logistic_regression.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "logistic_regression.ipynb",
      "version": "0.3.2",
      "provenance": [],
      "collapsed_sections": [
        "JndnmDMp66FL",
        "dPpJUV862FYI",
        "i2e3TlyL57Qs",
        "wCugvl0JdWYL"
      ],
      "include_colab_link": true
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "[View in Colaboratory](https://colab.research.google.com/github/diegoreico/machine-learning-crash-course/blob/master/logistic_regression.ipynb)"
      ]
    },
    {
      "metadata": {
        "id": "JndnmDMp66FL",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "#### Copyright 2017 Google LLC."
      ]
    },
    {
      "metadata": {
        "id": "hMqWDc_m6rUC",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "# Licensed under the Apache License, Version 2.0 (the \"License\");\n",
        "# you may not use this file except in compliance with the License.\n",
        "# You may obtain a copy of the License at\n",
        "#\n",
        "# https://www.apache.org/licenses/LICENSE-2.0\n",
        "#\n",
        "# Unless required by applicable law or agreed to in writing, software\n",
        "# distributed under the License is distributed on an \"AS IS\" BASIS,\n",
        "# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n",
        "# See the License for the specific language governing permissions and\n",
        "# limitations under the License."
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "g4T-_IsVbweU",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "# Logistic Regression"
      ]
    },
    {
      "metadata": {
        "id": "LEAHZv4rIYHX",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "**Learning Objectives:**\n",
        "  * Reframe the median house value predictor (from the preceding exercises) as a binary classification model\n",
        "  * Compare the effectiveness of logisitic regression vs linear regression for a binary classification problem"
      ]
    },
    {
      "metadata": {
        "id": "CnkCZqdIIYHY",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "As in the prior exercises, we're working with the [California housing data set](https://developers.google.com/machine-learning/crash-course/california-housing-data-description), but this time we will turn it into a binary classification problem by predicting whether a city block is a high-cost city block. We'll also revert to the default features, for now."
      ]
    },
    {
      "metadata": {
        "id": "9pltCyy2K3dd",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## Frame the Problem as Binary Classification\n",
        "\n",
        "The target of our dataset is `median_house_value` which is a numeric (continuous-valued) feature. We can create a boolean label by applying a threshold to this continuous value.\n",
        "\n",
        "Given features describing a city block, we wish to predict if it is a high-cost city block. To prepare the targets for train and eval data, we define a classification threshold of the 75%-ile for median house value (a value of approximately 265000). All house values above the threshold are labeled `1`, and all others are labeled `0`."
      ]
    },
    {
      "metadata": {
        "id": "67IJwZX1Vvjt",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## Setup\n",
        "\n",
        "Run the cells below to load the data and prepare the input features and targets."
      ]
    },
    {
      "metadata": {
        "id": "fOlbcJ4EIYHd",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "from __future__ import print_function\n",
        "\n",
        "import math\n",
        "\n",
        "from IPython import display\n",
        "from matplotlib import cm\n",
        "from matplotlib import gridspec\n",
        "from matplotlib import pyplot as plt\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "from sklearn import metrics\n",
        "import tensorflow as tf\n",
        "from tensorflow.python.data import Dataset\n",
        "\n",
        "tf.logging.set_verbosity(tf.logging.ERROR)\n",
        "pd.options.display.max_rows = 10\n",
        "pd.options.display.float_format = '{:.1f}'.format\n",
        "\n",
        "california_housing_dataframe = pd.read_csv(\"https://dl.google.com/mlcc/mledu-datasets/california_housing_train.csv\", sep=\",\")\n",
        "\n",
        "california_housing_dataframe = california_housing_dataframe.reindex(\n",
        "    np.random.permutation(california_housing_dataframe.index))"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "lTB73MNeIYHf",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "Note how the code below is slightly different from the previous exercises. Instead of using `median_house_value` as target, we create a new binary target, `median_house_value_is_high`."
      ]
    },
    {
      "metadata": {
        "id": "kPSqspaqIYHg",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "def preprocess_features(california_housing_dataframe):\n",
        "  \"\"\"Prepares input features from California housing data set.\n",
        "\n",
        "  Args:\n",
        "    california_housing_dataframe: A Pandas DataFrame expected to contain data\n",
        "      from the California housing data set.\n",
        "  Returns:\n",
        "    A DataFrame that contains the features to be used for the model, including\n",
        "    synthetic features.\n",
        "  \"\"\"\n",
        "  selected_features = california_housing_dataframe[\n",
        "    [\"latitude\",\n",
        "     \"longitude\",\n",
        "     \"housing_median_age\",\n",
        "     \"total_rooms\",\n",
        "     \"total_bedrooms\",\n",
        "     \"population\",\n",
        "     \"households\",\n",
        "     \"median_income\"]]\n",
        "  processed_features = selected_features.copy()\n",
        "  # Create a synthetic feature.\n",
        "  processed_features[\"rooms_per_person\"] = (\n",
        "    california_housing_dataframe[\"total_rooms\"] /\n",
        "    california_housing_dataframe[\"population\"])\n",
        "  return processed_features\n",
        "\n",
        "def preprocess_targets(california_housing_dataframe):\n",
        "  \"\"\"Prepares target features (i.e., labels) from California housing data set.\n",
        "\n",
        "  Args:\n",
        "    california_housing_dataframe: A Pandas DataFrame expected to contain data\n",
        "      from the California housing data set.\n",
        "  Returns:\n",
        "    A DataFrame that contains the target feature.\n",
        "  \"\"\"\n",
        "  output_targets = pd.DataFrame()\n",
        "  # Create a boolean categorical feature representing whether the\n",
        "  # median_house_value is above a set threshold.\n",
        "  output_targets[\"median_house_value_is_high\"] = (\n",
        "    california_housing_dataframe[\"median_house_value\"] > 265000).astype(float)\n",
        "  return output_targets"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "FwOYWmXqWA6D",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1229
        },
        "outputId": "788c20ad-b8be-4bf0-ad68-56f192366408"
      },
      "cell_type": "code",
      "source": [
        "# Choose the first 12000 (out of 17000) examples for training.\n",
        "training_examples = preprocess_features(california_housing_dataframe.head(12000))\n",
        "training_targets = preprocess_targets(california_housing_dataframe.head(12000))\n",
        "\n",
        "# Choose the last 5000 (out of 17000) examples for validation.\n",
        "validation_examples = preprocess_features(california_housing_dataframe.tail(5000))\n",
        "validation_targets = preprocess_targets(california_housing_dataframe.tail(5000))\n",
        "\n",
        "# Double-check that we've done the right thing.\n",
        "print(\"Training examples summary:\")\n",
        "display.display(training_examples.describe())\n",
        "print(\"Validation examples summary:\")\n",
        "display.display(validation_examples.describe())\n",
        "\n",
        "print(\"Training targets summary:\")\n",
        "display.display(training_targets.describe())\n",
        "print(\"Validation targets summary:\")\n",
        "display.display(validation_targets.describe())"
      ],
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Training examples summary:\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "       latitude  longitude  housing_median_age  total_rooms  total_bedrooms  \\\n",
              "count   12000.0    12000.0             12000.0      12000.0         12000.0   \n",
              "mean       35.6     -119.6                28.5       2663.4           542.5   \n",
              "std         2.1        2.0                12.6       2235.2           430.4   \n",
              "min        32.5     -124.3                 1.0          2.0             1.0   \n",
              "25%        33.9     -121.8                18.0       1469.0           297.0   \n",
              "50%        34.2     -118.5                29.0       2141.0           434.0   \n",
              "75%        37.7     -118.0                37.0       3157.0           652.0   \n",
              "max        42.0     -114.6                52.0      37937.0          6445.0   \n",
              "\n",
              "       population  households  median_income  rooms_per_person  \n",
              "count     12000.0     12000.0        12000.0           12000.0  \n",
              "mean       1441.5       504.7            3.9               2.0  \n",
              "std        1190.2       394.2            1.9               1.2  \n",
              "min           3.0         1.0            0.5               0.0  \n",
              "25%         791.0       282.8            2.6               1.5  \n",
              "50%        1172.0       409.0            3.5               1.9  \n",
              "75%        1737.0       607.0            4.8               2.3  \n",
              "max       35682.0      6082.0           15.0              55.2  "
            ],
            "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>latitude</th>\n",
              "      <th>longitude</th>\n",
              "      <th>housing_median_age</th>\n",
              "      <th>total_rooms</th>\n",
              "      <th>total_bedrooms</th>\n",
              "      <th>population</th>\n",
              "      <th>households</th>\n",
              "      <th>median_income</th>\n",
              "      <th>rooms_per_person</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>count</th>\n",
              "      <td>12000.0</td>\n",
              "      <td>12000.0</td>\n",
              "      <td>12000.0</td>\n",
              "      <td>12000.0</td>\n",
              "      <td>12000.0</td>\n",
              "      <td>12000.0</td>\n",
              "      <td>12000.0</td>\n",
              "      <td>12000.0</td>\n",
              "      <td>12000.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>mean</th>\n",
              "      <td>35.6</td>\n",
              "      <td>-119.6</td>\n",
              "      <td>28.5</td>\n",
              "      <td>2663.4</td>\n",
              "      <td>542.5</td>\n",
              "      <td>1441.5</td>\n",
              "      <td>504.7</td>\n",
              "      <td>3.9</td>\n",
