn-color-text);\n",
              "}\n",
              "\n",
              "#sk-container-id-17 pre {\n",
              "  padding: 0;\n",
              "}\n",
              "\n",
              "#sk-container-id-17 input.sk-hidden--visually {\n",
              "  border: 0;\n",
              "  clip: rect(1px 1px 1px 1px);\n",
              "  clip: rect(1px, 1px, 1px, 1px);\n",
              "  height: 1px;\n",
              "  margin: -1px;\n",
              "  overflow: hidden;\n",
              "  padding: 0;\n",
              "  position: absolute;\n",
              "  width: 1px;\n",
              "}\n",
              "\n",
              "#sk-container-id-17 div.sk-dashed-wrapped {\n",
              "  border: 1px dashed var(--sklearn-color-line);\n",
              "  margin: 0 0.4em 0.5em 0.4em;\n",
              "  box-sizing: border-box;\n",
              "  padding-bottom: 0.4em;\n",
              "  background-color: var(--sklearn-color-background);\n",
              "}\n",
              "\n",
              "#sk-container-id-17 div.sk-container {\n",
              "  /* jupyter's `normalize.less` sets `[hidden] { display: none; }`\n",
              "     but bootstrap.min.css set `[hidden] { display: none !important; }`\n",
              "     so we also need the `!important` here to be able to override the\n",
              "     default hidden behavior on the sphinx rendered scikit-learn.org.\n",
              "     See: https://github.com/scikit-learn/scikit-learn/issues/21755 */\n",
              "  display: inline-block !important;\n",
              "  position: relative;\n",
              "}\n",
              "\n",
              "#sk-container-id-17 div.sk-text-repr-fallback {\n",
              "  display: none;\n",
              "}\n",
              "\n",
              "div.sk-parallel-item,\n",
              "div.sk-serial,\n",
              "div.sk-item {\n",
              "  /* draw centered vertical line to link estimators */\n",
              "  background-image: linear-gradient(var(--sklearn-color-text-on-default-background), var(--sklearn-color-text-on-default-background));\n",
              "  background-size: 2px 100%;\n",
              "  background-repeat: no-repeat;\n",
              "  background-position: center center;\n",
              "}\n",
              "\n",
              "/* Parallel-specific style estimator block */\n",
              "\n",
              "#sk-container-id-17 div.sk-parallel-item::after {\n",
              "  content: \"\";\n",
              "  width: 100%;\n",
              "  border-bottom: 2px solid var(--sklearn-color-text-on-default-background);\n",
              "  flex-grow: 1;\n",
              "}\n",
              "\n",
              "#sk-container-id-17 div.sk-parallel {\n",
              "  display: flex;\n",
              "  align-items: stretch;\n",
              "  justify-content: center;\n",
              "  background-color: var(--sklearn-color-background);\n",
              "  position: relative;\n",
              "}\n",
              "\n",
              "#sk-container-id-17 div.sk-parallel-item {\n",
              "  display: flex;\n",
              "  flex-direction: column;\n",
              "}\n",
              "\n",
              "#sk-container-id-17 div.sk-parallel-item:first-child::after {\n",
              "  align-self: flex-end;\n",
              "  width: 50%;\n",
              "}\n",
              "\n",
              "#sk-container-id-17 div.sk-parallel-item:last-child::after {\n",
              "  align-self: flex-start;\n",
              "  width: 50%;\n",
              "}\n",
              "\n",
              "#sk-container-id-17 div.sk-parallel-item:only-child::after {\n",
              "  width: 0;\n",
              "}\n",
              "\n",
              "/* Serial-specific style estimator block */\n",
              "\n",
              "#sk-container-id-17 div.sk-serial {\n",
              "  display: flex;\n",
              "  flex-direction: column;\n",
              "  align-items: center;\n",
              "  background-color: var(--sklearn-color-background);\n",
              "  padding-right: 1em;\n",
              "  padding-left: 1em;\n",
              "}\n",
              "\n",
              "\n",
              "/* Toggleable style: style used for estimator/Pipeline/ColumnTransformer box that is\n",
              "clickable and can be expanded/collapsed.\n",
              "- Pipeline and ColumnTransformer use this feature and define the default style\n",
              "- Estimators will overwrite some part of the style using the `sk-estimator` class\n",
              "*/\n",
              "\n",
              "/* Pipeline and ColumnTransformer style (default) */\n",
              "\n",
              "#sk-container-id-17 div.sk-toggleable {\n",
              "  /* Default theme specific background. It is overwritten whether we have a\n",
              "  specific estimator or a Pipeline/ColumnTransformer */\n",
              "  background-color: var(--sklearn-color-background);\n",
              "}\n",
              "\n",
              "/* Toggleable label */\n",
              "#sk-container-id-17 label.sk-toggleable__label {\n",
              "  cursor: pointer;\n",
              "  display: block;\n",
              "  width: 100%;\n",
              "  margin-bottom: 0;\n",
              "  padding: 0.5em;\n",
              "  box-sizing: border-box;\n",
              "  text-align: center;\n",
              "}\n",
              "\n",
              "#sk-container-id-17 label.sk-toggleable__label-arrow:before {\n",
              "  /* Arrow on the left of the label */\n",
              "  content: \"▸\";\n",
              "  float: left;\n",
              "  margin-right: 0.25em;\n",
              "  color: var(--sklearn-color-icon);\n",
              "}\n",
              "\n",
              "#sk-container-id-17 label.sk-toggleable__label-arrow:hover:before {\n",
              "  color: var(--sklearn-color-text);\n",
              "}\n",
              "\n",
              "/* Toggleable content - dropdown */\n",
              "\n",
              "#sk-container-id-17 div.sk-toggleable__content {\n",
              "  max-height: 0;\n",
              "  max-width: 0;\n",
              "  overflow: hidden;\n",
              "  text-align: left;\n",
              "  /* unfitted */\n",
              "  background-color: var(--sklearn-color-unfitted-level-0);\n",
              "}\n",
              "\n",
              "#sk-container-id-17 div.sk-toggleable__content.fitted {\n",
              "  /* fitted */\n",
              "  background-color: var(--sklearn-color-fitted-level-0);\n",
              "}\n",
              "\n",
              "#sk-container-id-17 div.sk-toggleable__content pre {\n",
              "  margin: 0.2em;\n",
              "  border-radius: 0.25em;\n",
              "  color: var(--sklearn-color-text);\n",
              "  /* unfitted */\n",
              "  background-color: var(--sklearn-color-unfitted-level-0);\n",
              "}\n",
              "\n",
              "#sk-container-id-17 div.sk-toggleable__content.fitted pre {\n",
              "  /* unfitted */\n",
              "  background-color: var(--sklearn-color-fitted-level-0);\n",
              "}\n",
              "\n",
              "#sk-container-id-17 input.sk-toggleable__control:checked~div.sk-toggleable__content {\n",
              "  /* Expand drop-down */\n",
              "  max-height: 200px;\n",
              "  max-width: 100%;\n",
              "  overflow: auto;\n",
              "}\n",
              "\n",
              "#sk-container-id-17 input.sk-toggleable__control:checked~label.sk-toggleable__label-arrow:before {\n",
              "  content: \"▾\";\n",
              "}\n",
              "\n",
              "/* Pipeline/ColumnTransformer-specific style */\n",
              "\n",
              "#sk-container-id-17 div.sk-label input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
              "  color: var(--sklearn-color-text);\n",
              "  background-color: var(--sklearn-color-unfitted-level-2);\n",
              "}\n",
              "\n",
              "#sk-container-id-17 div.sk-label.fitted input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
              "  background-color: var(--sklearn-color-fitted-level-2);\n",
              "}\n",
              "\n",
              "/* Estimator-specific style */\n",
              "\n",
              "/* Colorize estimator box */\n",
              "#sk-container-id-17 div.sk-estimator input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
              "  /* unfitted */\n",
              "  background-color: var(--sklearn-color-unfitted-level-2);\n",
              "}\n",
              "\n",
              "#sk-container-id-17 div.sk-estimator.fitted input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
              "  /* fitted */\n",
              "  background-color: var(--sklearn-color-fitted-level-2);\n",
              "}\n",
              "\n",
              "#sk-container-id-17 div.sk-label label.sk-toggleable__label,\n",
              "#sk-container-id-17 div.sk-label label {\n",
              "  /* The background is the default theme color */\n",
              "  color: var(--sklearn-color-text-on-default-background);\n",
              "}\n",
              "\n",
              "/* On hover, darken the color of the background */\n",
              "#sk-container-id-17 div.sk-label:hover label.sk-toggleable__label {\n",
              "  color: var(--sklearn-color-text);\n",
              "  background-color: var(--sklearn-color-unfitted-level-2);\n",
              "}\n",
              "\n",
              "/* Label box, darken color on hover, fitted */\n",
              "#sk-container-id-17 div.sk-label.fitted:hover label.sk-toggleable__label.fitted {\n",
              "  color: var(--sklearn-color-text);\n",
              "  background-color: var(--sklearn-color-fitted-level-2);\n",
              "}\n",
              "\n",
              "/* Estimator label */\n",
              "\n",
              "#sk-container-id-17 div.sk-label label {\n",
              "  font-family: monospace;\n",
              "  font-weight: bold;\n",
              "  display: inline-block;\n",
              "  line-height: 1.2em;\n",
              "}\n",
              "\n",
              "#sk-container-id-17 div.sk-label-container {\n",
              "  text-align: center;\n",
              "}\n",
              "\n",
              "/* Estimator-specific */\n",
              "#sk-container-id-17 div.sk-estimator {\n",
              "  font-family: monospace;\n",
              "  border: 1px dotted var(--sklearn-color-border-box);\n",
              "  border-radius: 0.25em;\n",
              "  box-sizing: border-box;\n",
              "  margin-bottom: 0.5em;\n",
              "  /* unfitted */\n",
              "  background-color: var(--sklearn-color-unfitted-level-0);\n",
              "}\n",
              "\n",
              "#sk-container-id-17 div.sk-estimator.fitted {\n",
              "  /* fitted */\n",
              "  background-color: var(--sklearn-color-fitted-level-0);\n",
              "}\n",
              "\n",
              "/* on hover */\n",
              "#sk-container-id-17 div.sk-estimator:hover {\n",
              "  /* unfitted */\n",
              "  background-color: var(--sklearn-color-unfitted-level-2);\n",
              "}\n",
              "\n",
              "#sk-container-id-17 div.sk-estimator.fitted:hover {\n",
              "  /* fitted */\n",
              "  background-color: var(--sklearn-color-fitted-level-2);\n",
              "}\n",
              "\n",
              "/* Specification for estimator info (e.g. \"i\" and \"?\") */\n",
              "\n",
              "/* Common style for \"i\" and \"?\" */\n",
              "\n",
              ".sk-estimator-doc-link,\n",
              "a:link.sk-estimator-doc-link,\n",
              "a:visited.sk-estimator-doc-link {\n",
              "  float: right;\n",
              "  font-size: smaller;\n",
              "  line-height: 1em;\n",
              "  font-family: monospace;\n",
              "  background-color: var(--sklearn-color-background);\n",
              "  border-radius: 1em;\n",
              "  height: 1em;\n",
              "  width: 1em;\n",
              "  text-decoration: none !important;\n",
              "  margin-left: 1ex;\n",
              "  /* unfitted */\n",
              "  border: var(--sklearn-color-unfitted-level-1) 1pt solid;\n",
              "  color: var(--sklearn-color-unfitted-level-1);\n",
              "}\n",
              "\n",
              ".sk-estimator-doc-link.fitted,\n",
              "a:link.sk-estimator-doc-link.fitted,\n",
              "a:visited.sk-estimator-doc-link.fitted {\n",
              "  /* fitted */\n",
              "  border: var(--sklearn-color-fitted-level-1) 1pt solid;\n",
              "  color: var(--sklearn-color-fitted-level-1);\n",
              "}\n",
              "\n",
              "/* On hover */\n",
              "div.sk-estimator:hover .sk-estimator-doc-link:hover,\n",
              ".sk-estimator-doc-link:hover,\n",
              "div.sk-label-container:hover .sk-estimator-doc-link:hover,\n",
              ".sk-estimator-doc-link:hover {\n",
              "  /* unfitted */\n",
              "  background-color: var(--sklearn-color-unfitted-level-3);\n",
              "  color: var(--sklearn-color-background);\n",
              "  text-decoration: none;\n",
              "}\n",
              "\n",
              "div.sk-estimator.fitted:hover .sk-estimator-doc-link.fitted:hover,\n",
              ".sk-estimator-doc-link.fitted:hover,\n",
              "div.sk-label-container:hover .sk-estimator-doc-link.fitted:hover,\n",
              ".sk-estimator-doc-link.fitted:hover {\n",
              "  /* fitted */\n",
              "  background-color: var(--sklearn-color-fitted-level-3);\n",
              "  color: var(--sklearn-color-background);\n",
              "  text-decoration: none;\n",
              "}\n",
              "\n",
              "/* Span, style for the box shown on hovering the info icon */\n",
              ".sk-estimator-doc-link span {\n",
              "  display: none;\n",
              "  z-index: 9999;\n",
              "  position: relative;\n",
              "  font-weight: normal;\n",
              "  right: .2ex;\n",
              "  padding: .5ex;\n",
              "  margin: .5ex;\n",
              "  width: min-content;\n",
              "  min-width: 20ex;\n",
              "  max-width: 50ex;\n",
              "  color: var(--sklearn-color-text);\n",
              "  box-shadow: 2pt 2pt 4pt #999;\n",
              "  /* unfitted */\n",
              "  background: var(--sklearn-color-unfitted-level-0);\n",
              "  border: .5pt solid var(--sklearn-color-unfitted-level-3);\n",
              "}\n",
              "\n",
              ".sk-estimator-doc-link.fitted span {\n",
              "  /* fitted */\n",
              "  background: var(--sklearn-color-fitted-level-0);\n",
              "  border: var(--sklearn-color-fitted-level-3);\n",
              "}\n",
              "\n",
              ".sk-estimator-doc-link:hover span {\n",
              "  display: block;\n",
              "}\n",
              "\n",
              "/* \"?\"-specific style due to the `<a>` HTML tag */\n",
              "\n",
              "#sk-container-id-17 a.estimator_doc_link {\n",
              "  float: right;\n",
              "  font-size: 1rem;\n",
              "  line-height: 1em;\n",
              "  font-family: monospace;\n",
              "  background-color: var(--sklearn-color-background);\n",
              "  border-radius: 1rem;\n",
              "  height: 1rem;\n",
              "  width: 1rem;\n",
              "  text-decoration: none;\n",
              "  /* unfitted */\n",
              "  color: var(--sklearn-color-unfitted-level-1);\n",
              "  border: var(--sklearn-color-unfitted-level-1) 1pt solid;\n",
              "}\n",
              "\n",
              "#sk-container-id-17 a.estimator_doc_link.fitted {\n",
              "  /* fitted */\n",
              "  border: var(--sklearn-color-fitted-level-1) 1pt solid;\n",
              "  color: var(--sklearn-color-fitted-level-1);\n",
              "}\n",
              "\n",