              "      <td>2.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>std</th>\n",
              "      <td>2.1</td>\n",
              "      <td>2.0</td>\n",
              "      <td>12.6</td>\n",
              "      <td>2235.2</td>\n",
              "      <td>430.4</td>\n",
              "      <td>1190.2</td>\n",
              "      <td>394.2</td>\n",
              "      <td>1.9</td>\n",
              "      <td>1.2</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>min</th>\n",
              "      <td>32.5</td>\n",
              "      <td>-124.3</td>\n",
              "      <td>1.0</td>\n",
              "      <td>2.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>3.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.5</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>25%</th>\n",
              "      <td>33.9</td>\n",
              "      <td>-121.8</td>\n",
              "      <td>18.0</td>\n",
              "      <td>1469.0</td>\n",
              "      <td>297.0</td>\n",
              "      <td>791.0</td>\n",
              "      <td>282.8</td>\n",
              "      <td>2.6</td>\n",
              "      <td>1.5</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>50%</th>\n",
              "      <td>34.2</td>\n",
              "      <td>-118.5</td>\n",
              "      <td>29.0</td>\n",
              "      <td>2141.0</td>\n",
              "      <td>434.0</td>\n",
              "      <td>1172.0</td>\n",
              "      <td>409.0</td>\n",
              "      <td>3.5</td>\n",
              "      <td>1.9</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>75%</th>\n",
              "      <td>37.7</td>\n",
              "      <td>-118.0</td>\n",
              "      <td>37.0</td>\n",
              "      <td>3157.0</td>\n",
              "      <td>652.0</td>\n",
              "      <td>1737.0</td>\n",
              "      <td>607.0</td>\n",
              "      <td>4.8</td>\n",
              "      <td>2.3</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>max</th>\n",
              "      <td>42.0</td>\n",
              "      <td>-114.6</td>\n",
              "      <td>52.0</td>\n",
              "      <td>37937.0</td>\n",
              "      <td>6445.0</td>\n",
              "      <td>35682.0</td>\n",
              "      <td>6082.0</td>\n",
              "      <td>15.0</td>\n",
              "      <td>55.2</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "Validation examples summary:\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "       latitude  longitude  housing_median_age  total_rooms  total_bedrooms  \\\n",
              "count    5000.0     5000.0              5000.0       5000.0          5000.0   \n",
              "mean       35.6     -119.6                28.7       2596.2           531.9   \n",
              "std         2.1        2.0                12.5       2040.7           399.3   \n",
              "min        32.5     -124.3                 2.0         15.0             4.0   \n",
              "25%        33.9     -121.8                18.0       1444.5           296.0   \n",
              "50%        34.2     -118.5                29.0       2110.0           432.5   \n",
              "75%        37.7     -118.0                37.0       3129.2           644.0   \n",
              "max        41.9     -114.3                52.0      22128.0          4386.0   \n",
              "\n",
              "       population  households  median_income  rooms_per_person  \n",
              "count      5000.0      5000.0         5000.0            5000.0  \n",
              "mean       1400.9       493.0            3.9               2.0  \n",
              "std        1038.7       360.0            1.9               1.2  \n",
              "min          13.0         2.0            0.5               0.1  \n",
              "25%         787.8       279.0            2.5               1.5  \n",
              "50%        1152.0       409.0            3.5               1.9  \n",
              "75%        1689.2       598.0            4.7               2.3  \n",
              "max       12427.0      3886.0           15.0              52.0  "
            ],
            "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>latitude</th>\n",
              "      <th>longitude</th>\n",
              "      <th>housing_median_age</th>\n",
              "      <th>total_rooms</th>\n",
              "      <th>total_bedrooms</th>\n",
              "      <th>population</th>\n",
              "      <th>households</th>\n",
              "      <th>median_income</th>\n",
              "      <th>rooms_per_person</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>count</th>\n",
              "      <td>5000.0</td>\n",
              "      <td>5000.0</td>\n",
              "      <td>5000.0</td>\n",
              "      <td>5000.0</td>\n",
              "      <td>5000.0</td>\n",
              "      <td>5000.0</td>\n",
              "      <td>5000.0</td>\n",
              "      <td>5000.0</td>\n",
              "      <td>5000.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>mean</th>\n",
              "      <td>35.6</td>\n",
              "      <td>-119.6</td>\n",
              "      <td>28.7</td>\n",
              "      <td>2596.2</td>\n",
              "      <td>531.9</td>\n",
              "      <td>1400.9</td>\n",
              "      <td>493.0</td>\n",
              "      <td>3.9</td>\n",
              "      <td>2.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>std</th>\n",
              "      <td>2.1</td>\n",
              "      <td>2.0</td>\n",
              "      <td>12.5</td>\n",
              "      <td>2040.7</td>\n",
              "      <td>399.3</td>\n",
              "      <td>1038.7</td>\n",
              "      <td>360.0</td>\n",
              "      <td>1.9</td>\n",
              "      <td>1.2</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>min</th>\n",
              "      <td>32.5</td>\n",
              "      <td>-124.3</td>\n",
              "      <td>2.0</td>\n",
              "      <td>15.0</td>\n",
              "      <td>4.0</td>\n",
              "      <td>13.0</td>\n",
              "      <td>2.0</td>\n",
              "      <td>0.5</td>\n",
              "      <td>0.1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>25%</th>\n",
              "      <td>33.9</td>\n",
              "      <td>-121.8</td>\n",
              "      <td>18.0</td>\n",
              "      <td>1444.5</td>\n",
              "      <td>296.0</td>\n",
              "      <td>787.8</td>\n",
              "      <td>279.0</td>\n",
              "      <td>2.5</td>\n",
              "      <td>1.5</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>50%</th>\n",
              "      <td>34.2</td>\n",
              "      <td>-118.5</td>\n",
              "      <td>29.0</td>\n",
              "      <td>2110.0</td>\n",
              "      <td>432.5</td>\n",
              "      <td>1152.0</td>\n",
              "      <td>409.0</td>\n",
              "      <td>3.5</td>\n",
              "      <td>1.9</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>75%</th>\n",
              "      <td>37.7</td>\n",
              "      <td>-118.0</td>\n",
              "      <td>37.0</td>\n",
              "      <td>3129.2</td>\n",
              "      <td>644.0</td>\n",
              "      <td>1689.2</td>\n",
              "      <td>598.0</td>\n",
              "      <td>4.7</td>\n",
              "      <td>2.3</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>max</th>\n",
              "      <td>41.9</td>\n",
              "      <td>-114.3</td>\n",
              "      <td>52.0</td>\n",
              "      <td>22128.0</td>\n",
              "      <td>4386.0</td>\n",
              "      <td>12427.0</td>\n",
              "      <td>3886.0</td>\n",
              "      <td>15.0</td>\n",
              "      <td>52.0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "Training targets summary:\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "       median_house_value_is_high\n",
              "count                     12000.0\n",
              "mean                          0.2\n",
              "std                           0.4\n",
              "min                           0.0\n",
              "25%                           0.0\n",
              "50%                           0.0\n",
              "75%                           0.0\n",
              "max                           1.0"
            ],
            "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>median_house_value_is_high</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>count</th>\n",
              "      <td>12000.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>mean</th>\n",
              "      <td>0.2</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>std</th>\n",
              "      <td>0.4</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>min</th>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>25%</th>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>50%</th>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>75%</th>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>max</th>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "Validation targets summary:\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "       median_house_value_is_high\n",
              "count                      5000.0\n",
              "mean                          0.3\n",
              "std                           0.4\n",
              "min                           0.0\n",
              "25%                           0.0\n",
              "50%                           0.0\n",
              "75%                           1.0\n",
              "max                           1.0"
            ],
            "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>median_house_value_is_high</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>count</th>\n",
              "      <td>5000.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>mean</th>\n",
              "      <td>0.3</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>std</th>\n",
              "      <td>0.4</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>min</th>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>25%</th>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>50%</th>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>75%</th>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>max</th>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "metadata": {
        "id": "uon1LB3A31VN",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## How Would Linear Regression Fare?\n",
        "To see why logistic regression is effective, let us first train a naive model that uses linear regression. This model will use labels with values in the set `{0, 1}` and will try to predict a continuous value that is as close as possible to `0` or `1`. Furthermore, we wish to interpret the output as a probability, so it would be ideal if the output will be within the range `(0, 1)`. We would then apply a threshold of `0.5` to determine the label.\n",
        "\n",
        "Run the cells below to train the linear regression model using [LinearRegressor](https://www.tensorflow.org/api_docs/python/tf/estimator/LinearRegressor)."
      ]
    },
    {
      "metadata": {
        "id": "smmUYRDtWOV_",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "def construct_feature_columns(input_features):\n",
        "  \"\"\"Construct the TensorFlow Feature Columns.\n",
        "\n",
        "  Args:\n",
        "    input_features: The names of the numerical input features to use.\n",
        "  Returns:\n",
        "    A set of feature columns\n",
        "  \"\"\"\n",
        "  return set([tf.feature_column.numeric_column(my_feature)\n",
        "              for my_feature in input_features])"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "B5OwSrr1yIKD",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "def my_input_fn(features, targets, batch_size=1, shuffle=True, num_epochs=None):\n",
        "    \"\"\"Trains a linear regression model.\n",
        "  \n",
        "    Args:\n",
        "      features: pandas DataFrame of features\n",
        "      targets: pandas DataFrame of targets\n",
        "      batch_size: Size of batches to be passed to the model\n",
        "      shuffle: True or False. Whether to shuffle the data.\n",
        "      num_epochs: Number of epochs for which data should be repeated. None = repeat indefinitely\n",
        "    Returns:\n",
        "      Tuple of (features, labels) for next data batch\n",
        "    \"\"\"\n",
        "    \n",
        "    # Convert pandas data into a dict of np arrays.\n",
        "    features = {key:np.array(value) for key,value in dict(features).items()}                                            \n",
        " \n",
        "    # Construct a dataset, and configure batching/repeating.\n",
        "    ds = Dataset.from_tensor_slices((features,targets)) # warning: 2GB limit\n",
        "    ds = ds.batch(batch_size).repeat(num_epochs)\n",
        "    \n",
        "    # Shuffle the data, if specified.\n",
        "    if shuffle:\n",
        "      ds = ds.shuffle(10000)\n",
        "    \n",
        "    # Return the next batch of data.\n",
        "    features, labels = ds.make_one_shot_iterator().get_next()\n",
        "    return features, labels"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "SE2-hq8PIYHz",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "def train_linear_regressor_model(\n",
        "    learning_rate,\n",
        "    steps,\n",
        "    batch_size,\n",
        "    training_examples,\n",
        "    training_targets,\n",
        "    validation_examples,\n",
        "    validation_targets):\n",
        "  \"\"\"Trains a linear regression model.\n",
        "  \n",
        "  In addition to training, this function also prints training progress information,\n",
        "  as well as a plot of the training and validation loss over time.\n",
        "  \n",
        "  Args:\n",
        "    learning_rate: A `float`, the learning rate.\n",
        "    steps: A non-zero `int`, the total number of training steps. A training step\n",
        "      consists of a forward and backward pass using a single batch.\n",
        "    batch_size: A non-zero `int`, the batch size.\n",
        "    training_examples: A `DataFrame` containing one or more columns from\n",
        "      `california_housing_dataframe` to use as input features for training.\n",
        "    training_targets: A `DataFrame` containing exactly one column from\n",
        "      `california_housing_dataframe` to use as target for training.\n",
        "    validation_examples: A `DataFrame` containing one or more columns from\n",
        "      `california_housing_dataframe` to use as input features for validation.\n",
        "    validation_targets: A `DataFrame` containing exactly one column from\n",
        "      `california_housing_dataframe` to use as target for validation.\n",
        "      \n",
        "  Returns:\n",
        "    A `LinearRegressor` object trained on the training data.\n",
        "  \"\"\"\n",
        "\n",
        "  periods = 10\n",
        "  steps_per_period = steps / periods\n",
        "\n",
        "  # Create a linear regressor object.\n",
        "  my_optimizer = tf.train.GradientDescentOptimizer(learning_rate=learning_rate)\n",
        "  my_optimizer = tf.contrib.estimator.clip_gradients_by_norm(my_optimizer, 5.0)\n",
        "  linear_regressor = tf.estimator.LinearRegressor(\n",
        "      feature_columns=construct_feature_columns(training_examples),\n",
        "      optimizer=my_optimizer\n",
        "  )\n",
        "    \n",
        "  # Create input functions.\n",
        "  training_input_fn = lambda: my_input_fn(training_examples, \n",
        "                                          training_targets[\"median_house_value_is_high\"], \n",
        "                                          batch_size=batch_size)\n",
        "  predict_training_input_fn = lambda: my_input_fn(training_examples, \n",
        "                                                  training_targets[\"median_house_value_is_high\"], \n",
        "                                                  num_epochs=1, \n",
        "                                                  shuffle=False)\n",
        "  predict_validation_input_fn = lambda: my_input_fn(validation_examples, \n",
        "                                                    validation_targets[\"median_house_value_is_high\"], \n",
        "                                                    num_epochs=1, \n",
        "                                                    shuffle=False)\n",
        "\n",
        "  # Train the model, but do so inside a loop so that we can periodically assess\n",
        "  # loss metrics.\n",
        "  print(\"Training model...\")\n",
        "  print(\"RMSE (on training data):\")\n",
        "  training_rmse = []\n",
        "  validation_rmse = []\n",
        "  for period in range (0, periods):\n",
        "    # Train the model, starting from the prior state.\n",
        "    linear_regressor.train(\n",
        "        input_fn=training_input_fn,\n",
        "        steps=steps_per_period\n",
        "    )\n",
        "    \n",
        "    # Take a break and compute predictions.\n",
        "    training_predictions = linear_regressor.predict(input_fn=predict_training_input_fn)\n",
        "    training_predictions = np.array([item['predictions'][0] for item in training_predictions])\n",
        "    \n",
        "    validation_predictions = linear_regressor.predict(input_fn=predict_validation_input_fn)\n",
        "    validation_predictions = np.array([item['predictions'][0] for item in validation_predictions])\n",
        "    \n",
        "    # Compute training and validation loss.\n",
        "    training_root_mean_squared_error = math.sqrt(\n",
        "        metrics.mean_squared_error(training_predictions, training_targets))\n",
        "    validation_root_mean_squared_error = math.sqrt(\n",
        "        metrics.mean_squared_error(validation_predictions, validation_targets))\n",
        "    # Occasionally print the current loss.\n",
        "    print(\"  period %02d : %0.2f\" % (period, training_root_mean_squared_error))\n",
        "    # Add the loss metrics from this period to our list.\n",
        "    training_rmse.append(training_root_mean_squared_error)\n",
        "    validation_rmse.append(validation_root_mean_squared_error)\n",
        "  print(\"Model training finished.\")\n",
        "  \n",
        "  # Output a graph of loss metrics over periods.\n",
        "  plt.ylabel(\"RMSE\")\n",
        "  plt.xlabel(\"Periods\")\n",
        "  plt.title(\"Root Mean Squared Error vs. Periods\")\n",
        "  plt.tight_layout()\n",
        "  plt.plot(training_rmse, label=\"training\")\n",
        "  plt.plot(validation_rmse, label=\"validation\")\n",
        "  plt.legend()\n",
        "\n",
        "  return linear_regressor"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "TDBD8xeeIYH2",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 635
        },
        "outputId": "89564620-1832-4f61-8be6-55a7cd4f4235"
      },
      "cell_type": "code",
      "source": [
        "linear_regressor = train_linear_regressor_model(\n",
        "    learning_rate=0.000001,\n",
        "    steps=200,\n",
        "    batch_size=20,\n",
        "    training_examples=training_examples,\n",
        "    training_targets=training_targets,\n",
        "    validation_examples=validation_examples,\n",
        "    validation_targets=validation_targets)"
      ],
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Training model...\n",
            "RMSE (on training data):\n",
            "  period 00 : 0.45\n",
            "  period 01 : 0.45\n",
            "  period 02 : 0.45\n",
            "  period 03 : 0.44\n",
            "  period 04 : 0.45\n",
            "  period 05 : 0.44\n",
            "  period 06 : 0.44\n",
            "  period 07 : 0.44\n",
            "  period 08 : 0.45\n",
            "  period 09 : 0.44\n",
            "Model training finished.\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAjgAAAGACAYAAACgBBhzAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\ncgQOAFCOwAEAyhE4AEA5AgcAKOcHG8zKxnn9nG8AAAAASUVORK5CYII=\n",
            "text/plain": [
              "<matplotlib.figure.Figure at 0x7f68d70a3490>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "metadata": {
        "id": "JjBZ_q7aD9gh",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## Task 1: Can We Calculate LogLoss for These Predictions?\n",
        "\n",
        "**Examine the predictions and decide whether or not we can use them to calculate LogLoss.**\n",
        "\n",
        "`LinearRegressor` uses the L2 loss, which doesn't do a great job at penalizing misclassifications when the output is interpreted as a probability.  For example, there should be a huge difference whether a negative example is classified as positive with a probability of 0.9 vs 0.9999, but L2 loss doesn't strongly differentiate these cases.\n",
        "\n",
        "In contrast, `LogLoss` penalizes these \"confidence errors\" much more heavily.  Remember, `LogLoss` is defined as:\n",
        "\n",
        "$$Log Loss = \\sum_{(x,y)\\in D} -y \\cdot log(y_{pred}) - (1 - y) \\cdot log(1 - y_{pred})$$\n",
        "\n",
        "\n",
        "But first, we'll need to obtain the prediction values. We could use `LinearRegressor.predict` to obtain these.\n",
        "\n",
        "Given the predictions and the targets, can we calculate `LogLoss`?"
      ]
    },
    {
      "metadata": {
        "id": "dPpJUV862FYI",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "### Solution\n",
        "\n",
        "Click below to display the solution."