              "/* On hover */\n",
              "#sk-container-id-17 a.estimator_doc_link:hover {\n",
              "  /* unfitted */\n",
              "  background-color: var(--sklearn-color-unfitted-level-3);\n",
              "  color: var(--sklearn-color-background);\n",
              "  text-decoration: none;\n",
              "}\n",
              "\n",
              "#sk-container-id-17 a.estimator_doc_link.fitted:hover {\n",
              "  /* fitted */\n",
              "  background-color: var(--sklearn-color-fitted-level-3);\n",
              "}\n",
              "</style><div id=\"sk-container-id-17\" class=\"sk-top-container\"><div class=\"sk-text-repr-fallback\"><pre>Pipeline(steps=[(&#x27;columntransformer&#x27;,\n",
              "                 ColumnTransformer(remainder=Pipeline(steps=[(&#x27;standardscaler&#x27;,\n",
              "                                                              StandardScaler())]),\n",
              "                                   transformers=[(&#x27;monthly_credit_burden&#x27;,\n",
              "                                                  Pipeline(steps=[(&#x27;functiontransformer-1&#x27;,\n",
              "                                                                   FunctionTransformer(feature_names_out=&lt;function ratio_name at 0x78a662f00dc0&gt;,\n",
              "                                                                                       func=&lt;function ratio_pipeline.&lt;locals&gt;.&lt;lambda&gt; at 0x78a6624f3250&gt;)),\n",
              "                                                                  (&#x27;function...\n",
              "                                                   &#x27;credit_history&#x27;]),\n",
              "                                                 (&#x27;geo&#x27;,\n",
              "                                                  Pipeline(steps=[(&#x27;ordered_ordinal_encoder&#x27;,\n",
              "                                                                   OrderedOrdinalEncoder()),\n",
              "                                                                  (&#x27;cluster_simil&#x27;,\n",
              "                                                                   ClusterSimilarity(random_state=42))]),\n",
              "                                                  [&#x27;savings_status&#x27;,\n",
              "                                                   &#x27;checking_status&#x27;]),\n",
              "                                                 (&#x27;drop_out&#x27;, &#x27;drop&#x27;,\n",
              "                                                  [&#x27;num_dependents&#x27;,\n",
              "                                                   &#x27;foreign_worker&#x27;,\n",
              "                                                   &#x27;purpose&#x27;])])),\n",
              "                (&#x27;sgdclassifier&#x27;,\n",
              "                 SGDClassifier(class_weight=&#x27;balanced&#x27;, loss=&#x27;log_loss&#x27;,\n",
              "                               penalty=&#x27;elasticnet&#x27;, random_state=42))])</pre><b>In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook. <br />On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.</b></div><div class=\"sk-container\" hidden><div class=\"sk-item sk-dashed-wrapped\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-465\" type=\"checkbox\" ><label for=\"sk-estimator-id-465\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;&nbsp;Pipeline<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.pipeline.Pipeline.html\">?<span>Documentation for Pipeline</span></a><span class=\"sk-estimator-doc-link fitted\">i<span>Fitted</span></span></label><div class=\"sk-toggleable__content fitted\"><pre>Pipeline(steps=[(&#x27;columntransformer&#x27;,\n",
              "                 ColumnTransformer(remainder=Pipeline(steps=[(&#x27;standardscaler&#x27;,\n",
              "                                                              StandardScaler())]),\n",
              "                                   transformers=[(&#x27;monthly_credit_burden&#x27;,\n",
              "                                                  Pipeline(steps=[(&#x27;functiontransformer-1&#x27;,\n",
              "                                                                   FunctionTransformer(feature_names_out=&lt;function ratio_name at 0x78a662f00dc0&gt;,\n",
              "                                                                                       func=&lt;function ratio_pipeline.&lt;locals&gt;.&lt;lambda&gt; at 0x78a6624f3250&gt;)),\n",
              "                                                                  (&#x27;function...\n",
              "                                                   &#x27;credit_history&#x27;]),\n",
              "                                                 (&#x27;geo&#x27;,\n",
              "                                                  Pipeline(steps=[(&#x27;ordered_ordinal_encoder&#x27;,\n",
              "                                                                   OrderedOrdinalEncoder()),\n",
              "                                                                  (&#x27;cluster_simil&#x27;,\n",
              "                                                                   ClusterSimilarity(random_state=42))]),\n",
              "                                                  [&#x27;savings_status&#x27;,\n",
              "                                                   &#x27;checking_status&#x27;]),\n",
              "                                                 (&#x27;drop_out&#x27;, &#x27;drop&#x27;,\n",
              "                                                  [&#x27;num_dependents&#x27;,\n",
              "                                                   &#x27;foreign_worker&#x27;,\n",
              "                                                   &#x27;purpose&#x27;])])),\n",
              "                (&#x27;sgdclassifier&#x27;,\n",
              "                 SGDClassifier(class_weight=&#x27;balanced&#x27;, loss=&#x27;log_loss&#x27;,\n",
              "                               penalty=&#x27;elasticnet&#x27;, random_state=42))])</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item sk-dashed-wrapped\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-466\" type=\"checkbox\" ><label for=\"sk-estimator-id-466\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;columntransformer: ColumnTransformer<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.compose.ColumnTransformer.html\">?<span>Documentation for columntransformer: ColumnTransformer</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>ColumnTransformer(remainder=Pipeline(steps=[(&#x27;standardscaler&#x27;,\n",
              "                                             StandardScaler())]),\n",
              "                  transformers=[(&#x27;monthly_credit_burden&#x27;,\n",
              "                                 Pipeline(steps=[(&#x27;functiontransformer-1&#x27;,\n",
              "                                                  FunctionTransformer(feature_names_out=&lt;function ratio_name at 0x78a662f00dc0&gt;,\n",
              "                                                                      func=&lt;function ratio_pipeline.&lt;locals&gt;.&lt;lambda&gt; at 0x78a6624f3250&gt;)),\n",
              "                                                 (&#x27;functiontransformer-2&#x27;,\n",
              "                                                  FunctionTransformer(fe...\n",
              "                                                  OneHotEncoder(handle_unknown=&#x27;ignore&#x27;))]),\n",
              "                                 [&#x27;job&#x27;, &#x27;housing&#x27;, &#x27;other_payment_plans&#x27;,\n",
              "                                  &#x27;property_magnitude&#x27;, &#x27;other_parties&#x27;,\n",
              "                                  &#x27;personal_status&#x27;, &#x27;credit_history&#x27;]),\n",
              "                                (&#x27;geo&#x27;,\n",
              "                                 Pipeline(steps=[(&#x27;ordered_ordinal_encoder&#x27;,\n",
              "                                                  OrderedOrdinalEncoder()),\n",
              "                                                 (&#x27;cluster_simil&#x27;,\n",
              "                                                  ClusterSimilarity(random_state=42))]),\n",
              "                                 [&#x27;savings_status&#x27;, &#x27;checking_status&#x27;]),\n",
              "                                (&#x27;drop_out&#x27;, &#x27;drop&#x27;,\n",
              "                                 [&#x27;num_dependents&#x27;, &#x27;foreign_worker&#x27;,\n",
              "                                  &#x27;purpose&#x27;])])</pre></div> </div></div><div class=\"sk-parallel\"><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-467\" type=\"checkbox\" ><label for=\"sk-estimator-id-467\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">monthly_credit_burden</label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;credit_amount&#x27;, &#x27;duration&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-468\" type=\"checkbox\" ><label for=\"sk-estimator-id-468\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;FunctionTransformer<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.FunctionTransformer.html\">?<span>Documentation for FunctionTransformer</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>FunctionTransformer(feature_names_out=&lt;function ratio_name at 0x78a662f00dc0&gt;,\n",
              "                    func=&lt;function ratio_pipeline.&lt;locals&gt;.&lt;lambda&gt; at 0x78a6624f3250&gt;)</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-469\" type=\"checkbox\" ><label for=\"sk-estimator-id-469\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;FunctionTransformer<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.FunctionTransformer.html\">?<span>Documentation for FunctionTransformer</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>FunctionTransformer(feature_names_out=&#x27;one-to-one&#x27;, func=&lt;ufunc &#x27;log&#x27;&gt;,\n",
              "                    inverse_func=&lt;ufunc &#x27;exp&#x27;&gt;)</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-470\" type=\"checkbox\" ><label for=\"sk-estimator-id-470\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;StandardScaler<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.StandardScaler.html\">?<span>Documentation for StandardScaler</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>StandardScaler()</pre></div> </div></div></div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-471\" type=\"checkbox\" ><label for=\"sk-estimator-id-471\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">installment_per_credit</label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;installment_commitment&#x27;, &#x27;credit_amount&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-472\" type=\"checkbox\" ><label for=\"sk-estimator-id-472\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;FunctionTransformer<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.FunctionTransformer.html\">?<span>Documentation for FunctionTransformer</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>FunctionTransformer(feature_names_out=&lt;function ratio_name at 0x78a662f00dc0&gt;,\n",
              "                    func=&lt;function ratio_pipeline.&lt;locals&gt;.&lt;lambda&gt; at 0x78a6624f3400&gt;)</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-473\" type=\"checkbox\" ><label for=\"sk-estimator-id-473\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;FunctionTransformer<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.FunctionTransformer.html\">?<span>Documentation for FunctionTransformer</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>FunctionTransformer(feature_names_out=&#x27;one-to-one&#x27;, func=&lt;ufunc &#x27;log&#x27;&gt;,\n",
              "                    inverse_func=&lt;ufunc &#x27;exp&#x27;&gt;)</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-474\" type=\"checkbox\" ><label for=\"sk-estimator-id-474\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;StandardScaler<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.StandardScaler.html\">?<span>Documentation for StandardScaler</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>StandardScaler()</pre></div> </div></div></div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-475\" type=\"checkbox\" ><label for=\"sk-estimator-id-475\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">dependents_per_credit</label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;num_dependents&#x27;, &#x27;credit_amount&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-476\" type=\"checkbox\" ><label for=\"sk-estimator-id-476\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;FunctionTransformer<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.FunctionTransformer.html\">?<span>Documentation for FunctionTransformer</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>FunctionTransformer(feature_names_out=&lt;function ratio_name at 0x78a662f00dc0&gt;,\n",
              "                    func=&lt;function ratio_pipeline.&lt;locals&gt;.&lt;lambda&gt; at 0x78a6624f35b0&gt;)</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-477\" type=\"checkbox\" ><label for=\"sk-estimator-id-477\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;FunctionTransformer<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.FunctionTransformer.html\">?<span>Documentation for FunctionTransformer</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>FunctionTransformer(feature_names_out=&#x27;one-to-one&#x27;, func=&lt;ufunc &#x27;log&#x27;&gt;,\n",
              "                    inverse_func=&lt;ufunc &#x27;exp&#x27;&gt;)</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-478\" type=\"checkbox\" ><label for=\"sk-estimator-id-478\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;StandardScaler<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.StandardScaler.html\">?<span>Documentation for StandardScaler</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>StandardScaler()</pre></div> </div></div></div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-479\" type=\"checkbox\" ><label for=\"sk-estimator-id-479\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">log</label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;duration&#x27;, &#x27;credit_amount&#x27;, &#x27;installment_commitment&#x27;, &#x27;residence_since&#x27;, &#x27;age&#x27;, &#x27;existing_credits&#x27;, &#x27;num_dependents&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-480\" type=\"checkbox\" ><label for=\"sk-estimator-id-480\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;FunctionTransformer<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.FunctionTransformer.html\">?<span>Documentation for FunctionTransformer</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>FunctionTransformer(feature_names_out=&#x27;one-to-one&#x27;, func=&lt;ufunc &#x27;log&#x27;&gt;,\n",
              "                    inverse_func=&lt;ufunc &#x27;exp&#x27;&gt;)</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-481\" type=\"checkbox\" ><label for=\"sk-estimator-id-481\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;StandardScaler<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.StandardScaler.html\">?<span>Documentation for StandardScaler</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>StandardScaler()</pre></div> </div></div></div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-482\" type=\"checkbox\" ><label for=\"sk-estimator-id-482\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">cat_ordinal</label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;checking_status&#x27;, &#x27;savings_status&#x27;, &#x27;employment&#x27;, &#x27;own_telephone&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-483\" type=\"checkbox\" ><label for=\"sk-estimator-id-483\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">OrderedOrdinalEncoder</label><div class=\"sk-toggleable__content fitted\"><pre>OrderedOrdinalEncoder()</pre></div> </div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-484\" type=\"checkbox\" ><label for=\"sk-estimator-id-484\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">cat_onehot</label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;job&#x27;, &#x27;housing&#x27;, &#x27;other_payment_plans&#x27;, &#x27;property_magnitude&#x27;, &#x27;other_parties&#x27;, &#x27;personal_status&#x27;, &#x27;credit_history&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-485\" type=\"checkbox\" ><label for=\"sk-estimator-id-485\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;OneHotEncoder<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.OneHotEncoder.html\">?<span>Documentation for OneHotEncoder</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>OneHotEncoder(handle_unknown=&#x27;ignore&#x27;)</pre></div> </div></div></div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-486\" type=\"checkbox\" ><label for=\"sk-estimator-id-486\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">geo</label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;savings_status&#x27;, &#x27;checking_status&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-487\" type=\"checkbox\" ><label for=\"sk-estimator-id-487\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">OrderedOrdinalEncoder</label><div class=\"sk-toggleable__content fitted\"><pre>OrderedOrdinalEncoder()</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-488\" type=\"checkbox\" ><label for=\"sk-estimator-id-488\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">ClusterSimilarity</label><div class=\"sk-toggleable__content fitted\"><pre>ClusterSimilarity(random_state=42)</pre></div> </div></div></div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-489\" type=\"checkbox\" ><label for=\"sk-estimator-id-489\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">drop_out</label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;num_dependents&#x27;, &#x27;foreign_worker&#x27;, &#x27;purpose&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-490\" type=\"checkbox\" ><label for=\"sk-estimator-id-490\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">drop</label><div class=\"sk-toggleable__content fitted\"><pre>drop</pre></div> </div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-491\" type=\"checkbox\" ><label for=\"sk-estimator-id-491\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">remainder</label><div class=\"sk-toggleable__content fitted\"><pre>[]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-492\" type=\"checkbox\" ><label for=\"sk-estimator-id-492\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;StandardScaler<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.StandardScaler.html\">?<span>Documentation for StandardScaler</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>StandardScaler()</pre></div> </div></div></div></div></div></div></div></div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-493\" type=\"checkbox\" ><label for=\"sk-estimator-id-493\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;SGDClassifier<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.linear_model.SGDClassifier.html\">?<span>Documentation for SGDClassifier</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>SGDClassifier(class_weight=&#x27;balanced&#x27;, loss=&#x27;log_loss&#x27;, penalty=&#x27;elasticnet&#x27;,\n",
              "              random_state=42)</pre></div> </div></div></div></div></div></div>"
            ]
          },
          "metadata": {},
          "execution_count": 91
        }
      ],
      "source": [
        "from sklearn.linear_model import SGDClassifier\n",
        "from sklearn.pipeline import make_pipeline\n",
        "\n",
        "sgd_model = make_pipeline(preprocessing, SGDClassifier(random_state=42, loss='log_loss', penalty='elasticnet', learning_rate='optimal', class_weight='balanced'))\n",
        "sgd_model.fit(X_train, y_train)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "by_bXfER79w8"
      },
      "source": [
        "Testing the accuracy of the model:"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 93,
      "metadata": {
        "id": "faxRdoB_79w8",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "eeb11545-fad7-4d00-d9c7-4d148642a24e"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Training Set Performance:\n",
            "Accuracy: 0.6850\n",
            "Precision: 0.8548\n",
            "Recall: 0.6625\n",
            "F1-Score: 0.7465\n",
            "ROC-AUC: 0.7771\n",
            "------------------------------\n"
          ]
        }
      ],
      "source": [
        "from sklearn.metrics import accuracy_score, precision_score, recall_score, f1_score, roc_auc_score\n",
        "\n",
        "# Make predictions\n",
        "train_predictions = sgd_model.predict(X_train)\n",
        "\n",
        "# Evaluation metrics\n",
        "train_accuracy = accuracy_score(y_train, train_predictions)\n",
        "train_precision = precision_score(y_train, train_predictions, pos_label='good')\n",
        "train_recall = recall_score(y_train, train_predictions, pos_label='good')\n",
        "train_f1 = f1_score(y_train, train_predictions, pos_label='good')\n",
        "train_roc_auc = roc_auc_score(y_train, sgd_model.predict_proba(X_train)[:, 1])\n",
        "\n",
        "# Print training set results\n",
        "print(\"Training Set Performance:\")\n",
        "print(f\"Accuracy: {train_accuracy:.4f}\")\n",
        "print(f\"Precision: {train_precision:.4f}\")\n",
        "print(f\"Recall: {train_recall:.4f}\")\n",
        "print(f\"F1-Score: {train_f1:.4f}\")\n",
        "print(f\"ROC-AUC: {train_roc_auc:.4f}\")\n",
        "print(\"-\" * 30)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "mpb0vYG00CYE"
      },
      "source": [
        "### Ensemble Method"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "_uL4suol0CYE"
      },
      "source": [
        "We are now going to be training a random forest model."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 97,
      "metadata": {
        "id": "Us-xaF6N0CYF",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 467
        },
        "outputId": "79769c46-2454-473f-8398-3b4f5f8e37e5"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Pipeline(steps=[('preprocessing',\n",
              "                 ColumnTransformer(remainder=Pipeline(steps=[('standardscaler',\n",
              "                                                              StandardScaler())]),\n",
              "                                   transformers=[('monthly_credit_burden',\n",
              "                                                  Pipeline(steps=[('functiontransformer-1',\n",
              "                                                                   FunctionTransformer(feature_names_out=<function ratio_name at 0x78a662f00dc0>,\n",
              "                                                                                       func=<function ratio_pipeline.<locals>.<lambda> at 0x78a6624f3250>)),\n",
              "                                                                  ('functiontran...\n",
              "                                                               interaction_constraints=None,\n",
              "                                                               learning_rate=0.1,\n",
              "                                                               max_bin=None,\n",
              "                                                               max_cat_threshold=None,\n",
              "                                                               max_cat_to_onehot=None,\n",
              "                                                               max_delta_step=None,\n",
              "                                                               max_depth=5,\n",
              "                                                               max_leaves=None,\n",
              "                                                               min_child_weight=None,\n",
              "                                                               missing=nan,\n",
              "                                                               monotone_constraints=None,\n",
              "                                                               multi_strategy=None,\n",
              "                                                               n_estimators=100,\n",
              "                                                               n_jobs=None,\n",
              "                                                               num_parallel_tree=None,\n",
              "                                                               random_state=42, ...))],\n",
              "                                    final_estimator=LogisticRegression(random_state=42)))])"
            ],
            "text/html": [
              "<style>#sk-container-id-18 {\n",
              "  /* Definition of color scheme common for light and dark mode */\n",
              "  --sklearn-color-text: black;\n",
              "  --sklearn-color-line: gray;\n",
              "  /* Definition of color scheme for unfitted estimators */\n",
              "  --sklearn-color-unfitted-level-0: #fff5e6;\n",
              "  --sklearn-color-unfitted-level-1: #f6e4d2;\n",
              "  --sklearn-color-unfitted-level-2: #ffe0b3;\n",
              "  --sklearn-color-unfitted-level-3: chocolate;\n",
              "  /* Definition of color scheme for fitted estimators */\n",
              "  --sklearn-color-fitted-level-0: #f0f8ff;\n",
              "  --sklearn-color-fitted-level-1: #d4ebff;\n",
              "  --sklearn-color-fitted-level-2: #b3dbfd;\n",
              "  --sklearn-color-fitted-level-3: cornflowerblue;\n",
              "\n",
              "  /* Specific color for light theme */\n",
              "  --sklearn-color-text-on-default-background: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, black)));\n",
              "  --sklearn-color-background: var(--sg-background-color, var(--theme-background, var(--jp-layout-color0, white)));\n",
              "  --sklearn-color-border-box: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, black)));\n",
              "  --sklearn-color-icon: #696969;\n",
              "\n",
              "  @media (prefers-color-scheme: dark) {\n",
              "    /* Redefinition of color scheme for dark theme */\n",
              "    --sklearn-color-text-on-default-background: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, white)));\n",
              "    --sklearn-color-background: var(--sg-background-color, var(--theme-background, var(--jp-layout-color0, #111)));\n",
              "    --sklearn-color-border-box: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, white)));\n",
              "    --sklearn-color-icon: #878787;\n",
              "  }\n",
              "}\n",
              "\n",
              "#sk-container-id-18 {\n",
              "  color: var(--sklearn-color-text);\n",
              "}\n",
              "\n",
              "#sk-container-id-18 pre {\n",
              "  padding: 0;\n",
              "}\n",
              "\n",
              "#sk-container-id-18 input.sk-hidden--visually {\n",
              "  border: 0;\n",
              "  clip: rect(1px 1px 1px 1px);\n",
              "  clip: rect(1px, 1px, 1px, 1px);\n",
              "  height: 1px;\n",
              "  margin: -1px;\n",
              "  overflow: hidden;\n",
              "  padding: 0;\n",
              "  position: absolute;\n",
              "  width: 1px;\n",
              "}\n",
              "\n",
              "#sk-container-id-18 div.sk-dashed-wrapped {\n",
              "  border: 1px dashed var(--sklearn-color-line);\n",
              "  margin: 0 0.4em 0.5em 0.4em;\n",
              "  box-sizing: border-box;\n",
              "  padding-bottom: 0.4em;\n",
              "  background-color: var(--sklearn-color-background);\n",
              "}\n",
              "\n",
              "#sk-container-id-18 div.sk-container {\n",
              "  /* jupyter's `normalize.less` sets `[hidden] { display: none; }`\n",
              "     but bootstrap.min.css set `[hidden] { display: none !important; }`\n",
              "     so we also need the `!important` here to be able to override the\n",
              "     default hidden behavior on the sphinx rendered scikit-learn.org.\n",
              "     See: https://github.com/scikit-learn/scikit-learn/issues/21755 */\n",
              "  display: inline-block !important;\n",
              "  position: relative;\n",
              "}\n",
              "\n",
              "#sk-container-id-18 div.sk-text-repr-fallback {\n",
              "  display: none;\n",
              "}\n",
              "\n",
              "div.sk-parallel-item,\n",
              "div.sk-serial,\n",
              "div.sk-item {\n",
              "  /* draw centered vertical line to link estimators */\n",
              "  background-image: linear-gradient(var(--sklearn-color-text-on-default-background), var(--sklearn-color-text-on-default-background));\n",
              "  background-size: 2px 100%;\n",
              "  background-repeat: no-repeat;\n",
              "  background-position: center center;\n",
              "}\n",
              "\n",
              "/* Parallel-specific style estimator block */\n",
              "\n",
              "#sk-container-id-18 div.sk-parallel-item::after {\n",
              "  content: \"\";\n",
              "  width: 100%;\n",
              "  border-bottom: 2px solid var(--sklearn-color-text-on-default-background);\n",
              "  flex-grow: 1;\n",
              "}\n",
              "\n",
              "#sk-container-id-18 div.sk-parallel {\n",
              "  display: flex;\n",
              "  align-items: stretch;\n",
              "  justify-content: center;\n",
              "  background-color: var(--sklearn-color-background);\n",
              "  position: relative;\n",
              "}\n",
              "\n",
              "#sk-container-id-18 div.sk-parallel-item {\n",
              "  display: flex;\n",
              "  flex-direction: column;\n",
              "}\n",
              "\n",
              "#sk-container-id-18 div.sk-parallel-item:first-child::after {\n",
              "  align-self: flex-end;\n",
              "  width: 50%;\n",
              "}\n",
              "\n",
              "#sk-container-id-18 div.sk-parallel-item:last-child::after {\n",
              "  align-self: flex-start;\n",
              "  width: 50%;\n",
              "}\n",
              "\n",
              "#sk-container-id-18 div.sk-parallel-item:only-child::after {\n",
              "  width: 0;\n",
              "}\n",
              "\n",
              "/* Serial-specific style estimator block */\n",
              "\n",
              "#sk-container-id-18 div.sk-serial {\n",
              "  display: flex;\n",
              "  flex-direction: column;\n",
              "  align-items: center;\n",
              "  background-color: var(--sklearn-color-background);\n",
              "  padding-right: 1em;\n",
              "  padding-left: 1em;\n",
              "}\n",
              "\n",
              "\n",
              "/* Toggleable style: style used for estimator/Pipeline/ColumnTransformer box that is\n",
              "clickable and can be expanded/collapsed.\n",
              "- Pipeline and ColumnTransformer use this feature and define the default style\n",
              "- Estimators will overwrite some part of the style using the `sk-estimator` class\n",
              "*/\n",
              "\n",
              "/* Pipeline and ColumnTransformer style (default) */\n",
              "\n",
              "#sk-container-id-18 div.sk-toggleable {\n",
              "  /* Default theme specific background. It is overwritten whether we have a\n",
              "  specific estimator or a Pipeline/ColumnTransformer */\n",
              "  background-color: var(--sklearn-color-background);\n",
              "}\n",
              "\n",
              "/* Toggleable label */\n",
              "#sk-container-id-18 label.sk-toggleable__label {\n",
              "  cursor: pointer;\n",
              "  display: block;\n",
              "  width: 100%;\n",
              "  margin-bottom: 0;\n",