      ]
    },
    {
      "metadata": {
        "id": "kXFQ5uig2RoP",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 347
        },
        "outputId": "656d199a-8116-41fc-8cd6-a63a58f31d99"
      },
      "cell_type": "code",
      "source": [
        "predict_validation_input_fn = lambda: my_input_fn(validation_examples, \n",
        "                                                  validation_targets[\"median_house_value_is_high\"], \n",
        "                                                  num_epochs=1, \n",
        "                                                  shuffle=False)\n",
        "\n",
        "validation_predictions = linear_regressor.predict(input_fn=predict_validation_input_fn)\n",
        "validation_predictions = np.array([item['predictions'][0] for item in validation_predictions])\n",
        "\n",
        "_ = plt.hist(validation_predictions)"
      ],
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAeoAAAFKCAYAAADScRzUAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nDx7z4yOrfp9bAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<matplotlib.figure.Figure at 0x7f68e3d54f10>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "metadata": {
        "id": "rYpy336F9wBg",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## Task 2: Train a Logistic Regression Model and Calculate LogLoss on the Validation Set\n",
        "\n",
        "To use logistic regression, simply use [LinearClassifier](https://www.tensorflow.org/api_docs/python/tf/estimator/LinearClassifier) instead of `LinearRegressor`. Complete the code below.\n",
        "\n",
        "**NOTE**: When running `train()` and `predict()` on a `LinearClassifier` model, you can access the real-valued predicted probabilities via the `\"probabilities\"` key in the returned dict—e.g., `predictions[\"probabilities\"]`. Sklearn's [log_loss](http://scikit-learn.org/stable/modules/generated/sklearn.metrics.log_loss.html) function is handy for calculating LogLoss using these probabilities.\n"
      ]
    },
    {
      "metadata": {
        "id": "JElcb--E9wBm",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "def train_linear_classifier_model(\n",
        "    learning_rate,\n",
        "    steps,\n",
        "    batch_size,\n",
        "    training_examples,\n",
        "    training_targets,\n",
        "    validation_examples,\n",
        "    validation_targets):\n",
        "  \"\"\"Trains a linear classification model.\n",
        "  \n",
        "  In addition to training, this function also prints training progress information,\n",
        "  as well as a plot of the training and validation loss over time.\n",
        "  \n",
        "  Args:\n",
        "    learning_rate: A `float`, the learning rate.\n",
        "    steps: A non-zero `int`, the total number of training steps. A training step\n",
        "      consists of a forward and backward pass using a single batch.\n",
        "    batch_size: A non-zero `int`, the batch size.\n",
        "    training_examples: A `DataFrame` containing one or more columns from\n",
        "      `california_housing_dataframe` to use as input features for training.\n",
        "    training_targets: A `DataFrame` containing exactly one column from\n",
        "      `california_housing_dataframe` to use as target for training.\n",
        "    validation_examples: A `DataFrame` containing one or more columns from\n",
        "      `california_housing_dataframe` to use as input features for validation.\n",
        "    validation_targets: A `DataFrame` containing exactly one column from\n",
        "      `california_housing_dataframe` to use as target for validation.\n",
        "      \n",
        "  Returns:\n",
        "    A `LinearClassifier` object trained on the training data.\n",
        "  \"\"\"\n",
        "\n",
        "  periods = 10\n",
        "  steps_per_period = steps / periods\n",
        "  \n",
        "  # Create a linear classifier object.\n",
        "  my_optimizer = tf.train.GradientDescentOptimizer(learning_rate=learning_rate)\n",
        "  my_optimizer = tf.contrib.estimator.clip_gradients_by_norm(my_optimizer, 5.0)\n",
        "  linear_classifier = tf.estimator.LinearClassifier(\n",
        "      feature_columns=construct_feature_columns(training_examples),\n",
        "      optimizer=my_optimizer\n",
        "  )\n",
        "  \n",
        "  # Create input functions.\n",
        "  training_input_fn = lambda: my_input_fn(training_examples, \n",
        "                                          training_targets[\"median_house_value_is_high\"], \n",
        "                                          batch_size=batch_size)\n",
        "  predict_training_input_fn = lambda: my_input_fn(training_examples, \n",
        "                                                  training_targets[\"median_house_value_is_high\"], \n",
        "                                                  num_epochs=1, \n",
        "                                                  shuffle=False)\n",
        "  predict_validation_input_fn = lambda: my_input_fn(validation_examples, \n",
        "                                                    validation_targets[\"median_house_value_is_high\"], \n",
        "                                                    num_epochs=1, \n",
        "                                                    shuffle=False)\n",
        "  \n",
        "  # Train the model, but do so inside a loop so that we can periodically assess\n",
        "  # loss metrics.\n",
        "  print(\"Training model...\")\n",