              "  padding: 0.5em;\n",
              "  box-sizing: border-box;\n",
              "  text-align: center;\n",
              "}\n",
              "\n",
              "#sk-container-id-18 label.sk-toggleable__label-arrow:before {\n",
              "  /* Arrow on the left of the label */\n",
              "  content: \"▸\";\n",
              "  float: left;\n",
              "  margin-right: 0.25em;\n",
              "  color: var(--sklearn-color-icon);\n",
              "}\n",
              "\n",
              "#sk-container-id-18 label.sk-toggleable__label-arrow:hover:before {\n",
              "  color: var(--sklearn-color-text);\n",
              "}\n",
              "\n",
              "/* Toggleable content - dropdown */\n",
              "\n",
              "#sk-container-id-18 div.sk-toggleable__content {\n",
              "  max-height: 0;\n",
              "  max-width: 0;\n",
              "  overflow: hidden;\n",
              "  text-align: left;\n",
              "  /* unfitted */\n",
              "  background-color: var(--sklearn-color-unfitted-level-0);\n",
              "}\n",
              "\n",
              "#sk-container-id-18 div.sk-toggleable__content.fitted {\n",
              "  /* fitted */\n",
              "  background-color: var(--sklearn-color-fitted-level-0);\n",
              "}\n",
              "\n",
              "#sk-container-id-18 div.sk-toggleable__content pre {\n",
              "  margin: 0.2em;\n",
              "  border-radius: 0.25em;\n",
              "  color: var(--sklearn-color-text);\n",
              "  /* unfitted */\n",
              "  background-color: var(--sklearn-color-unfitted-level-0);\n",
              "}\n",
              "\n",
              "#sk-container-id-18 div.sk-toggleable__content.fitted pre {\n",
              "  /* unfitted */\n",
              "  background-color: var(--sklearn-color-fitted-level-0);\n",
              "}\n",
              "\n",
              "#sk-container-id-18 input.sk-toggleable__control:checked~div.sk-toggleable__content {\n",
              "  /* Expand drop-down */\n",
              "  max-height: 200px;\n",
              "  max-width: 100%;\n",
              "  overflow: auto;\n",
              "}\n",
              "\n",
              "#sk-container-id-18 input.sk-toggleable__control:checked~label.sk-toggleable__label-arrow:before {\n",
              "  content: \"▾\";\n",
              "}\n",
              "\n",
              "/* Pipeline/ColumnTransformer-specific style */\n",
              "\n",
              "#sk-container-id-18 div.sk-label input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
              "  color: var(--sklearn-color-text);\n",
              "  background-color: var(--sklearn-color-unfitted-level-2);\n",
              "}\n",
              "\n",
              "#sk-container-id-18 div.sk-label.fitted input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
              "  background-color: var(--sklearn-color-fitted-level-2);\n",
              "}\n",
              "\n",
              "/* Estimator-specific style */\n",
              "\n",
              "/* Colorize estimator box */\n",
              "#sk-container-id-18 div.sk-estimator input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
              "  /* unfitted */\n",
              "  background-color: var(--sklearn-color-unfitted-level-2);\n",
              "}\n",
              "\n",
              "#sk-container-id-18 div.sk-estimator.fitted input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
              "  /* fitted */\n",
              "  background-color: var(--sklearn-color-fitted-level-2);\n",
              "}\n",
              "\n",
              "#sk-container-id-18 div.sk-label label.sk-toggleable__label,\n",
              "#sk-container-id-18 div.sk-label label {\n",
              "  /* The background is the default theme color */\n",
              "  color: var(--sklearn-color-text-on-default-background);\n",
              "}\n",
              "\n",
              "/* On hover, darken the color of the background */\n",
              "#sk-container-id-18 div.sk-label:hover label.sk-toggleable__label {\n",
              "  color: var(--sklearn-color-text);\n",
              "  background-color: var(--sklearn-color-unfitted-level-2);\n",
              "}\n",
              "\n",
              "/* Label box, darken color on hover, fitted */\n",
              "#sk-container-id-18 div.sk-label.fitted:hover label.sk-toggleable__label.fitted {\n",
              "  color: var(--sklearn-color-text);\n",
              "  background-color: var(--sklearn-color-fitted-level-2);\n",
              "}\n",
              "\n",
              "/* Estimator label */\n",
              "\n",
              "#sk-container-id-18 div.sk-label label {\n",
              "  font-family: monospace;\n",
              "  font-weight: bold;\n",
              "  display: inline-block;\n",
              "  line-height: 1.2em;\n",
              "}\n",
              "\n",
              "#sk-container-id-18 div.sk-label-container {\n",
              "  text-align: center;\n",
              "}\n",
              "\n",
              "/* Estimator-specific */\n",
              "#sk-container-id-18 div.sk-estimator {\n",
              "  font-family: monospace;\n",
              "  border: 1px dotted var(--sklearn-color-border-box);\n",
              "  border-radius: 0.25em;\n",
              "  box-sizing: border-box;\n",
              "  margin-bottom: 0.5em;\n",
              "  /* unfitted */\n",
              "  background-color: var(--sklearn-color-unfitted-level-0);\n",
              "}\n",
              "\n",
              "#sk-container-id-18 div.sk-estimator.fitted {\n",
              "  /* fitted */\n",
              "  background-color: var(--sklearn-color-fitted-level-0);\n",
              "}\n",
              "\n",
              "/* on hover */\n",
              "#sk-container-id-18 div.sk-estimator:hover {\n",
              "  /* unfitted */\n",
              "  background-color: var(--sklearn-color-unfitted-level-2);\n",
              "}\n",
              "\n",
              "#sk-container-id-18 div.sk-estimator.fitted:hover {\n",
              "  /* fitted */\n",
              "  background-color: var(--sklearn-color-fitted-level-2);\n",
              "}\n",
              "\n",
              "/* Specification for estimator info (e.g. \"i\" and \"?\") */\n",
              "\n",
              "/* Common style for \"i\" and \"?\" */\n",
              "\n",
              ".sk-estimator-doc-link,\n",
              "a:link.sk-estimator-doc-link,\n",
              "a:visited.sk-estimator-doc-link {\n",
              "  float: right;\n",
              "  font-size: smaller;\n",
              "  line-height: 1em;\n",
              "  font-family: monospace;\n",
              "  background-color: var(--sklearn-color-background);\n",
              "  border-radius: 1em;\n",
              "  height: 1em;\n",
              "  width: 1em;\n",
              "  text-decoration: none !important;\n",
              "  margin-left: 1ex;\n",
              "  /* unfitted */\n",
              "  border: var(--sklearn-color-unfitted-level-1) 1pt solid;\n",
              "  color: var(--sklearn-color-unfitted-level-1);\n",
              "}\n",
              "\n",
              ".sk-estimator-doc-link.fitted,\n",
              "a:link.sk-estimator-doc-link.fitted,\n",
              "a:visited.sk-estimator-doc-link.fitted {\n",
              "  /* fitted */\n",
              "  border: var(--sklearn-color-fitted-level-1) 1pt solid;\n",
              "  color: var(--sklearn-color-fitted-level-1);\n",
              "}\n",
              "\n",
              "/* On hover */\n",
              "div.sk-estimator:hover .sk-estimator-doc-link:hover,\n",
              ".sk-estimator-doc-link:hover,\n",
              "div.sk-label-container:hover .sk-estimator-doc-link:hover,\n",
              ".sk-estimator-doc-link:hover {\n",
              "  /* unfitted */\n",
              "  background-color: var(--sklearn-color-unfitted-level-3);\n",
              "  color: var(--sklearn-color-background);\n",
              "  text-decoration: none;\n",
              "}\n",
              "\n",
              "div.sk-estimator.fitted:hover .sk-estimator-doc-link.fitted:hover,\n",
              ".sk-estimator-doc-link.fitted:hover,\n",
              "div.sk-label-container:hover .sk-estimator-doc-link.fitted:hover,\n",
              ".sk-estimator-doc-link.fitted:hover {\n",
              "  /* fitted */\n",
              "  background-color: var(--sklearn-color-fitted-level-3);\n",
              "  color: var(--sklearn-color-background);\n",
              "  text-decoration: none;\n",
              "}\n",
              "\n",
              "/* Span, style for the box shown on hovering the info icon */\n",
              ".sk-estimator-doc-link span {\n",
              "  display: none;\n",
              "  z-index: 9999;\n",
              "  position: relative;\n",
              "  font-weight: normal;\n",
              "  right: .2ex;\n",
              "  padding: .5ex;\n",
              "  margin: .5ex;\n",
              "  width: min-content;\n",
              "  min-width: 20ex;\n",
              "  max-width: 50ex;\n",
              "  color: var(--sklearn-color-text);\n",
              "  box-shadow: 2pt 2pt 4pt #999;\n",
              "  /* unfitted */\n",
              "  background: var(--sklearn-color-unfitted-level-0);\n",
              "  border: .5pt solid var(--sklearn-color-unfitted-level-3);\n",
              "}\n",
              "\n",
              ".sk-estimator-doc-link.fitted span {\n",
              "  /* fitted */\n",
              "  background: var(--sklearn-color-fitted-level-0);\n",
              "  border: var(--sklearn-color-fitted-level-3);\n",
              "}\n",
              "\n",
              ".sk-estimator-doc-link:hover span {\n",
              "  display: block;\n",
              "}\n",
              "\n",
              "/* \"?\"-specific style due to the `<a>` HTML tag */\n",
              "\n",
              "#sk-container-id-18 a.estimator_doc_link {\n",
              "  float: right;\n",
              "  font-size: 1rem;\n",
              "  line-height: 1em;\n",
              "  font-family: monospace;\n",
              "  background-color: var(--sklearn-color-background);\n",
              "  border-radius: 1rem;\n",
              "  height: 1rem;\n",
              "  width: 1rem;\n",
              "  text-decoration: none;\n",
              "  /* unfitted */\n",
              "  color: var(--sklearn-color-unfitted-level-1);\n",
              "  border: var(--sklearn-color-unfitted-level-1) 1pt solid;\n",
              "}\n",
              "\n",
              "#sk-container-id-18 a.estimator_doc_link.fitted {\n",
              "  /* fitted */\n",
              "  border: var(--sklearn-color-fitted-level-1) 1pt solid;\n",
              "  color: var(--sklearn-color-fitted-level-1);\n",
              "}\n",
              "\n",
              "/* On hover */\n",
              "#sk-container-id-18 a.estimator_doc_link:hover {\n",
              "  /* unfitted */\n",
              "  background-color: var(--sklearn-color-unfitted-level-3);\n",
              "  color: var(--sklearn-color-background);\n",
              "  text-decoration: none;\n",
              "}\n",
              "\n",
              "#sk-container-id-18 a.estimator_doc_link.fitted:hover {\n",
              "  /* fitted */\n",
              "  background-color: var(--sklearn-color-fitted-level-3);\n",
              "}\n",
              "</style><div id=\"sk-container-id-18\" class=\"sk-top-container\"><div class=\"sk-text-repr-fallback\"><pre>Pipeline(steps=[(&#x27;preprocessing&#x27;,\n",
              "                 ColumnTransformer(remainder=Pipeline(steps=[(&#x27;standardscaler&#x27;,\n",
              "                                                              StandardScaler())]),\n",
              "                                   transformers=[(&#x27;monthly_credit_burden&#x27;,\n",
              "                                                  Pipeline(steps=[(&#x27;functiontransformer-1&#x27;,\n",
              "                                                                   FunctionTransformer(feature_names_out=&lt;function ratio_name at 0x78a662f00dc0&gt;,\n",
              "                                                                                       func=&lt;function ratio_pipeline.&lt;locals&gt;.&lt;lambda&gt; at 0x78a6624f3250&gt;)),\n",
              "                                                                  (&#x27;functiontran...\n",
              "                                                               interaction_constraints=None,\n",
              "                                                               learning_rate=0.1,\n",
              "                                                               max_bin=None,\n",
              "                                                               max_cat_threshold=None,\n",
              "                                                               max_cat_to_onehot=None,\n",
              "                                                               max_delta_step=None,\n",
              "                                                               max_depth=5,\n",
              "                                                               max_leaves=None,\n",
              "                                                               min_child_weight=None,\n",
              "                                                               missing=nan,\n",
              "                                                               monotone_constraints=None,\n",
              "                                                               multi_strategy=None,\n",
              "                                                               n_estimators=100,\n",
              "                                                               n_jobs=None,\n",
              "                                                               num_parallel_tree=None,\n",
              "                                                               random_state=42, ...))],\n",
              "                                    final_estimator=LogisticRegression(random_state=42)))])</pre><b>In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook. <br />On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.</b></div><div class=\"sk-container\" hidden><div class=\"sk-item sk-dashed-wrapped\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-494\" type=\"checkbox\" ><label for=\"sk-estimator-id-494\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;&nbsp;Pipeline<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.pipeline.Pipeline.html\">?<span>Documentation for Pipeline</span></a><span class=\"sk-estimator-doc-link fitted\">i<span>Fitted</span></span></label><div class=\"sk-toggleable__content fitted\"><pre>Pipeline(steps=[(&#x27;preprocessing&#x27;,\n",
              "                 ColumnTransformer(remainder=Pipeline(steps=[(&#x27;standardscaler&#x27;,\n",
              "                                                              StandardScaler())]),\n",
              "                                   transformers=[(&#x27;monthly_credit_burden&#x27;,\n",
              "                                                  Pipeline(steps=[(&#x27;functiontransformer-1&#x27;,\n",
              "                                                                   FunctionTransformer(feature_names_out=&lt;function ratio_name at 0x78a662f00dc0&gt;,\n",
              "                                                                                       func=&lt;function ratio_pipeline.&lt;locals&gt;.&lt;lambda&gt; at 0x78a6624f3250&gt;)),\n",
              "                                                                  (&#x27;functiontran...\n",
              "                                                               interaction_constraints=None,\n",
              "                                                               learning_rate=0.1,\n",
              "                                                               max_bin=None,\n",
              "                                                               max_cat_threshold=None,\n",
              "                                                               max_cat_to_onehot=None,\n",
              "                                                               max_delta_step=None,\n",
              "                                                               max_depth=5,\n",
              "                                                               max_leaves=None,\n",
              "                                                               min_child_weight=None,\n",
              "                                                               missing=nan,\n",
              "                                                               monotone_constraints=None,\n",
              "                                                               multi_strategy=None,\n",
              "                                                               n_estimators=100,\n",
              "                                                               n_jobs=None,\n",
              "                                                               num_parallel_tree=None,\n",
              "                                                               random_state=42, ...))],\n",
              "                                    final_estimator=LogisticRegression(random_state=42)))])</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item sk-dashed-wrapped\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-495\" type=\"checkbox\" ><label for=\"sk-estimator-id-495\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;preprocessing: ColumnTransformer<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.compose.ColumnTransformer.html\">?<span>Documentation for preprocessing: ColumnTransformer</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>ColumnTransformer(remainder=Pipeline(steps=[(&#x27;standardscaler&#x27;,\n",
              "                                             StandardScaler())]),\n",
              "                  transformers=[(&#x27;monthly_credit_burden&#x27;,\n",
              "                                 Pipeline(steps=[(&#x27;functiontransformer-1&#x27;,\n",
              "                                                  FunctionTransformer(feature_names_out=&lt;function ratio_name at 0x78a662f00dc0&gt;,\n",
              "                                                                      func=&lt;function ratio_pipeline.&lt;locals&gt;.&lt;lambda&gt; at 0x78a6624f3250&gt;)),\n",
              "                                                 (&#x27;functiontransformer-2&#x27;,\n",
              "                                                  FunctionTransformer(fe...\n",
              "                                                  OneHotEncoder(handle_unknown=&#x27;ignore&#x27;))]),\n",
              "                                 [&#x27;job&#x27;, &#x27;housing&#x27;, &#x27;other_payment_plans&#x27;,\n",
              "                                  &#x27;property_magnitude&#x27;, &#x27;other_parties&#x27;,\n",
              "                                  &#x27;personal_status&#x27;, &#x27;credit_history&#x27;]),\n",
              "                                (&#x27;geo&#x27;,\n",
              "                                 Pipeline(steps=[(&#x27;ordered_ordinal_encoder&#x27;,\n",
              "                                                  OrderedOrdinalEncoder()),\n",
              "                                                 (&#x27;cluster_simil&#x27;,\n",
              "                                                  ClusterSimilarity(random_state=42))]),\n",
              "                                 [&#x27;savings_status&#x27;, &#x27;checking_status&#x27;]),\n",
              "                                (&#x27;drop_out&#x27;, &#x27;drop&#x27;,\n",
              "                                 [&#x27;num_dependents&#x27;, &#x27;foreign_worker&#x27;,\n",
              "                                  &#x27;purpose&#x27;])])</pre></div> </div></div><div class=\"sk-parallel\"><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-496\" type=\"checkbox\" ><label for=\"sk-estimator-id-496\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">monthly_credit_burden</label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;credit_amount&#x27;, &#x27;duration&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-497\" type=\"checkbox\" ><label for=\"sk-estimator-id-497\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;FunctionTransformer<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.FunctionTransformer.html\">?<span>Documentation for FunctionTransformer</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>FunctionTransformer(feature_names_out=&lt;function ratio_name at 0x78a662f00dc0&gt;,\n",
              "                    func=&lt;function ratio_pipeline.&lt;locals&gt;.&lt;lambda&gt; at 0x78a6624f3250&gt;)</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-498\" type=\"checkbox\" ><label for=\"sk-estimator-id-498\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;FunctionTransformer<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.FunctionTransformer.html\">?<span>Documentation for FunctionTransformer</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>FunctionTransformer(feature_names_out=&#x27;one-to-one&#x27;, func=&lt;ufunc &#x27;log&#x27;&gt;,\n",
              "                    inverse_func=&lt;ufunc &#x27;exp&#x27;&gt;)</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-499\" type=\"checkbox\" ><label for=\"sk-estimator-id-499\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;StandardScaler<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.StandardScaler.html\">?<span>Documentation for StandardScaler</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>StandardScaler()</pre></div> </div></div></div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-500\" type=\"checkbox\" ><label for=\"sk-estimator-id-500\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">installment_per_credit</label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;installment_commitment&#x27;, &#x27;credit_amount&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-501\" type=\"checkbox\" ><label for=\"sk-estimator-id-501\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;FunctionTransformer<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.FunctionTransformer.html\">?<span>Documentation for FunctionTransformer</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>FunctionTransformer(feature_names_out=&lt;function ratio_name at 0x78a662f00dc0&gt;,\n",
              "                    func=&lt;function ratio_pipeline.&lt;locals&gt;.&lt;lambda&gt; at 0x78a6624f3400&gt;)</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-502\" type=\"checkbox\" ><label for=\"sk-estimator-id-502\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;FunctionTransformer<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.FunctionTransformer.html\">?<span>Documentation for FunctionTransformer</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>FunctionTransformer(feature_names_out=&#x27;one-to-one&#x27;, func=&lt;ufunc &#x27;log&#x27;&gt;,\n",
              "                    inverse_func=&lt;ufunc &#x27;exp&#x27;&gt;)</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-503\" type=\"checkbox\" ><label for=\"sk-estimator-id-503\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;StandardScaler<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.StandardScaler.html\">?<span>Documentation for StandardScaler</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>StandardScaler()</pre></div> </div></div></div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-504\" type=\"checkbox\" ><label for=\"sk-estimator-id-504\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">dependents_per_credit</label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;num_dependents&#x27;, &#x27;credit_amount&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-505\" type=\"checkbox\" ><label for=\"sk-estimator-id-505\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;FunctionTransformer<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.FunctionTransformer.html\">?<span>Documentation for FunctionTransformer</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>FunctionTransformer(feature_names_out=&lt;function ratio_name at 0x78a662f00dc0&gt;,\n",
              "                    func=&lt;function ratio_pipeline.&lt;locals&gt;.&lt;lambda&gt; at 0x78a6624f35b0&gt;)</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-506\" type=\"checkbox\" ><label for=\"sk-estimator-id-506\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;FunctionTransformer<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.FunctionTransformer.html\">?<span>Documentation for FunctionTransformer</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>FunctionTransformer(feature_names_out=&#x27;one-to-one&#x27;, func=&lt;ufunc &#x27;log&#x27;&gt;,\n",
              "                    inverse_func=&lt;ufunc &#x27;exp&#x27;&gt;)</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-507\" type=\"checkbox\" ><label for=\"sk-estimator-id-507\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;StandardScaler<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.StandardScaler.html\">?<span>Documentation for StandardScaler</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>StandardScaler()</pre></div> </div></div></div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-508\" type=\"checkbox\" ><label for=\"sk-estimator-id-508\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">log</label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;duration&#x27;, &#x27;credit_amount&#x27;, &#x27;installment_commitment&#x27;, &#x27;residence_since&#x27;, &#x27;age&#x27;, &#x27;existing_credits&#x27;, &#x27;num_dependents&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-509\" type=\"checkbox\" ><label for=\"sk-estimator-id-509\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;FunctionTransformer<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.FunctionTransformer.html\">?<span>Documentation for FunctionTransformer</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>FunctionTransformer(feature_names_out=&#x27;one-to-one&#x27;, func=&lt;ufunc &#x27;log&#x27;&gt;,\n",
              "                    inverse_func=&lt;ufunc &#x27;exp&#x27;&gt;)</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-510\" type=\"checkbox\" ><label for=\"sk-estimator-id-510\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;StandardScaler<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.StandardScaler.html\">?<span>Documentation for StandardScaler</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>StandardScaler()</pre></div> </div></div></div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-511\" type=\"checkbox\" ><label for=\"sk-estimator-id-511\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">cat_ordinal</label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;checking_status&#x27;, &#x27;savings_status&#x27;, &#x27;employment&#x27;, &#x27;own_telephone&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-512\" type=\"checkbox\" ><label for=\"sk-estimator-id-512\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">OrderedOrdinalEncoder</label><div class=\"sk-toggleable__content fitted\"><pre>OrderedOrdinalEncoder()</pre></div> </div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-513\" type=\"checkbox\" ><label for=\"sk-estimator-id-513\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">cat_onehot</label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;job&#x27;, &#x27;housing&#x27;, &#x27;other_payment_plans&#x27;, &#x27;property_magnitude&#x27;, &#x27;other_parties&#x27;, &#x27;personal_status&#x27;, &#x27;credit_history&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-514\" type=\"checkbox\" ><label for=\"sk-estimator-id-514\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;OneHotEncoder<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.OneHotEncoder.html\">?<span>Documentation for OneHotEncoder</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>OneHotEncoder(handle_unknown=&#x27;ignore&#x27;)</pre></div> </div></div></div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-515\" type=\"checkbox\" ><label for=\"sk-estimator-id-515\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">geo</label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;savings_status&#x27;, &#x27;checking_status&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-516\" type=\"checkbox\" ><label for=\"sk-estimator-id-516\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">OrderedOrdinalEncoder</label><div class=\"sk-toggleable__content fitted\"><pre>OrderedOrdinalEncoder()</pre></div> </div></div><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-517\" type=\"checkbox\" ><label for=\"sk-estimator-id-517\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">ClusterSimilarity</label><div class=\"sk-toggleable__content fitted\"><pre>ClusterSimilarity(random_state=42)</pre></div> </div></div></div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-518\" type=\"checkbox\" ><label for=\"sk-estimator-id-518\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">drop_out</label><div class=\"sk-toggleable__content fitted\"><pre>[&#x27;num_dependents&#x27;, &#x27;foreign_worker&#x27;, &#x27;purpose&#x27;]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-519\" type=\"checkbox\" ><label for=\"sk-estimator-id-519\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">drop</label><div class=\"sk-toggleable__content fitted\"><pre>drop</pre></div> </div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-520\" type=\"checkbox\" ><label for=\"sk-estimator-id-520\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">remainder</label><div class=\"sk-toggleable__content fitted\"><pre>[]</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-521\" type=\"checkbox\" ><label for=\"sk-estimator-id-521\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;StandardScaler<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.preprocessing.StandardScaler.html\">?<span>Documentation for StandardScaler</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>StandardScaler()</pre></div> </div></div></div></div></div></div></div></div></div><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-522\" type=\"checkbox\" ><label for=\"sk-estimator-id-522\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;stacking: StackingClassifier<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.ensemble.StackingClassifier.html\">?<span>Documentation for stacking: StackingClassifier</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>StackingClassifier(estimators=[(&#x27;random_forest&#x27;,\n",
              "                                RandomForestClassifier(class_weight=&#x27;balanced&#x27;,\n",
              "                                                       random_state=42)),\n",
              "                               (&#x27;xgboost&#x27;,\n",
              "                                XGBClassifier(base_score=None, booster=None,\n",
              "                                              callbacks=None,\n",
              "                                              colsample_bylevel=None,\n",
              "                                              colsample_bynode=None,\n",
              "                                              colsample_bytree=None,\n",
              "                                              device=None,\n",
              "                                              early_stopping_rounds=None,\n",
              "                                              enable_categorical=False,\n",
              "                                              eval_metric=&#x27;logloss&#x27;,\n",
              "                                              feature_types=N...\n",
              "                                              interaction_constraints=None,\n",
              "                                              learning_rate=0.1, max_bin=None,\n",
              "                                              max_cat_threshold=None,\n",
              "                                              max_cat_to_onehot=None,\n",