        "  print(\"LogLoss (on training data):\")\n",
        "  training_log_losses = []\n",
        "  validation_log_losses = []\n",
        "  for period in range (0, periods):\n",
        "    # Train the model, starting from the prior state.\n",
        "    linear_classifier.train(\n",
        "        input_fn=training_input_fn,\n",
        "        steps=steps_per_period\n",
        "    )\n",
        "    # Take a break and compute predictions.    \n",
        "    training_probabilities = linear_classifier.predict(input_fn=predict_training_input_fn)\n",
        "    training_probabilities = np.array([item['probabilities'] for item in training_probabilities])\n",
        "    \n",
        "    validation_probabilities = linear_classifier.predict(input_fn=predict_validation_input_fn)\n",
        "    validation_probabilities = np.array([item['probabilities'] for item in validation_probabilities])\n",
        "    \n",
        "    training_log_loss = metrics.log_loss(training_targets, training_probabilities)\n",
        "    validation_log_loss = metrics.log_loss(validation_targets, validation_probabilities)\n",
        "    # Occasionally print the current loss.\n",
        "    print(\"  period %02d : %0.2f\" % (period, training_log_loss))\n",
        "    # Add the loss metrics from this period to our list.\n",
        "    training_log_losses.append(training_log_loss)\n",
        "    validation_log_losses.append(validation_log_loss)\n",
        "  print(\"Model training finished.\")\n",
        "  \n",
        "  # Output a graph of loss metrics over periods.\n",
        "  plt.ylabel(\"LogLoss\")\n",
        "  plt.xlabel(\"Periods\")\n",
        "  plt.title(\"LogLoss vs. Periods\")\n",
        "  plt.tight_layout()\n",
        "  plt.plot(training_log_losses, label=\"training\")\n",
        "  plt.plot(validation_log_losses, label=\"validation\")\n",
        "  plt.legend()\n",
        "\n",
        "  return linear_classifier"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "VM0wmnFUIYH9",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 635
        },
        "outputId": "5ef473e0-22d2-43f2-da4f-56628316ea5b"
      },
      "cell_type": "code",
      "source": [
        "linear_classifier = train_linear_classifier_model(\n",
        "    learning_rate=0.000005,\n",
        "    steps=500,\n",
        "    batch_size=20,\n",
        "    training_examples=training_examples,\n",
        "    training_targets=training_targets,\n",
        "    validation_examples=validation_examples,\n",
        "    validation_targets=validation_targets)"
      ],
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Training model...\n",
            "LogLoss (on training data):\n",
            "  period 00 : 0.59\n",
            "  period 01 : 0.58\n",
            "  period 02 : 0.57\n",
            "  period 03 : 0.55\n",
            "  period 04 : 0.54\n",
            "  period 05 : 0.55\n",
            "  period 06 : 0.54\n",
            "  period 07 : 0.54\n",
            "  period 08 : 0.54\n",
            "  period 09 : 0.53\n",
            "Model training finished.\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAjMAAAGACAYAAABY5OOEAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4yLCBo\ndHRwOi8vbWF0cGxvdGxpYi5vcmcvNQv5yAAAIABJREFUeJzs3Xd4VHXaxvHvmUnvCUkIvYQaOghI\nR4oEXRQEIaCIiroqCiq6CqsL6squu9jbrlhfC0YQEVGIAlJEBGkhBEIJEDqk9z7z/oFGWUgYQiYz\nSe7PdXnJzCnzzDwzePs7v3OOYbVarYiIiIjUUCZHFyAiIiJyJRRmREREpEZTmBEREZEaTWFGRERE\najSFGREREanRFGZERESkRlOYEanl2rZty+nTp6tkX8ePHyciIqJK9uUIkydPpn///kRGRjJixAiu\nu+46Pvzww8vez65du5g6deplbxcREcHx48cvezsRqZiLowsQEalOjz32GDfeeCMAycnJTJgwgRYt\nWjBw4ECb99G5c2feffdde5UoIpdJIzMidVRhYSF/+9vfGDFiBCNHjuSf//wnpaWlAGzYsIFBgwYx\ncuRIoqOj6d69+yVHFDIyMpgxY0bZiMfbb79dtuyll15ixIgRjBgxgttuu40zZ85U+Pxv1q1bx6hR\no8577sYbb2T9+vVs2bKFMWPGcN111zFy5EhWrFhx2Z9BSEgIkZGRbNy4EYCDBw9y6623MmLECEaN\nGkVcXBwAmzdvJioqihkzZjBz5kw2b97M8OHDL/k5rlu3juHDhzNy5EjeeeedstfNzc1l2rRpjBw5\nkqFDh/Lkk09SXFx82fWLyDkKMyJ11Icffsjp06f55ptv+PLLL9m6dSvLly+ntLSUJ554gmeeeYYV\nK1Zw5MgR8vPzL7m/F198EX9/f2JiYvj0009ZuHAhW7du5cCBA6xcuZLly5cTExPD8OHD2bRpU7nP\n/1GfPn04ffo0x44dA+DYsWOcPn2avn378vzzzzNr1iy+/fZb3nrrLVatWlWpz6GkpAQ3NzcsFgvT\npk3jxhtvJCYmhrlz53L//fdTUlICwJ49e4iKiuKFF16w+XP861//ypw5c1ixYgUmk6ks5CxduhQ/\nPz9WrFhBTEwMZrOZgwcPVqp+EVGYEamz1q5dy/jx43FxccHDw4NRo0axceNGjhw5QlFREYMGDQLO\nzTOxWCyX3N+6deuYNGkSAAEBAQwfPpyNGzfi5+dHWloaX3/9NZmZmUyePJnRo0eX+/wfubm5cc01\n17BmzRoAVq1axbBhw3BxcaFevXosXbqUxMREmjdvfkHIsMWxY8dYuXIlw4cP59ChQ6SmpjJu3DgA\nevToQVBQEDt27ADAw8ODPn36XPbn2L9/fwDGjBlTts1v+/3xxx+xWCw8/fTTtG/f/rLrF5FzFGZE\n6qi0tDT8/f3LHvv7+5OamkpmZiZ+fn5lz4eGhtq8vz9u5+fnR2pqKvXr1+e1115j5cqVDB48mHvu\nuYdTp06V+/z/GjFixHlh5rrrrgNg3rx5eHp6cscdd3DttdeycuVKm+r897//XTYB+JFHHuGJJ56g\nc+fOZGVlUVBQwMiRI4mMjCQyMpLU1FQyMjLKPp/y3nd5n6OPj895z/9m5MiR3H777bzyyiv06dOH\np59+mqKiIpvqF5ELKcyI1FHBwcFl/6GGc3NegoOD8fHxIS8vr+z5lJSUK9ofwNVXX83bb7/Nxo0b\nadCgAfPnz6/w+T8aMGAACQkJHDlyhCNHjnD11VeXvd5TTz3F+vXr+dvf/sasWbPIzc29ZJ2PPfYY\nK1euJCYmhkWLFpWFo9DQULy9vVm5cmXZPz/++GPZ3JjLfd/+/v7k5OSUPZ+WlnbedlFRUSxatIhv\nv/2W+Ph4li5desnaReTiFGZE6qjBgwezePFiSktLycvL46uvvmLQoEE0b96ckpISNm/eDMDChQsx\nDMOm/UVHRwPn/sP9/fffM3jwYH788UeefvppLBYLXl5etGvXDsMwyn3+f7m5udG/f3/+/e9/M3To\nUMxmM8XFxUyePJmzZ88C0KFDB1xcXDCZKv9XWqNGjQgLCysb4UlLS+ORRx45L9iV974v9jk2bdoU\ns9lc9jkuWbKk7P298cYbLF68GID69evTuHFjmz5jEbk4nZotUgdMnjwZs9lc9vjvf/87kydP5tix\nY1x//fUYhkFkZCQjR47EMAzmzp3LrFmz8PX15Y477sBkMmEYBlarldLSUiIjI8/b/4IFC3jooYeY\nO3cukZGRmEwm7rnnHjp37kxhYSHffPMNI0aMwM3NjaCgIObNm0doaOhFn7+YESNG8OCDD/LBBx8A\n4Orqyrhx47j99tsBMJlMPPnkk3