              "                                              max_delta_step=None, max_depth=5,\n",
              "                                              max_leaves=None,\n",
              "                                              min_child_weight=None,\n",
              "                                              missing=nan,\n",
              "                                              monotone_constraints=None,\n",
              "                                              multi_strategy=None,\n",
              "                                              n_estimators=100, n_jobs=None,\n",
              "                                              num_parallel_tree=None,\n",
              "                                              random_state=42, ...))],\n",
              "                   final_estimator=LogisticRegression(random_state=42))</pre></div> </div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-parallel\"><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><label>random_forest</label></div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-523\" type=\"checkbox\" ><label for=\"sk-estimator-id-523\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;RandomForestClassifier<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.ensemble.RandomForestClassifier.html\">?<span>Documentation for RandomForestClassifier</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>RandomForestClassifier(class_weight=&#x27;balanced&#x27;, random_state=42)</pre></div> </div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><label>xgboost</label></div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-524\" type=\"checkbox\" ><label for=\"sk-estimator-id-524\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">XGBClassifier</label><div class=\"sk-toggleable__content fitted\"><pre>XGBClassifier(base_score=None, booster=None, callbacks=None,\n",
              "              colsample_bylevel=None, colsample_bynode=None,\n",
              "              colsample_bytree=None, device=None, early_stopping_rounds=None,\n",
              "              enable_categorical=False, eval_metric=&#x27;logloss&#x27;,\n",
              "              feature_types=None, gamma=None, grow_policy=None,\n",
              "              importance_type=None, interaction_constraints=None,\n",
              "              learning_rate=0.1, max_bin=None, max_cat_threshold=None,\n",
              "              max_cat_to_onehot=None, max_delta_step=None, max_depth=5,\n",
              "              max_leaves=None, min_child_weight=None, missing=nan,\n",
              "              monotone_constraints=None, multi_strategy=None, n_estimators=100,\n",
              "              n_jobs=None, num_parallel_tree=None, random_state=42, ...)</pre></div> </div></div></div></div></div></div></div><div class=\"sk-item\"><div class=\"sk-parallel\"><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label fitted sk-toggleable\"><label>final_estimator</label></div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-525\" type=\"checkbox\" ><label for=\"sk-estimator-id-525\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow fitted\">&nbsp;LogisticRegression<a class=\"sk-estimator-doc-link fitted\" rel=\"noreferrer\" target=\"_blank\" href=\"https://scikit-learn.org/1.5/modules/generated/sklearn.linear_model.LogisticRegression.html\">?<span>Documentation for LogisticRegression</span></a></label><div class=\"sk-toggleable__content fitted\"><pre>LogisticRegression(random_state=42)</pre></div> </div></div></div></div></div></div></div></div></div></div></div></div></div>"
            ]
          },
          "metadata": {},
          "execution_count": 97
        }
      ],
      "source": [
        "from sklearn.ensemble import RandomForestClassifier, StackingClassifier\n",
        "from sklearn.model_selection import train_test_split, cross_val_score\n",
        "from sklearn.metrics import accuracy_score, precision_score, recall_score\n",
        "from xgboost import XGBClassifier\n",
        "from sklearn.linear_model import LogisticRegression\n",
        "\n",
        "# Random Forest model\n",
        "random_forest_em = RandomForestClassifier(\n",
        "    n_estimators=100,\n",
        "    random_state=42,\n",
        "    class_weight='balanced'\n",
        ")\n",
        "\n",
        "# XGBoost model\n",
        "xgboost = XGBClassifier(\n",
        "    n_estimators=100,\n",
        "    max_depth=5,\n",
        "    learning_rate=0.1,\n",
        "    random_state=42,\n",
        "    eval_metric='logloss'\n",
        ")\n",
        "\n",
        "# Meta-learner (Logistic Regression to combine the outputs)\n",
        "meta_learner = LogisticRegression(random_state=42)\n",
        "\n",
        "# Create a stacking ensemble that uses Random Forest and XGBoost as base models\n",
        "stacking_classifier = StackingClassifier(\n",
        "    estimators=[\n",
        "        ('random_forest', random_forest_em),\n",
        "        ('xgboost', xgboost)\n",
        "    ],\n",
        "    final_estimator=meta_learner,\n",
        "    passthrough=False  # Set to True if you want to pass original features alongside model predictions\n",
        ")\n",
        "\n",
        "# Full pipeline including preprocessing and stacking classifier\n",
        "ensemble_model = Pipeline(steps=[\n",
        "    ('preprocessing', preprocessing),  # Your custom preprocessing step goes here\n",
        "    ('stacking', stacking_classifier)  # Stacking ensemble with RF and XGBoost\n",
        "])\n",
        "\n",
        "# Train the model\n",
        "ensemble_model.fit(X_train, y_train)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "L9Q4nOgI0CYF"
      },
      "source": [
        "Testing the accuracy of the model:"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 100,
      "metadata": {
        "id": "hQF1smYt0CYG",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "a746a75a-7ec7-4384-bc81-90a72be786cc"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Training Set Performance:\n",
            "Accuracy: 0.9988\n",
            "Precision: 0.9982\n",
            "Recall: 1.0000\n",
            "F1-Score: 0.9991\n",
            "ROC-AUC: 1.0000\n",
            "------------------------------\n"
          ]
        }
      ],
      "source": [
        "from sklearn.metrics import accuracy_score, precision_score, recall_score, f1_score, roc_auc_score\n",
        "\n",
        "# Make predictions\n",
        "train_predictions = ensemble_model.predict(X_train)\n",
        "\n",
        "# Evaluation metrics\n",
        "train_accuracy = accuracy_score(y_train, train_predictions)\n",
        "train_precision = precision_score(y_train, train_predictions, pos_label='good')\n",
        "train_recall = recall_score(y_train, train_predictions, pos_label='good')\n",
        "train_f1 = f1_score(y_train, train_predictions, pos_label='good')\n",
        "train_roc_auc = roc_auc_score(y_train, ensemble_model.predict_proba(X_train)[:, 1])\n",
        "\n",
        "# Print training set results\n",
        "print(\"Training Set Performance:\")\n",
        "print(f\"Accuracy: {train_accuracy:.4f}\")\n",
        "print(f\"Precision: {train_precision:.4f}\")\n",
        "print(f\"Recall: {train_recall:.4f}\")\n",
        "print(f\"F1-Score: {train_f1:.4f}\")\n",
        "print(f\"ROC-AUC: {train_roc_auc:.4f}\")\n",
        "print(\"-\" * 30)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "bfuZJ8k-mFjZ"
      },
      "source": [
        "# Model Evaluation"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ySehtiQHbkqi"
      },
      "source": [
        "After training our models with Logistic Regression, Random Forest, Gradient Boosting, and SVM, we need to evaluate their performance on the testing data. We utiliz metrics such as accuracy, precision, recall, F1-score, and ROC-AUC to determine the model's effectiveness in predicting credit risk. Here's the code to evaluate the models:"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 105,
      "metadata": {
        "collapsed": true,
        "id": "S1MIBU06mInU",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "ba48d1d3-a160-41ea-a612-f3e715e1e0cd"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Logistic Regression Evaluation on Training Set:\n",
            "Accuracy: 0.7425\n",
            "Precision: 0.8766\n",
            "Recall: 0.7357\n",
            "F1-Score: 0.8000\n",
            "ROC-AUC: 0.8123\n",
            "\n",
            "\n",
            "Random Forest Evaluation on Training Set:\n",
            "Accuracy: 1.0000\n",
            "Precision: 1.0000\n",
            "Recall: 1.0000\n",
            "F1-Score: 1.0000\n",
            "ROC-AUC: 1.0000\n",
            "\n",
            "\n",
            "Gradient Boosting Evaluation on Training Set:\n",
            "Accuracy: 0.9250\n",
            "Precision: 0.9153\n",
            "Recall: 0.9839\n",
            "F1-Score: 0.9484\n",
            "ROC-AUC: 0.9743\n",
            "\n",
            "\n",
            "SVM Evaluation on Training Set:\n",
            "Accuracy: 0.7913\n",
            "Precision: 0.9119\n",
            "Recall: 0.7768\n",
            "F1-Score: 0.8390\n",
            "ROC-AUC: 0.8964\n",
            "\n",
            "\n",
            "SGD Evaluation on Training Set:\n",
            "Accuracy: 0.6850\n",
            "Precision: 0.8548\n",
            "Recall: 0.6625\n",
            "F1-Score: 0.7465\n",
            "ROC-AUC: 0.7771\n",
            "\n",
            "\n",
            "Ensemble Method Evaluation on Training Set:\n",
            "Accuracy: 0.9988\n",
            "Precision: 0.9982\n",
            "Recall: 1.0000\n",
            "F1-Score: 0.9991\n",
            "ROC-AUC: 1.0000\n",
            "\n",
            "\n"
          ]
        }
      ],
      "source": [
        "from sklearn.metrics import accuracy_score, precision_score, recall_score, f1_score, roc_auc_score\n",
        "\n",
        "# Function to evaluate the model's performance on the training set\n",
        "def evaluate_model_on_train(model, X_train, y_train):\n",
        "    # Make predictions on the training set\n",
        "    y_pred = model.predict(X_train)\n",
        "\n",
        "    # Calculate metrics using y_train for true labels\n",
        "    accuracy = accuracy_score(y_train, y_pred)\n",
        "    precision = precision_score(y_train, y_pred, pos_label=\"good\")\n",
        "    recall = recall_score(y_train, y_pred, pos_label=\"good\")\n",
        "    f1 = f1_score(y_train, y_pred, pos_label=\"good\")\n",
        "    roc_auc = roc_auc_score(y_train, model.predict_proba(X_train)[:, 1])\n",
        "\n",
        "    # Display metrics\n",
        "    print(f\"Accuracy: {accuracy:.4f}\")\n",
        "    print(f\"Precision: {precision:.4f}\")\n",
        "    print(f\"Recall: {recall:.4f}\")\n",
        "    print(f\"F1-Score: {f1:.4f}\")\n",
        "    print(f\"ROC-AUC: {roc_auc:.4f}\")\n",
        "    print(\"\\n\")\n",
        "\n",
        "# Logistic Regression Model\n",
        "print(\"Logistic Regression Evaluation on Training Set:\")\n",
        "evaluate_model_on_train(log_reg, X_train, y_train)\n",
        "\n",
        "# Random Forest Model\n",
        "print(\"Random Forest Evaluation on Training Set:\")\n",
        "evaluate_model_on_train(random_forest, X_train, y_train)\n",
        "\n",
        "# Gradient Boosting Model\n",
        "print(\"Gradient Boosting Evaluation on Training Set:\")\n",
        "evaluate_model_on_train(grad_boost, X_train, y_train)\n",
        "\n",
        "# SVM Model\n",
        "print(\"SVM Evaluation on Training Set:\")\n",
        "evaluate_model_on_train(svm_model, X_train, y_train)\n",
        "\n",
        "# Stochastic Gradient Descent\n",
        "print(\"SGD Evaluation on Training Set:\")\n",
        "evaluate_model_on_train(sgd_model, X_train, y_train)\n",
        "\n",
        "# Ensemble Method\n",
        "print(\"Ensemble Method Evaluation on Training Set:\")\n",
        "evaluate_model_on_train(ensemble_model, X_train, y_train)\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "BFYFsbgYd2oT"
      },
      "source": [
        "Let's verify our results by checking the confusion matrices. The following code will use cross validation to predict some values and then compute a confusion matrix. This will help us better understand the errors each model is making."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 109,
      "metadata": {
        "id": "JUWkVWa2AO7O",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "88b92b1b-65d6-4ad1-9f12-6cf6e92a2a5c"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Logistic Regression Evaluation (Training Set):\n",
            "[[155  85]\n",
            " [175 385]]\n",
            "Accuracy: 0.675\n",
            "Precision: 0.819\n",
            "Recall: 0.688\n",
            "\n",
            "Random Forest Evaluation (Training Set):\n",
            "[[111 129]\n",
            " [ 61 499]]\n",
            "Accuracy: 0.762\n",
            "Precision: 0.795\n",
            "Recall: 0.891\n",
            "\n",
            "Gradient Boosting Evaluation (Training Set):\n",
            "[[110 130]\n",
            " [ 68 492]]\n",
            "Accuracy: 0.752\n",
            "Precision: 0.791\n",
            "Recall: 0.879\n",
            "\n",
            "SVM Evaluation (Training Set):\n",
            "[[166  74]\n",
            " [167 393]]\n",
            "Accuracy: 0.699\n",
            "Precision: 0.842\n",
            "Recall: 0.702\n",
            "\n",
            "SGD Evaluation (Training Set):\n",
            "[[174  66]\n",
            " [268 292]]\n",
            "Accuracy: 0.583\n",
            "Precision: 0.816\n",
            "Recall: 0.521\n",
            "\n",
            "Ensemble Method Evaluation (Training Set):\n",
            "[[104 136]\n",
            " [ 49 511]]\n",
            "Accuracy: 0.769\n",
            "Precision: 0.790\n",
            "Recall: 0.912\n",
            "\n"
          ]
        }
      ],
      "source": [
        "from sklearn.metrics import confusion_matrix\n",
        "from sklearn.model_selection import cross_val_predict\n",
        "\n",
        "# Function to calculate and return the confusion matrix\n",
        "def getConfusionMatrix(model, X_train, y_train):\n",
        "    y_pred = cross_val_predict(model, X_train, y_train, cv=3)\n",
        "    return confusion_matrix(y_train, y_pred)\n",
        "\n",
        "# Function to calculate accuracy, precision, and recall from confusion matrix\n",
        "def calcStatsFromCM(conf_matrix):\n",
        "    accuracy = (conf_matrix[0][0] + conf_matrix[1][1]) / (conf_matrix.sum())\n",
        "    precision = conf_matrix[1][1] / (conf_matrix[0][1] + conf_matrix[1][1])\n",
        "    recall = conf_matrix[1][1] / (conf_matrix[1][0] + conf_matrix[1][1])\n",
        "\n",
        "    print(f\"Accuracy: {accuracy:.3f}\")\n",
        "    print(f\"Precision: {precision:.3f}\")\n",
        "    print(f\"Recall: {recall:.3f}\")\n",
        "    # Add F1 score if required\n",
        "\n",
        "# Evaluate Logistic Regression\n",
        "print(\"Logistic Regression Evaluation (Training Set):\")\n",
        "cf_toprint = getConfusionMatrix(log_reg, X_train, y_train)\n",
        "print(cf_toprint)\n",
        "calcStatsFromCM(cf_toprint)\n",
        "print()\n",
        "\n",
        "# Evaluate Random Forest\n",
        "print(\"Random Forest Evaluation (Training Set):\")\n",
        "cf_toprint = getConfusionMatrix(random_forest, X_train, y_train)\n",
        "print(cf_toprint)\n",
        "calcStatsFromCM(cf_toprint)\n",
        "print()\n",
        "\n",
        "# Evaluate Gradient Boosting\n",
        "print(\"Gradient Boosting Evaluation (Training Set):\")\n",
        "cf_toprint = getConfusionMatrix(grad_boost, X_train, y_train)\n",
        "print(cf_toprint)\n",
        "calcStatsFromCM(cf_toprint)\n",
        "print()\n",
        "\n",
        "# Evaluate SVM\n",
        "print(\"SVM Evaluation (Training Set):\")\n",
        "cf_toprint = getConfusionMatrix(svm_model, X_train, y_train)\n",
        "print(cf_toprint)\n",
        "calcStatsFromCM(cf_toprint)\n",
        "print()\n",
        "\n",
        "# Evaluate SGD\n",
        "print(\"SGD Evaluation (Training Set):\")\n",
        "cf_toprint = getConfusionMatrix(sgd_model, X_train, y_train)\n",
        "print(cf_toprint)\n",
        "calcStatsFromCM(cf_toprint)\n",
        "print()\n",
        "\n",
        "# Evaluate Ensemble Method\n",
        "print(\"Ensemble Method Evaluation (Training Set):\")\n",
        "cf_toprint = getConfusionMatrix(ensemble_model, X_train, y_train)\n",
        "print(cf_toprint)\n",
        "calcStatsFromCM(cf_toprint)\n",
        "print()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "yuZTkgZhoEp1"
      },
      "source": [
        "Here we are seeing that the accuracy of our model has significantly dropped from our previous evaluation this is beacuse our previous model was most likely overfitted with the training data. The current evaluation is more robust using cross validation which avoids bias from overfitting. Still, in both cases the RandomForest model seems to perform the best. Let's look at the feature importances to understand how it works."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 110,
      "metadata": {
        "id": "ksKq3VqGOwoo",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "94ba0a2e-bfa9-457d-894a-a25cb875afd4"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([0.07266337, 0.06580943, 0.06830424, 0.05986567, 0.07055637,\n",
              "       0.02145402, 0.02436791, 0.06165939, 0.0123794 , 0.00863905,\n",
              "       0.03460524, 0.00794557, 0.0268563 , 0.01163128, 0.00798102,\n",
              "       0.01029658, 0.0013283 , 0.0089704 , 0.00517736, 0.00961986,\n",
              "       0.0071636 , 0.01309946, 0.01871864, 0.00426377, 0.00949016,\n",
              "       0.01275909, 0.00726657, 0.01267891, 0.00412174, 0.01136062,\n",
              "       0.01011154, 0.01122694, 0.00528049, 0.00566991, 0.01205744,\n",
              "       0.0034243 , 0.01677702, 0.0065119 , 0.01139894, 0.0044158 ,\n",
              "       0.02771799, 0.01567494, 0.01726939, 0.02186727, 0.03416902,\n",
              "       0.01592579, 0.03299702, 0.01013927, 0.01679317, 0.02953851])"
            ]
          },
          "metadata": {},
          "execution_count": 110
        }
      ],
      "source": [
        "random_forest.named_steps[\"randomforestclassifier\"].feature_importances_"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "K1j40ly4oXam"
      },
      "source": [
        "These numbers are not very useful on their own, so let's make a function to assign names to them and place them in a dataframe."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 111,
      "metadata": {
        "id": "XwcfuLyTP3VX",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "f07f1fbd-3b8d-4733-eac3-cc1563c36b90"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "                                              feature  importance\n",
            "0                        monthly_credit_burden__ratio    0.072663\n",
            "1                                  log__credit_amount    0.070556\n",
            "2                        dependents_per_credit__ratio    0.068304\n",
            "3                       installment_per_credit__ratio    0.065809\n",
            "4                                            log__age    0.061659\n",
            "5                                       log__duration    0.059866\n",
            "6                        cat_ordinal__checking_status    0.034605\n",
            "7                           geo__Cluster 4 similarity    0.034169\n",
            "8                           geo__Cluster 6 similarity    0.032997\n",
            "9                           geo__Cluster 9 similarity    0.029539\n",
            "10                          geo__Cluster 0 similarity    0.027718\n",
            "11                            cat_ordinal__employment    0.026856\n",
            "12                               log__residence_since    0.024368\n",
            "13                          geo__Cluster 3 similarity    0.021867\n",
            "14                        log__installment_commitment    0.021454\n",
            "15               cat_onehot__other_payment_plans_none    0.018719\n",
            "16                          geo__Cluster 2 similarity    0.017269\n",
            "17                          geo__Cluster 8 similarity    0.016793\n",
            "18  cat_onehot__credit_history_critical/other exis...    0.016777\n",
            "19                          geo__Cluster 5 similarity    0.015926\n",
            "20                          geo__Cluster 1 similarity    0.015675\n",
            "21               cat_onehot__other_payment_plans_bank    0.013099\n",
            "22      cat_onehot__property_magnitude_life insurance    0.012759\n",
            "23         cat_onehot__property_magnitude_real estate    0.012679\n",
            "24                              log__existing_credits    0.012379\n",
            "25            cat_onehot__personal_status_male single    0.012057\n",
            "26                         cat_ordinal__own_telephone    0.011631\n",
            "27           cat_onehot__credit_history_existing paid    0.011399\n",
            "28                cat_onehot__other_parties_guarantor    0.011361\n",
            "29     cat_onehot__personal_status_female div/dep/mar    0.011227\n",
            "30                            cat_onehot__job_skilled    0.010297\n",
            "31                          geo__Cluster 7 similarity    0.010139\n",
            "32                     cat_onehot__other_parties_none    0.010112\n",
            "33                            cat_onehot__housing_own    0.009620\n",
            "34                 cat_onehot__property_magnitude_car    0.009490\n",
            "35                 cat_onehot__job_unskilled resident    0.008970\n",
            "36                                log__num_dependents    0.008639\n",
            "37          cat_onehot__job_high qualif/self emp/mgmt    0.007981\n",
            "38                        cat_ordinal__savings_status    0.007946\n",
            "39   cat_onehot__property_magnitude_no known property    0.007267\n",
            "40                           cat_onehot__housing_rent    0.007164\n",
            "41      cat_onehot__credit_history_delayed previously    0.006512\n",
            "42           cat_onehot__personal_status_male mar/wid    0.005670\n",
            "43           cat_onehot__personal_status_male div/sep    0.005280\n",
            "44                       cat_onehot__housing_for free    0.005177\n",
            "45     cat_onehot__credit_history_no credits/all paid    0.004416\n",
            "46             cat_onehot__other_payment_plans_stores    0.004264\n",
            "47             cat_onehot__other_parties_co applicant    0.004122\n",
            "48                cat_onehot__credit_history_all paid    0.003424\n",
            "49            cat_onehot__job_unemp/unskilled non res    0.001328\n"
          ]
        }
      ],
      "source": [
        "# Extract feature names after preprocessing\n",
        "feature_names = preprocessing.get_feature_names_out ()\n",
        "\n",
        "# Get feature improtances from the RandomForest model\n",
        "rf_model = random_forest.named_steps['randomforestclassifier']\n",
        "importances = rf_model.feature_importances_\n",
        "\n",
        "# Pair feature names with importance values\n",
        "feature_importance = pd.DataFrame ({\n",
        "    'feature': feature_names,\n",
        "    'importance': importances\n",
        "})\n",
        "\n",
        "# Sort by importance in descending order\n",
        "feature_importance = feature_importance.sort_values ('importance', ascending=False).reset_index (drop=True)\n",
        "\n",
        "# Display results\n",
        "print(feature_importance.head (60))"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "pFH1bcZTot59"
      },
      "source": [
        "It looks like the ratio features are extremely useful. You may have noticed earlier that we dropped some columns with our ColumnTransformer. This was based on previous iterations of the model that told us those features were not helpful to the model. We removed them to reduce its complexity and allow it to focus on more important features."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ZoAh3qyCmJCE"
      },
      "source": [
        "# Hyperparameter Tuning"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "4aPlwRq4iYQH"
      },
      "source": [
        "## Grid Search"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "BJzP7gg5Al8s"
      },
      "source": [
        "Below we are first going to be creating a pipeline that includes preprocessing steps and a Random Forest model. We want to perform hyperparameter tuning using GridSearchCV.\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 112,
      "metadata": {
        "id": "FSDLYRX5AbMK",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "1b8b9437-0cd7-4177-fc2b-ce95dc942f6e"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Best Hyperparameters: {'random_forest__max_depth': None, 'random_forest__min_samples_leaf': 2, 'random_forest__min_samples_split': 2, 'random_forest__n_estimators': 100}\n"
          ]
        }
      ],
      "source": [
        "from sklearn.model_selection import GridSearchCV\n",
        "from sklearn.ensemble import RandomForestClassifier\n",
        "from sklearn.pipeline import Pipeline\n",
        "\n",
        "# Define the pipeline\n",
        "full_pipeline = Pipeline([\n",
        "    (\"preprocessing\", preprocessing),\n",
        "    (\"random_forest\", RandomForestClassifier(random_state=42)),\n",
        "])\n",
        "\n",
        "# Define the parameter grid with the correct prefix 'random_forest__'\n",
        "param_grid = {\n",
        "    'random_forest__n_estimators': [100, 200, 300],\n",
        "    'random_forest__max_depth': [None, 10, 20, 30],\n",
        "    'random_forest__min_samples_split': [2, 5, 10],\n",
        "    'random_forest__min_samples_leaf': [1, 2, 4],\n",
        "}\n",
        "\n",
        "# Perform GridSearch\n",
        "grid_search = GridSearchCV(full_pipeline, param_grid, cv=5, scoring='accuracy', n_jobs=-1)\n",
        "grid_search.fit(X_train, y_train)\n",
        "\n",
        "# Best hyperparameters\n",
        "print(\"Best Hyperparameters:\", grid_search.best_params_)\n",
        "\n",
        "# Re-fit the model with the best parameters\n",
        "best_rf_model = grid_search.best_estimator_"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "sATBFeKBCH0u"
      },
      "source": [
        "We can now re-evaluate the tuned Random Forest model on the test set using accuracy, precision, recall, F1-score, and ROC-AUC."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 114,
      "metadata": {
        "id": "WPd-1eVvCL5p",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "4d8d738a-25bd-48f4-afa0-8c571d36f49b"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Training Set Evaluation:\n",
            "Accuracy: 0.9663\n",
            "Precision: 0.9540\n",
            "Recall: 1.0000\n",
            "F1-Score: 0.9765\n",
            "ROC-AUC: 0.9987\n"
          ]
        }
      ],
      "source": [
        "from sklearn.metrics import accuracy_score, precision_score, recall_score, f1_score, roc_auc_score\n",
        "\n",
        "# Make predictions on the training set\n",
        "y_train_pred = best_rf_model.predict(X_train)\n",
        "\n",
        "# Calculate evaluation metrics\n",
        "train_accuracy = accuracy_score(y_train, y_train_pred)\n",
        "train_precision = precision_score(y_train, y_train_pred, pos_label=\"good\")\n",
        "train_recall = recall_score(y_train, y_train_pred, pos_label=\"good\")\n",
        "train_f1 = f1_score(y_train, y_train_pred, pos_label=\"good\")\n",
        "train_roc_auc = roc_auc_score(y_train, best_rf_model.predict_proba(X_train)[:, 1])\n",
        "\n",
        "# Print results\n",
        "print(f\"Training Set Evaluation:\")\n",
        "print(f\"Accuracy: {train_accuracy:.4f}\")\n",
        "print(f\"Precision: {train_precision:.4f}\")\n",
        "print(f\"Recall: {train_recall:.4f}\")\n",
        "print(f\"F1-Score: {train_f1:.4f}\")\n",
        "print(f\"ROC-AUC: {train_roc_auc:.4f}\")\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 116,
      "metadata": {
        "id": "lUMemcLyMSf4",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 395
        },
        "outputId": "4b15224e-7d0e-4034-bb06-2699a3f0cfe1"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 600x400 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ],
      "source": [
        "# Importing necessary libraries for confusion matrix\n",
        "from sklearn.metrics import confusion_matrix\n",
        "import matplotlib.pyplot as plt\n",
        "import seaborn as sns\n",
        "\n",
        "# Predicting using the best_rf_model on the training set\n",
        "y_train_pred = best_rf_model.predict(X_train)\n",
        "\n",
        "# Creating the confusion matrix using the training set\n",
        "cm = confusion_matrix(y_train, y_train_pred)\n",
        "\n",
        "# Plotting the confusion matrix\n",
        "plt.figure(figsize=(6, 4))\n",
        "sns.heatmap(cm, annot=True, fmt='d', cmap='Blues', cbar=False)\n",
        "plt.title('Confusion Matrix for Random Forest (Training Set)')\n",
        "plt.xlabel('Predicted Labels')\n",
        "plt.ylabel('True Labels')\n",
        "plt.show()\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "TDJHTVzqElXx"
      },
      "source": [
        "We are now able to see a slight uptick in performance after hyperparameter tuning with GridSearchCV compared to the original untuned Random Forest model."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "DpvuRbB0CWrv"
      },
      "source": [
        "We can better visualize the results of the grid search with some plots below:"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import matplotlib.pyplot as plt\n",
        "import seaborn as sns\n",
        "import pandas as pd\n",
        "\n",
        "# Extract the results from GridSearchCV\n",
        "grid_results = pd.DataFrame(grid_search.cv_results_)\n",
        "\n",
        "# Pivot the data to create a matrix format suitable for heatmap\n",
        "heatmap_data = grid_results.pivot_table(index='param_random_forest__max_depth',\n",
        "                                        columns='param_random_forest__n_estimators',\n",
        "                                        values='mean_test_score')\n",
        "\n",
        "# Plotting the heatmap\n",
        "plt.figure(figsize=(10, 6))\n",
        "sns.heatmap(heatmap_data, annot=True, cmap=\"YlGnBu\", fmt=\".4f\")\n",
        "plt.title(\"GridSearchCV - Random Forest Mean Test Scores\")\n",
        "plt.xlabel('Number of Estimators')\n",
        "plt.ylabel('Max Depth')\n",
        "plt.show()\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 549
        },
        "id": "mDaNHRsnzfKo",
        "outputId": "4941a6a5-f618-4907-c16d-72a00ee2c567"
      },
      "execution_count": 125,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1000x600 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "do67meDRAS6g"
      },
      "source": [
        "## Randomized Search"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "G8mxTz8UDRr4"
      },
      "source": [
        "Below we are first going to be creating a pipeline that includes preprocessing steps and a Random Forest model. We want to perform hyperparameter tuning using RandomizedSearchCV."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 118,
      "metadata": {
        "id": "D4PR1_voC1l4",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "171ed334-ba02-4660-fac1-aee326833ae5"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Fitting 5 folds for each of 100 candidates, totalling 500 fits\n",
            "Best Hyperparameters from RandomizedSearchCV: {'random_forest__n_estimators': 200, 'random_forest__min_samples_split': 10, 'random_forest__min_samples_leaf': 4, 'random_forest__max_depth': 50, 'random_forest__bootstrap': False}\n"
          ]
        }
      ],
      "source": [
        "from sklearn.model_selection import RandomizedSearchCV\n",
        "from sklearn.ensemble import RandomForestClassifier\n",
        "import numpy as np\n",
        "from sklearn.pipeline import Pipeline\n",
        "\n",
        "# Define the pipeline\n",
        "full_pipeline = Pipeline([\n",
        "    (\"preprocessing\", preprocessing),\n",
        "    (\"random_forest\", RandomForestClassifier(random_state=42)),\n",
        "])\n",
        "\n",
        "# Define the hyperparameter grid with correct prefix 'random_forest__'\n",
        "param_dist = {\n",
        "    'random_forest__n_estimators': [int(x) for x in np.linspace(start=100, stop=1000, num=10)],\n",
        "    'random_forest__max_depth': [None, 10, 20, 30, 40, 50],\n",
        "    'random_forest__min_samples_split': [2, 5, 10],\n",
        "    'random_forest__min_samples_leaf': [1, 2, 4],\n",
        "    'random_forest__bootstrap': [True, False]\n",
        "}\n",
        "\n",
        "# Initialize RandomizedSearchCV\n",
        "random_search = RandomizedSearchCV(full_pipeline, param_distributions=param_dist,\n",
        "                                   n_iter=100, cv=5, verbose=2, random_state=42, n_jobs=-1)\n",
        "\n",
        "# Fit the RandomizedSearchCV model\n",
        "random_search.fit(X_train, y_train)\n",
        "\n",
        "# Best hyperparameters\n",
        "print(\"Best Hyperparameters from RandomizedSearchCV:\", random_search.best_params_)\n",
        "\n",
        "# Re-fit the model with the best parameters\n",
        "best_rf_random_model = random_search.best_estimator_"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "EVz07QxyDohS"
      },
      "source": [
        "We can now re-evaluate the tuned Random Forest model on the test set using accuracy, precision, recall, F1-score, and ROC-AUC."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 119,
      "metadata": {
        "id": "UXuoobJNmSEb",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "7d0ecdff-1603-4895-c35e-7bfaa463a211"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Training Set Evaluation:\n",
            "Accuracy: 0.9663\n",
            "Precision: 0.9540\n",
            "Recall: 1.0000\n",
            "F1-Score: 0.9765\n",
            "ROC-AUC: 0.9987\n"
          ]
        }
      ],
      "source": [
        "from sklearn.metrics import accuracy_score, precision_score, recall_score, f1_score, roc_auc_score\n",
        "\n",
        "# Make predictions on the training set\n",
        "y_train_pred = best_rf_model.predict(X_train)\n",
        "\n",
        "# Calculate evaluation metrics\n",
        "train_accuracy = accuracy_score(y_train, y_train_pred)\n",
        "train_precision = precision_score(y_train, y_train_pred, pos_label=\"good\")\n",
        "train_recall = recall_score(y_train, y_train_pred, pos_label=\"good\")\n",
        "train_f1 = f1_score(y_train, y_train_pred, pos_label=\"good\")\n",
        "train_roc_auc = roc_auc_score(y_train, best_rf_model.predict_proba(X_train)[:, 1])\n",
        "\n",
        "# Print results\n",
        "print(f\"Training Set Evaluation:\")\n",
        "print(f\"Accuracy: {train_accuracy:.4f}\")\n",
        "print(f\"Precision: {train_precision:.4f}\")\n",
        "print(f\"Recall: {train_recall:.4f}\")\n",
        "print(f\"F1-Score: {train_f1:.4f}\")\n",
        "print(f\"ROC-AUC: {train_roc_auc:.4f}\")\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 122,
      "metadata": {
        "id": "J8ZjjCNlORt9",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 395
        },
        "outputId": "2e01639c-968d-43c3-cc1e-75b39e3e0e91"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 600x400 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ],
      "source": [
        "# Importing necessary libraries for confusion matrix\n",
        "from sklearn.metrics import confusion_matrix\n",
        "import matplotlib.pyplot as plt\n",
        "import seaborn as sns\n",
        "\n",
        "# Predicting using the best_rf_random_model on the training set\n",
        "y_train_pred = best_rf_random_model.predict(X_train)\n",
        "\n",
        "# Creating the confusion matrix using the training set\n",
        "cm = confusion_matrix(y_train, y_train_pred)\n",
        "\n",
        "# Plotting the confusion matrix\n",
        "plt.figure(figsize=(6, 4))\n",
        "sns.heatmap(cm, annot=True, fmt='d', cmap='Blues', cbar=False)\n",
        "plt.title('Confusion Matrix for Random Forest (RandomizedSearchCV - Training Set)')\n",
        "plt.xlabel('Predicted Labels')\n",
        "plt.ylabel('True Labels')\n",
        "plt.show()\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "FqyOA2-LHZYg"
      },
      "source": [
        "RandomizedSearchCV provides a model that has better precision than both the default model and the GridSearchCV-tuned model, with comparable recall and F1-score to the default model. It’s a strong performer in terms of overall balance between precision and recall."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "E_eWlcL2I1jP"
      },
      "source": [
        "We can better visualize the results of the randomized search using the following graph."
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import matplotlib.pyplot as plt\n",
        "import pandas as pd\n",
        "\n",
        "# Extract the results from the RandomizedSearchCV object\n",
        "results = pd.DataFrame(random_search.cv_results_)\n",
        "\n",
        "# Sort by the mean test score to make it easier to visualize the best performing models\n",
        "results = results.sort_values(by='mean_test_score', ascending=False)\n",
        "\n",
        "# Plot the top 10 models tested by RandomizedSearchCV based on mean test score\n",
        "plt.figure(figsize=(10, 6))\n",
        "plt.plot(range(1, 11), results['mean_test_score'].head(10), marker='o', linestyle='-', color='b')\n",
        "plt.title('Top 10 Model Scores from RandomizedSearchCV')\n",
        "plt.xlabel('Model Rank (Based on Test Score)')\n",
        "plt.ylabel('Mean Test Score (Accuracy)')\n",
        "plt.grid(True)\n",
        "plt.xticks(range(1, 11))  # Show ticks for top 10 models\n",
        "plt.show()\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 549
        },
        "id": "SzUfNXTc0mEv",
        "outputId": "708dc5cc-17db-4b87-9832-16f21d232c78"
      },
      "execution_count": 123,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1000x600 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "EAe0LjUvmXqf"
      },
      "source": [
        "# Prediction"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "After the tuning we can now make predictions using the test set. We ge the following results."
      ],
      "metadata": {
        "id": "RxwlEzsF5Psk"
      }
    },
    {
      "cell_type": "code",
      "execution_count": 126,
      "metadata": {
        "id": "6DfnWq2NmbT2",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 502
        },
        "outputId": "8816e946-66df-46c9-c972-c00a0b8bbe0f"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Evaluation on Test Set:\n",
            "Accuracy: 0.7800\n",
            "Precision: 0.8000\n",
            "Recall: 0.9143\n",
            "F1-Score: 0.8533\n",
            "ROC-AUC: 0.7915\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 600x400 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ],
      "source": [
        "from sklearn.metrics import accuracy_score, precision_score, recall_score, f1_score, roc_auc_score\n",
        "from sklearn.metrics import confusion_matrix\n",
        "import seaborn as sns\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "# Make predictions on the test set\n",
        "y_test_pred = best_rf_random_model.predict(X_test)\n",
        "\n",
        "# Calculate evaluation metrics\n",
        "test_accuracy = accuracy_score(y_test, y_test_pred)\n",
        "test_precision = precision_score(y_test, y_test_pred, pos_label=\"good\")\n",
        "test_recall = recall_score(y_test, y_test_pred, pos_label=\"good\")\n",
        "test_f1 = f1_score(y_test, y_test_pred, pos_label=\"good\")\n",
        "test_roc_auc = roc_auc_score(y_test, best_rf_random_model.predict_proba(X_test)[:, 1])\n",
        "\n",
        "# Print the evaluation metrics\n",
        "print(\"Evaluation on Test Set:\")\n",
        "print(f\"Accuracy: {test_accuracy:.4f}\")\n",
        "print(f\"Precision: {test_precision:.4f}\")\n",
        "print(f\"Recall: {test_recall:.4f}\")\n",
        "print(f\"F1-Score: {test_f1:.4f}\")\n",
        "print(f\"ROC-AUC: {test_roc_auc:.4f}\")\n",
        "\n",
        "# Confusion Matrix\n",
        "cm = confusion_matrix(y_test, y_test_pred)\n",
        "plt.figure(figsize=(6, 4))\n",
        "sns.heatmap(cm, annot=True, fmt='d', cmap='Blues', cbar=False)\n",
        "plt.title('Confusion Matrix for Random Forest (RandomizedSearchCV)')\n",
        "plt.xlabel('Predicted Labels')\n",
        "plt.ylabel('True Labels')\n",
        "plt.show()\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "bCOfkO5mmbz3"
      },
      "source": [
        "# Interpretation"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "AwWAZNE8e0QD"
      },
      "source": [
        " Random Forest was selected as the final model due to its superior balance of performance and interpretability. The Random Forest model was used to predict credit risk, classifying applicants as either good or bad credit risks. To better understand the factors influencing these predictions, we analyzed the model's performance and the importance of key features that drive its decision-making. Here is a detailed interpretation of the results, insights, and potential improvements:\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "en1av7Mhss8r"
      },
      "source": [
        "Model Predictions\n",
        "\n",
        "The Random Forest model, tuned with RandomizedSearchCV, produced the following performance metrics:\n",
        "\n",
        "1. Accuracy: 78% ; This indicates that the model correctly classified credit applicants 78% of the time.\n",
        "\n",
        "2. Precision: 80%; Out of all applicants predicted as good credit risks, 79.75% were correctly classified. This shows the model is reasonably good at avoiding false positives (approving bad credit risks).\n",
        "\n",
        "3. Recall: 91.43%; The model identified 91.43% of actual good credit risks correctly. This high recall means the model effectively minimizes false negatives, ensuring that most good credit applicants are approved.\n",
        "\n",
        "4. F1-Score: 85.33%; The F1-score combines both precision and recall, providing a balanced view of the model’s performance. With an 85.81% score, the model strikes a good balance between correctly identifying good credit risks and avoiding risky approvals.\n",
        "\n",
        "5. ROC-AUC: 79.15%; This score reflects how well the model distinguishes between good and bad credit risks. With a score of 79.15%, the model shows a solid ability to separate the two classes, though there is room for improvement."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "7ooV_z7m5hi5"
      },
      "source": [
        "Confusion Matrix:"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "k3kY4exB5sbR"
      },
      "source": [
        "The confusion matrix provides a detailed breakdown of the model's performance in classifying credit applicants:\n",
        "\n",
        "1. True Positives (TP): The model correctly identified applicants as good credit risks.\n",
        "\n",
        "2. True Negatives (TN): The model correctly identified applicants as bad credit risks.\n",
        "\n",
        "3. False Positives (FP): The model incorrectly classified bad credit risks as good, potentially leading to financial losses.\n",
        "\n",
        "4. False Negatives (FN): The model incorrectly classified good credit risks as bad, resulting in missed business opportunities."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "R-vz0gEX5_00"
      },
      "source": [
        "The model has high true positives and true negatives, but the presence of false positives means that the model occasionally approves credit to risky applicants.The presence of false positives suggests that more data could improve its ability to accurately classify credit applicants. Additional training data can help the model learn more nuanced patterns in credit behavior, which will reduce false positives"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Feature Importance:\n",
        "\n",
        "The Random Forest model highlights key features that strongly impact credit risk classification. These features help determine whether an applicant is considered a good or bad credit risk based on their financial status. Below are the features from the German Credit Risk dataset contribute to the model’s predictions:"
      ],
      "metadata": {
        "id": "SEsT7Pr7cXWb"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "1. Monthly Credit: This is the ratio of Credit Amount to Duration.This is an important feature as it reflects the financial pressure on the applicant. A higher monthly burden suggests a greater likelihood of default.\n",
        "\n",
        "2. Credit Amount: The total amount of credit or loan requested plays a significant role in determining credit risk. Larger loan amounts typically carry higher risks, especially if the applicant has limited income\n",
        "\n",
        "3. Duration: The length of the loan is an important factor. While longer loan terms allow for smaller payments over time, they also increase the risk of non-payment as the applicant's financial situation may become uncertain in the long run.\n",
        "\n",
        "4. Age of Applicant: The age of the applicant influences the risk assessment. Younger applicants may have less stable income, while older applicants could face retirement and lower income levels.\n",
        "\n",
        "5. Employment Status: An applicant’s savings and employment history also significantly impact the prediction. Applicants with steady employment with good savings are considered to be at lower risk.\n",
        "\n",
        "6. Checking Staus: The state of the applicant's checking account shows their available funds and plays an important role in evaluating credit risk.\n",
        "\n",
        "7. Installment Commitment: The proportion of the applicant’s income that goes toward loan payments. A higher installment commitment leaves less room for financial flexibility, increasing the risk of default."
      ],
      "metadata": {
        "id": "JthYVtMecy4_"
      }
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "U4ABPxMbmf5C"
      },
      "source": [
        "The follow code indicates when the notebook was last run and signals when execution is complete."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 127,
      "metadata": {
        "id": "-PSPvYM_S0to",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "f04f47a8-acef-4184-a885-b363972c3c9d"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Execution completed at:  2024-10-22 01:48:28.269161\n"
          ]
        }
      ],
      "source": [
        "from datetime import datetime\n",
        "\n",
        "# Get current time\n",
        "current_time = datetime.now()\n",
        "\n",
        "# Print current time\n",
        "print(\"Execution completed at: \", current_time)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 128,
      "metadata": {
        "id": "TnD77j-5S_Jy",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 17
        },
        "outputId": "62afd31a-3e91-4215-9520-684120ff28eb"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.Javascript object>"
            ],
            "application/javascript": [
              "alert(\"Execution is complete!\")"
            ]
          },
          "metadata": {}
        }
      ],
      "source": [
        "from IPython.display import display, Javascript\n",
        "\n",
        "# Display an alert when execution is complete\n",
        "def alert_completion():\n",
        "    display(Javascript('alert(\"Execution is complete!\")'))\n",
        "\n",
        "# Signal completion\n",
        "alert_completion()\n"
      ]
    }
  ],
  "metadata": {
    "colab": {
      "collapsed_sections": [
        "XbAypvmqQEuG",
        "meb_w3kDU6IJ",
        "ISqW97Mlwo4i",
        "7wKTSZEO7jBU",
        "Wt45INQvoth8"
      ],
      "provenance": []
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}