h6evL999+zZs0a/vGPf1zWZ2QYBi+++CJz587l5ZdfxmQycccd\nd+Dl5XXJz7a8z/HZZ59l9uzZuLm5cdNNN5Xt68Ybb2TWrFksWLAAwzDo0qVL2eniInL5DKvVanV0\nESLivPLy8ujWrRtbt27F19fX0eWIiFxAh5lE5AJjx47l22+/BeDbb78lPDxcQUZEnJZGZkTkAlu3\nbuWZZ56hsLAQb29v5s6dS+fOnR1dlojIRSnMiIiISI2mw0wiIiJSoynMiIiISI1W40/NTk7Ottu+\nAwO9SE+v+BoT4hjqjXNSX5yXeuOc1BfbhYSUfxKCRmYq4OJivvRK4hDqjXNSX5yXeuOc1JeqoTAj\nIiIiNZrCjIiIiNRoCjMiIiJSoynMiIiISI2mMCMiIiI1msKMiIiI1GgKMyIiIlKjKcyIiIjUYmvX\nrrZpvVdeeYGTJ0+Uu/yJJx6pqpKqnMKMiIhILXXq1ElWrYqxad0ZM2bSsGGjcpf/858vVlVZVa7G\n385ARERELu7FF59n7954BgzoybXXjuTUqZO8/PKb/OMfz5CcfJb8/HzuvPMe+vUbwAMP3MMjj/yF\nH35YTW5uDkePJnHixHGmT59Jnz79uP76oXzzzWoeeOAeevbszfbtW8nIyOD5518iODiYZ555itOn\nT9GpU2fWrFnFl19+W23vU2FGRESkGny+5iC/JJw97zmz2aC01FrpffZsF8r4Ia3KXT5x4mSWLPmc\nFi3COXr0CG+++Q7p6Wn06nU1I0f+iRMnjvPUU0/Qr9+A87Y7e/YM8+e/ys8//8RXX31Bnz79zlvu\n7e3NK6+8xVtvvcb69Wto2LAxRUWFvP32B2zcuIHPP19Y6fdUGQoz5Th4PJO8UiteZsPRpYiIiFyx\n9u07AODr68fevfEsW7YEwzCRlZV5wbqdO3cFIDQ0lJycnAuWd+nSrWx5ZmYmSUmH6dSpCwB9+vTD\nbK7ee04pzJTjnW/2kJFTxAM3daRji3qOLkdERGq48UNaXTCKEhLiS3JydrW8vqurKwDff7+SrKws\n3njjHbKysrjrrskXrPvHMGK1Xjhy9L/LrVYrJtO55wzDwDCqdyBAE4DLMWlYa6xWK68u3sXOAymO\nLkdEROSymUwmSktLz3suIyODBg0aYjKZWLduDcXFxVf8Oo0aNWbfvj0AbNny8wWvaW8KM+XoHB7M\nnKlXYzIZvPFl3AXHOUVERJxds2Yt2Lcvgdzc3w8VDR48hJ9+2sCMGffh6elJaGgo77+/4Ipep2/f\nAeTm5nLffVOJjd2Bn5//lZZ+WQzrxcaPahB7Dc9lFWXTIDSQHXEpvLwolsLiUqZe356+HRvY5fXk\n8lTn0KzYTn1xXuqNc6otfcnKymT79q0MHjyU5OSzzJhxH59++kWVvkZIiG+5yzRnphwvbX+LIksR\n93a6g0ejuvFi9E7eXb6X4hILg7qWfx6+iIhIXePl5c2aNav49NOPsFotPPhg9V5gTyMz5Vh/fBOf\n71+Km9mVuzpOxqekIfM/20lOfjETh7Vm+FVN7PK6Ypva8n8ztY364rzUG+ekvtiuopEZzZkpx8DG\nfXi4712UWi28tet9jpcm8Pgt3fH3dmPhqgN8+3OSo0sUERERFGYqdHWT7kzveg+eZg8+SVjEzqyN\nPD6pG0F+7ixem8jSDYcuesqaiIiIVB+FmUsID2jOzB73U88jiG+PrOL7s9/w2MQuBPt7sGzjERav\nTVSgERERcSCFGRvU9w7l0aum0dS3MT+f2sqipM94eGIHwoK8WLH5KJ9+fwCLAo2IiIhDKMzYyM/N\nl4e630vHeu3Ym7af9/e/x303t6JRiDertx/n/1YmYLEo0IiISM0zbtwo8vLy+OijD9i9e9d5y/Ly\n8hg3blSF269duxqAb7/9mnXrfrBbneVRmLkM7mY37uk0hf4Ne3Mi5xT/3fM2U25sSLP6vqyPPcW7\n3+yh1GJxdJkiIiKVMnny7XTs2Pmytjl16iSrVsUAcN11oxg06Bp7lFYhXWfmMplNZqLa3kSQRyDL\nDq3kP3sWMOW6W1gWY7Ap/gxFJRb+fEMHXMzKiSIi4lh33nkL8+a9QFhYGKdPn2LWrJmEhISSn59P\nQUEBDz/8GBERHcvWf+65uQwePJSuXbvx17/+haKiorKbTgJ8990KFi+Oxmw20bx5OI8//ldefPF5\n9u6N5/33F2CxWAgICGDs2Am8+eYrxMXFUlJSytix44mMvJ4HHriHnj17s337VjIyMnj++ZcICwu7\n4vepMFMJhmEwovkQAj0C+HjvIt7Z8wEThozDZV0A2/Yl88aSOO4f0xFXl+q9a6iIiDivJQeXs+Ns\n3HnPmU0GpVcwRaFbaCduavWncpcPHHgNGzeuZ+zY8WzYsI6BA68hPLw1AwcOZtu2X/jkkw957rl/\nX7BdTMwKWrYMZ/r0maxe/V3ZyEt+fj4vvPAavr6+TJt2N4mJB5k4cTJLlnzOHXfczbvv/heAnTu3\nc+hQIm+99R75+flMmRLFwIGDAfD29uaVV97irbdeY/36NYwfP6nS7/83Gj64Ar3CujOty1RcTK58\nsj+aTn0yiGgRSGxiKq8u3kVhcfXeaEtEROSPzoWZDQD8+OM6+vcfxLp1q7nvvqm89dZrZGZmXnS7\nI0cO0bFjFwC6detR9ryfnx+zZs3kgQfuISnpMJmZGRfdPiFhD127dgfA09OT5s1bcuzYMQC6dOkG\nQGhoKDk5ORfd/nJpZOYKtQ1qxSM97uPN2Pf45shK+nXujas5nNiDabz0eSwzxnXG010fs4hIXXdT\nqz9dMIpi7ysAt2wZTmpqMmfOnCY7O5sNG9YSHBzKU089S0LCHl5//eWLbme1gslkAJSd3FJcXMyL\nL/6LDz74lHr1gvnLXx4q93UNw+CPJ/mWlBSX7c9s/v2oRVVd2kQjM1WgkU8DHu0xjUY+Ddh4ajNu\nrXfQvV0Q+49l8EL0TvIKrvz26iIiIpXRp09/3n77TQYMGERmZgaNGjUGYN26HygpKbnoNk2bNiMh\nYS8A27dvBSAvLxez2Uy9esGcOXOahIS9lJSUYDKZKC09/0hEu3Yd2LFj26/b5XHixHEaN25qr7eo\nMFNVAj0CeLj7vbQNbMXu1L3kNdrAVR38OXQyi38t3EF2XpGjSxQRkTpo0KBrWLUqhsGDhxIZeT3R\n0Z/w8MPT6NChI6mpqXzzzbILtomMvJ74+DhmzLiPY8eSMAwDf/8AevbszV133cb77y9g0qTJvPrq\nizRr1oJ9+xJ49dUXyrbv0qUrbdu2Y9q0u3n44Wnce+8DeHp62u096kaTFajM8F+JpYRPE75g8+lt\nBHsE0SDrGrbszKVRsDePRnXF38fdTtXWLbo5m3NSX5yXeuOc1BfbVXSjSbtO5pg3bx6xsbEYhsHs\n2bPp3Pn3c9eHDBlCWFhY2bGz+fPnExISwpw5czhw4ACurq7MnTuX8PBwe5ZY5VxMLkxuP55AjwBW\nHllNvvcKevUYypZtufzz0x08FtWVID8PR5cpIiJSa9gtzGzZsoWkpCSio6NJTExk9uzZREdHn7fO\nggUL8Pb2Lnv8/fffk52dzWeffcbRo0d57rnn+O9//2uvEu3GMAxGtRxBkHsAn+3/kgTXFVzV+xq2\nbs7jn59s57GJ3QgJsN9wm4iISF1itzkzmzZtYtiwYQCEh4eTmZl5yVOwjhw5UjZ607RpU06ePHnB\npKKapF+j3vy50xQMw8Qe6yq69ckhJbOAf36yndNpeY4uT0REpFaw28hMSkoKHTp0KHscFBREcnIy\nPj4+Zc/NmTOHEydO0KNHD2bOnEmbNm348MMPmTJlCklJSRw7doz09HSCg4PLfZ3AQC9c7HhxuoqO\n0dnimpBeNKtfn39seJOEgh/pOqgHO9cF86+FO/j7vX1pFuZXRZXWPVfaG7EP9cV5qTfOSX25ctV2\nAZT/nWc8ffp0BgwYgL+/P9OmTSMmJobIyEi2b9/OLbfcQtu2bWnZsuUlz0FPT7ffCEdVTczyJYhH\nut3Pm7Hvsi9vG636tebgT8154vUfmTmhK83C9EW+XJo055zUF+el3jgn9cV2FYU+ux1mCg0NJSUl\npezx2bNnCQkJKXs8evRo6tWrh4uLCwMHDmT//v0APPzww3z22Wc8/fTTZGVlUa9ePXuVWK2CPYOY\n2WMa4f7NOVF8gKZ99pBblMu/F+4g8eTFr8AoIiIil2a3MNOvXz9iYs7dyyE+Pp7Q0NCyQ0zZ2dlM\nnTqVoqJz11755ZdfaN26NQkJCcyaNQuA9evXExERgclUey6F4+3qxYNd76ZbaGeSS04Q2msnBWQz\n/7Od7Dua7ujyREREaiS7HWbq3r07HTp0ICoqCsMwmDNnDkuWLMHX15fhw4czcOBAJkyYgLu7OxER\nEURGRmK1WrFarYwbNw53d3fmz59vr/IcxtXsyp0dJvGluz9rjm0goNtWsnZ34aXPY3lwXGc6NA9y\ndIkiIiI1ii6aVwF7H8v84diPfHHga8yGCwX7u2DNCmXamI50aVX+hGc5R8eZnZP64rzUG+ekvtjO\nIXNm5NKuadKfuzreimFYcW29DXPwMV5fEsfWhLOOLk1ERKTGUJhxsK6hnZje7R68XD0xNYvDtfFB\n/vPVbjbFn3Z0aSIiIjWCwowTaOnfnJk9phHsEYQRdgD38N28s3w362NPOro0ERERp6cw4yTqe4Xw\n6FUP0My3CQQdx6Pddj6I2c3qbccdXZqIiIhTU5hxIr5uPszo/mc6BbcH3xQ8O2zhk7W7WLn5qKNL\nExERcVoKM07G3ezG3R1vY0CjPuCZhWfHzSz6eQfLfjx8yashi4iI1EXVdjsDsZ3ZZGZCm9EEuQfw\n1aEVeERsZllsEUUlFsYOaolhGI4uUURExGkozDgpwzC4tvk1BHoE8NHez3Fvt5WYA4UUlZQycWhr\nBRoREZFf6TCTk+sZ1o0Huk7Fw8Udt/BdrD2xng9jErDokJOIiAigMFMjtAlsxcwe9+Pv5o9rk/1s\nSl/FO8vjKbVYHF2aiIiIwynM1BANfcL4S88HaOAVhkv9Y2wvWsFby2IpKVWgERGRuk1hpgYJcPdn\n5lX30yagFebAZOJN3/Lq0q0UlyjQiIhI3aUwU8N4unjwQNep9AztjsknkwOe3zB/6Y8UFpc6ujQR\nERGHUJipgcwmM1M6TGBE0yGYPPI55hvD80tXk19Y4ujSREREqp3CTA1lGAY3tIokqs1YDJcSTgeu\n4blly8krUKAREZG6RWGmhhvQuDf3dr4ds2EiLegnnl7+OTn5xY4uS0REpNoozNQCnULaM/Oq+3DF\nk5ygncxZ+QHpOQWOLktERKRaKMzUEs39m/Bkn+l4Wv0p8D/A06v/y9nMHEeXJSIiYncKM7VIiFc9\n5g54CH/CKPY9wbMb3uBoaqqjyxIREbErhZlaxsfNm7kDH6C+EY7FK5V/bXmDfadPOLosERERu1GY\nqYXcXNx4cvDdtHDpitU9h1d3/YetRw84uiwRERG7UJippUyGiUcHTqKTx0Cs5kLe3/ceB5JPOros\nERGRKqcwU8vd2/dPdPcYBuZS3tr+MSWlug6NiIjULgozdcDUvsPxKWhGoWsKb2/+2tHliIiIVCmF\nmTrAMAxm9J0Exe7szvuZ2BOHHF2SiIhIlVGYqSMaBgQyLPR6DJOV9+IWUlBS5OiSREREqoTCTB0y\nptvVBBW3ocQtk9c3fuHockRERKqEwkwd83D/KCjy4lDJDjYe2uPockRERK6YwkwdE+Ttw+hmYwD4\n7MAicgrzHFyRiIjIlVGYqYOGt+9CQ2snLK65vLwx2tHliIiIXBGFmTrq4f7jMRX6cYq9fLd3m6PL\nERERqTSFmTrK28OdSW1uxmoxWHb0K1Jzsx1dkoiISKUozNRhfcLb0srlKqyuBbz80yeOLkdERKRS\nFGbquGn9R+NSGESa+RBfxP7o6HJEREQum8JMHefu4spdnSZitZhYc2YFxzNSHV2SiIjIZVGYETo1\nbkYnj/7gUsxrmz/GYrE4uiQRERGbudhz5/PmzSM2NhbDMJg9ezadO3cuWzZkyBDCwsIwm80AzJ8/\nHx8fHx5//HEyMzMpLi5m2rRpDBgwwJ4lyq/u7juSx2P2keN+go+3rua2XsMdXZKIiIhN7BZmtmzZ\nQlJSEtHR0SQmJjJ79myio8+/psmCBQvw9vYue/zxxx/TokULZs6cyZkzZ5gyZQorV660V4nyBy4m\nM/f3uIUXYl9lc+Ya+p7tSKvQBo4uS0RE5JLsdphp06ZNDBs2DIDw8HAyMzPJycmpcJvAwEAyMjIA\nyMrKIjAw0F7lyUWEh4bRx38omEt5c/vHlFhKHV2SiIjIJdktzKSkpJwXRoKCgkhOTj5vnTlz5jBx\n4kTmz5+P1Wrl+uuv5+TJkwwfPpxbb72Vxx9/3F7lSTluueoafIqaUOiWzIJN3zi6HBERkUuy65yZ\nP7Jarec9nj59OgMGDMDf359p06YRExNDYWEhDRs25N133yUhIYHZs2ezZMmSCvcbGOiFi4vZbnWH\nhPjabd/Oau7Ie3g05lniLD9xJLsXPVu2dnRJF1UXe1MTqC/OS71xTurLlbNbmAkNDSUlJaXs8dmz\nZwkJCSl7PHr06LI/Dxw4kP3795Oamkr//v0BaNeuHWfPnqW0tLRskvDFpKfb70aJISG+JCfXvSvj\neuLOsNDrWZX2JS9tfI9/eTyGu6uro8s6T13tjbNTX5yXeuOc1BfbVRT67HaYqV+/fsTExAAQHx9P\naGgoPj4+AGRnZzN16lSKiooA+OWXX2jdujXNmjUjNjYWgBMnTuDt7V1hkBH7GdO1D/VKWlHilsHr\nGyseHRMREXEku43MdO/enQ4dOhAVFYVhGMyZM4clS5bg6+vL8OHDGThwIBMmTMDd3Z2IiAgiIyPJ\ny8tj9uzZ3HrrrZSUlDB37lx7lSc2eKjvJP628d8kumznp0Nd6NuynaNLEhERuYBh/d/JLDWMPYfn\nNPwH3+/dwdJTCzEV+fCPwX/Bx8PD0SUB6o2zUl+cl3rjnNQX2znkMJPUDsPbd6OBtSMWtxxe/jH6\n0huIiIhUM4UZuaSH+o3HVOTLKVM83+3d4ehyREREzqMwI5fk4+HBxDY3Y7UafHV0Kak5GhIVERHn\noTAjNunbsh3h5u7gms/LGxc6uhwREZEyCjNiswf63YRLUQBprgf5YsdPji5HREQEUJiRy+Du6srU\nTpOwWkysOfstJ9LTHF2SiIiIwoxcns6NmtPJsy+4FvHqz59gsVgcXZKIiNRxCjNy2e7ucz3uRaHk\nuB/j460/OLocERGp4xRm5LK5mMxM63ELlJr5OXM1B8+ednRJIiJShynMSKWEhzSgd8AQDHMJb277\nhJLSUkeXJCIidZTCjFTarT2G4lPciEL3MyzYtMLR5YiISB2lMCOVZjKZmN77VihxI67gR2KPJzm6\nJBERqYMUZuSKNAqox9CwkRgmC+/FLaSwpNjRJYmISB2jMCNX7KbO/QgqbUmJexqvb1jq6HJERKSO\nUZiRKvFQ31swij1ILN3KT4n7HF2OiIjUIQozUiXqeftyY7PRGCYrC/cvIreg0NEliYhIHaEwI1Vm\neLvuNCQCi3sWL2343NHliIhIHaEwI1VqRr8JmIq9OWmK4/s9sY4uR0RE6gCFGalSPu6eTGxzMwBL\nj35Jak6OgysSEZHaTmFGqlzfFhGEu3YDtzxe/vEzR5cjIiK1nMKM2MW0vmNxKfYnzW0/S3b87Ohy\nRESkFlOYEbvwcHHjzo4TsVoMVp/9hhPp6Y4uSUREaimFGbGbLo1a0smrD7gW8uqmT7FarY4uSURE\naiGFGbGru3v/CffiYHI8kvhoy1pHlyMiIrWQwozYlYvZhfu63woWMz9nruLg2dOOLklERGoZhRmx\nu9YhDekdMBjDpZg3t35KSWmpo0sSEZFaRGFGqsXkHtfiU9KAQo/TLPgpxtHliIhILaIwI9XCMAwe\n7DUZSl2IK9jArmNHHV2SiIjUEgozUm0aBwQztH4khrmUd+MWUlhc7OiSRESkFlCYkWo1ptMAgizN\nKfFI5fUNyxxdjoiI1AIKM1KtDMPgoT63YJS4k2jZwk8H9zu6JBERqeEUZqTa1fP2Z1SzGzBMVhbu\nX0R2foGjSxIRkRpMYUYcYkTbnjQw2mLxyOSVDV84uhwREanBFGbEYR7qOxFTiRcnzbF8F7/L0eWI\niEgNpTAjDuPj7kVU63EYBnx19EtSsnMcXZKIiNRACjPiUP1adKSlSxdwz+WVH6N1M0oREblsLvbc\n+bx584iNjcUwDGbPnk3nzp3Llg0ZMoSwsDDMZjMA8+fPZ/369Sxb9vvpurt372bHjh32LFGcwLQ+\nN/P4msOkue/jyx2/cFP3Xo4uSUREahC7hZktW7aQlJREdHQ0iYmJzJ49m+jo6PPWWbBgAd7e3mWP\nb775Zm6++eay7VesWGGv8sSJeLi6cUfHKN5OWMDqs8vpndaaRkGBji5LRERqCLsdZtq0aRPDhg0D\nIDw8nMzMTHJybJ8T8cYbb3D//ffbqzxxMl0btaKjV29wK+CVTQux6HCTiIjYyG5hJiUlhcDA3//v\nOigoiOTk5PPWmTNnDhMnTmT+/PnnzZXYtWsXDRo0ICQkxF7liRO6p9cNuJcEket5hI9/XufockRE\npIaw65yZP/rfiZ3Tp09nwIAB+Pv7M23aNGJiYoiMjARg8eLFjBkzxqb9BgZ64eJirvJ6fxMS4mu3\nfcuFZg25h7lr/8XPWasYWXDaj6JFAAAgAElEQVQVEU0alLuueuOc1Bfnpd44J/XlytktzISGhpKS\nklL2+OzZs+eNtIwePbrszwMHDmT//v1lYWbz5s08+eSTNr1OenpeFVV8oZAQX5KTs+22f7lQiEsQ\nvQIGsiVrLf9Y9Q7PRz6I60XCqnrjnNQX56XeOCf1xXYVhT67HWbq168fMTExAMTHxxMaGoqPjw8A\n2dnZTJ06laKiIgB++eUXWrduDcCZM2fw9vbGzc3NXqWJk5vcIxLv0voUep3knY3fObocERFxcnYb\nmenevTsdOnQgKioKwzCYM2cOS5YswdfXl+HDhzNw4EAmTJiAu7s7ERERZaMyycnJBAUF2assqQFM\nhonpPSfzj19eJq5kA7FHO9KlaRNHlyUiIk7KsNbwq5TZc3hOw3+OtXjXD/yQsgKXvBCev/YhPNxc\ny5apN85JfXFe6o1zUl9s55DDTCJXamynwQRam1Lilcwb65c7uhwREXFSCjPitAzD4KGrb8UodSOR\nzfy0/4CjSxIRESekMCNOLdg7gFFNR2GYLHx6YBHZ+YWOLklERJyMwow4vRFtexNmtMbqmcEr65Y4\nuhwREXEyCjNSI8zoMxFTiScnXXYSE7fb0eWIiIgTUZiRGsHPw4cJrcdimKwsO7aEsxma/S8iIuco\nzEiN0b9FZ1q4dgKPHJ786l1KLRZHlyQiIk5AYUZqlAf6jMet1J8Mj338d+0aR5cjIiJOQGFGahQP\nF3ce7DEFLCZ2l/zA2vhER5ckIiIOpjAjNU7LoMbc1PYGDJdiPj+4iGO6eqaISJ2mMCM10oRu19LM\nozWGbxovrV1MfmGJo0sSEREHsTnM5OTkAJCSksLWrVuxaPKlOJBhGDzQ8xbc8aEgaC+vrlxLDb/N\nmIiIVJJNYebZZ59lxYoVZGRkEBUVxUcffcTcuXPtXJpIxbxcvbiv22QMDJLc17F0U4KjSxIREQew\nKczs2bOHm2++mRUrVjBmzBheeeUVkpKS7F2byCW1DmzBiCbDMNwKiTm9nN2HUh1dkoiIVDObwsxv\nw/dr165lyJAhABQVFdmvKpHL8KfWQ2nq1RxzQDL/+Wk5KZn5ji5JRESqkU1hpkWLFlx33XXk5ubS\nvn17li5dir+/v71rE7GJyTBxb7dbcTc8sYTt4eXlGyguKXV0WSIiUk0Mqw2zJktLS9m/fz/h4eG4\nubkRHx9PkyZN8PPzq44aK5Rsx9NyQ0J87bp/qbyL9SY+JYE3d72HpcCL7tzEXSM7YRiGgyqsm/Sb\ncV7qjXNSX2wXEuJb7jKbRmb27t3L6dOncXNz46WXXuJf//oX+/fvr7ICRapCh+B2DGk8EJNHHltz\n1rAu9qSjSxIRkWpgU5j5+9//TosWLdi6dStxcXE89dRTvPrqq/auTeSy3dgqkkZejXAJPsnCbT+Q\neDLT0SWJiIid2RRm3N3dad68OatXr2b8+PG0atUKk0nX2xPn42Jy4Z4ut+JmuGNuGs/r32wiK1eT\n1UVEajObEkl+fj4rVqxg1apV9O/fn4yMDLKysuxdm0ilBHvW49aIsRjmUgrCfuGtr2J1h20RkVrM\npjDzyCOP8PXXX/PII4/g4+PDRx99xO23327n0kQqr0f9rvRt0BOTdzaH2MIXaw85uiQREbETF1tW\nuvrqq+ncuTOHDx9mz5493HXXXXh6etq7NpErcnObGzmYcYSzYUl8t38rLRr60bNdqKPLEhGRKmbT\nyMyqVau49tprmTNnDk8++SQjRoxg3bp19q5N5Iq4md24q9OtuBguuLWM473vtnMiOcfRZYmISBWz\nKcy88847LFu2jMWLF7NkyRIWLVrEW2+9Ze/aRK5YI58GjGtzA4ZLMdamO3jty13kFegO2yIitYlN\nYcbV1ZWgoKCyx/Xr18fV1dVuRYlUpf4Ne9MtpBNmv3TSPON495s9WHSHbRGRWsOmMOPt7c17771H\nQkICCQkJvPPOO3h7e9u7NpEqYRgGk9qNI8g9ANeGicSe3seKn3WjVBGR2sKmMPPcc89x5MgRnnji\nCWbNmsWJEyeYN2+evWsTqTJerp7c0fEWTIYJ91ZxLNmYwO7DusO2iEhtYNPZTPXq1eOZZ54577nE\nxMTzDj2JOLuW/s0Y1XIEXx1agVvLOP7zlSdzb+9FcIDOzBMRqckqfRnfp59+uirrEKkWw5oNol1g\na0wByRT6J/LGl7spKtYdtkVEarJKhxkbbrYt4nRMhonbIqLwdfXBvel+jmYf5+Pv9uv7LCJSg1U6\nzBiGUZV1iFQbf3dfpkREYTUseLeN48f4Y6zbqTtsi4jUVBXOmVm8eHG5y5KTk6u8GJHq0r5eG4Y3\nHcz3R9fiFb6XT753oUmoD+GN/B1dmoiIXKYKw8y2bdvKXda1a9cqL0akOo1qOYIDGYc4wlEICuLN\npe787fae+Hu7Obo0ERG5DIa1hk8WSE7Ottu+Q0J87bp/qbyq6k1Kfhr/2PIyxaUl5O66mrahjZkZ\n1RWzqdJHYOs0/Wacl3rjnNQX24WE+Ja7zKZTsydNmnTBHBmz2UyLFi24//77qV+//pVVKOIgwZ5B\n3NJ+HO/u/hj/DvEkbPdk8dpEJgxp7ejSRETERjaFmb59+3L48GFGjBiByWRi1apVNGjQAH9/f2bN\nmsV777130e3mzZtHbGwshmEwe/ZsOnfuXLZsyJAhhIWFYTabAZg/fz7169dn2bJlvPPOO7i4uDB9\n+nQGDx585e9SpALdQzuT0LA3G09uxr9NIjFbzLRo4Eev9grpIiI1gU1hZtu2bbz//vtlj4cNG8Y9\n99zD22+/zerVqy+6zZYtW0hKSiI6OprExERmz55NdHT0eessWLDgvNsipKen88Ybb/DFF1+Ql5fH\na6+9pjAj1WJc61EcyjzCKQ7hHhLI+98m0CjYm0YhPo4uTURELsGmiQGpqamkpaWVPc7OzubkyZNk\nZWWRnX3xY32bNm1i2LBhAISHh5OZmUlOTk6Fr7Np0yb69OmDj48PoaGhPPvss7a+D5Er4mZ2484O\nt+BqcsGjZTxFRg6vL4nTHbZFRGoAm8LMbbfdxsiRI7npppsYO3Ysw4YN46abbuKHH35gwoQJF90m\nJSWFwMDAssdBQUEXnM49Z84cJk6cyPz587FarRw/fpyCggLuvfdeJk2axKZNm67grYlcnoY+YYxr\nfQNF1gLqd93HmfRc3WFbRKQGsOkw07hx44iMjOTIkSNYLBaaNm1KQEDAZb3Q/540NX36dAYMGIC/\nvz/Tpk0jJiYGgIyMDF5//XVOnjzJbbfdxg8//FDhBfoCA71wcTFfVi2Xo6LZ0+JY9ujN6OBhHM47\nws/HttOoUwg74kys23Wa8cPaVPlr1Vb6zTgv9cY5qS9XzqYwk5uby4cffkhcXByGYdC1a1emTJmC\nh4dHuduEhoaSkpJS9vjs2bOEhISUPR49enTZnwcOHMj+/ftp1KgR3bp1w8XFhaZNm+Lt7U1aWhr1\n6tUr93XS0/NseQuVolPmnJc9ezO2+Y0cSD5MGrvxD/Pl4xV7CfF1o2PL8r+Hco5+M85LvXFO6ovt\nKgp9Nh1meuqpp8jJySEqKorx48eTkpLCk08+WeE2/fr1KxttiY+PJzQ0FB+fc5Mps7OzmTp1KkVF\nRQD88ssvtG7dmv79+/Pzzz9jsVhIT08nLy/vvENVItXBy9WTOzqcuxyBa8tdmN2L+O+yeFIy8h1d\nmoiIXIRNIzMpKSm8+OKLZY+vueYaJk+eXOE23bt3p0OHDkRFRWEYBnPmzGHJkiX4+voyfPhwBg4c\nyIQJE3B3dyciIoLIyEgMw2DEiBGMHz8egCeffBKTLl4mDtDCvxmjWo7gq8QVNL3qEIc2tuX1L+OY\nfWsP3Fztd1hTREQun01hJj8/n/z8fDw9PQHIy8ujsLDwkts9+uij5z1u165d2Z+nTJnClClTLtgm\nKiqKqKgoW8oSsathTQexL+0gCekHaNM9lP3bDT6K2ced17fXjVZFRJyITWFmwoQJjBw5ko4dOwLn\nDhvNmDHDroWJOJrJMDGlQxTztrzESWMbjZoOZuPu07Rs6Mc13Rs7ujwREfmVTcdwxo0bx8KFCxk9\nejRjxozhs88+4+DBg/auTcTh/Nx8mRIRRam1FJrtwNsbPl11gIMnMh1dmoiI/MrmCSkNGjRg2LBh\nDB06lPr167Nr1y571iXiNNoHteHaZteQVphGeK+jWKwW3vwyjszcIkeXJiIiXEaY+V81/GbbIpfl\nTy2upYVfUw7k7qF3v2Iycor4z9LdlFosji5NRKTOq3SY0QRIqUvMJjN3dJiEp4sHe0s20LG9G/uO\nZbDoh0RHlyYiUudVOAF40KBBFw0tVquV9PR0uxUl4ozqeQYxqd043t39MXn1txCW3JvvfjlGiwZ+\n9I7QHbZFRBylwjDz6aefVlcdIjVC99DO7GvYmx9PbuaqnidJXx3K+yv20ijEm8a6w7aIiENUGGYa\nNWpUXXWI1BhjW9/Aocwktqb+wrBrbuSbmELeWBLHU1N64uVh09UORESkCunyuiKXyc3syh0dJuFq\ncmVT9ndc0zuQM+n5vLNcd9gWEXEEhRmRSmjoE8bNbW4gvySfM74/0a6ZPzsPpvDNT0ccXZqISJ2j\nMCNSSX0b9KJ7aGcOZyXRsvtpgvzcWbrhMHGHUh1dmohInaIwI1JJhmEwqd1Y6nkEsvbEekZd64PZ\nbPD2sniSdYdtEZFqozAjcgU8XTy5o8MtGIbBytPLGDesCbkFJbyxJI6i4lJHlyciUicozIhcoRb+\nTbmhZSRZRdkcMK1jQJcGHD2bw//F7NOVskVEqoHCjEgVGNp0IO2D2rAndR+N2p+hRQNfftp9mh92\nnHB0aSIitZ7CjEgVMBkmpkRE4efmy9eHV3LjtUH4eLqyUHfYFhGxO4UZkSri6+bDlIgorFYrXxxe\nxNRRrbBYrefusJ1T6OjyRERqLYUZkSrULqg1w5sNJqUgjW15axg7sCUZOUW89VU8JaW6w7aIiD0o\nzIhUsT+1uJYWfs3YdjaWgGZn6dE2hP26w7aIiN0ozIhUMbPJzB0dJuLp4snn+7/i+sH1aFDPi++3\nHuPnPacdXZ6ISK2jMCNiB/U8g7il3TiKLcV8sv8z/jy6HR5uZj5YkcDxszmOLk9EpFZRmBGxk26h\nnRjQqA8nc0/zU9oPTL2+PUXFFl7/Mo68gmJHlyciUmsozIjY0U2t/kRD7zA2nNiEOfAM113djLPp\n+byzfK/usC0iUkUUZkTsyM3syp0db8HV5MrHCYsZ3CuQiOaB7DyYwnLdYVtEpEoozIjYWQPv+oxv\ncyP5Jfl8sGchd/2pHfX83Plqw2F2JeoO2yIiV0phRqQa9GnQkx6hXTiclcT6M+u4f0wnzGYTC76O\n56zusC0ickUUZkSqgWEYTGx3E8EeQXyX9AOF7me49do25BaU8Pwn29mblO7oEkVEaiyFGZFq4uni\nyZ0db8EwDD7c8xnd2vsxdlBLMnOKmL9wB5+vOUhxia4SLCJyuRRmRKpRM78m3Bg+kqyibP5vTzQj\nr27KrMndCQ30ZOWWozz74VZdh0ZE5DIpzIhUsyFNBhAR1JY9aftYc2wD4Q39mXtHLwZ3bcjx5Bye\n+fAXYrYc1anbIiI2UpgRqWYmw8RtERPwc/Plq8QVHM48irubmdsi2zF9bGe83F2IXnOQFz7bSVpW\ngaPLFRFxegozIg7g6+bDlIgorFYrr+58m/XHf8JitdC1dTDPTO1N11bB7E1K52/vbmHznjOOLldE\nxKkpzIg4SLug1tzRYSIuhpno/Ut5bccCUvLT8PN248GxnZgS2ZYSi4X/Lovn7WXxugWCiEg5zHPn\nzp3r6CKuRF5ekd327e3tbtf9S+XVlt409Amjd1gPzuYnszdtPz+d2oKXiwdN/RrTooE/PduFcvhU\nFnGH0vh5zxmahvoSHODp6LLLVVv6UhupN85JfbGdt7d7ucsUZiqgL5nzqk298XBxp0doV0K8gklI\n28/O5N0kZhymVUBLQv386NcpDJNhsOtgKhvjTlFYVEqbJgGYTYajS79AbepLbaPeOCf1xXYKM5Wk\nL5nzqm29MQyDRj4NLjpK08yvMe2bBdGhZRAJRzOITUxl54EUWjfxx8/bzdGln6e29aU2UW+ck/pi\nu4rCjGG12u/8z3nz5hEbG4thGMyePZvOnTuXLRsyZAhhYWGYzWYA5s+fz5EjR5gxYwatW7cGoE2b\nNjz11FMVvkZycra9yickxNeu+5fKq829sVqt/HJmB4v2f0VeST5tAsK5pf3NBHsGUVBUQvSag6zb\neRIXs4lxg1oyrGcTTIZzjNLU5r7UdOqNc1JfbBcS4lvuMhd7veiWLVtISkoiOjqaxMREZs+eTXR0\n9HnrLFiwAG9v77LHR44coVevXrz66qv2KkvE6RmGQa+w7rQNbMXCfV8Ql7KX57a8yJjw6+jf6Gqm\nRLajS3gw76/Yy2drDhKbmMrU69sT5Ofh6NJFRBzCbmczbdq0iWHDhgEQHh5OZmYmOTm6sqmIrfzd\n/fhzp9uZEhF1wRlPv53C3SW8Xtkp3Fv26hRuEamb7BZmUlJSCAwMLHscFBREcnLyeevMmTOHiRMn\nMn/+fH472nXw4EHuvfdeJk6cyMaNG+1VnkiN8NsozZO9Z9IpuD37MxJ5bsuLrD/+E75eLkwf15nb\nfj2F+z9fxfP21zqFW0TqHrsdZvpf/zs1Z/r06QwYMAB/f3+mTZtGTEwM3bp144EHHmDkyJEcO3aM\n2267je+++w43t/InOQYGeuHiYrZb3RUdoxPHqku9CcGXJxs9yIakLby/43Oi9y9ld8Ye7us5mZuH\nt6Nf18a88Ok2fo4/w8ETWTwysTudWgU7ptY61JeaRr1xTurLlbPbBODXXnuNkJAQoqKiABg6dChf\nffUVPj4+F6z7ySefkJqayvTp0897fty4cbz00ks0adKk3NfRBOC6qS73JrMwq2wujZvZrWwujcUC\ny386wvKfkrBarYzo3ZQxA1ri6lJ918asy31xduqNc1JfbFdR6LPb33L9+vUjJiYGgPj4eEJDQ8uC\nTHZ2NlOnTqWo6NzpaL/88gutW7dm2bJlvPvuuwAkJyeTmppK/fr17VWiSI1U3lyajKIMRg9oyazJ\n3QkJ9GTl5qP8/f+2cjxZc9VEpHaz66nZ8+fPZ+vWrRiGwZw5c9izZw++vr4MHz6cDz/8kKVLl+Lu\n7k5ERARPPfUUubm5PProo2RlZVFcXMwDDzzAoEGDKnwNjczUTerNOReO0lxP/0a9KSq28Nnqg6yP\n/fUU7sHhDLuqsd1P4VZfnJd645zUF9tVNDJj1zBTHRRm6ib15ncVXZdmx4FkPliRQHZeMe2bBdr9\nFG71xXmpN85JfbFdRWFGVwCugK7M6LzUm9+Vf/VgT3o1b02/Tg05lZpL/OE0NsadItjfg0YhF85d\nqwrqi/NSb5yT+mK7iq4ArLtmi9QSF86l+ZLXdiyg2JTDjHGduW1EW4pLz53CveDrePIKShxdsohI\nlai2U7NFxP7Kv3rw9Qzs2pt2zQJZ8HU8m+LPsP9YBnf9KYK2TQMvvWMRESemkRmRWqi8URoXzwJm\n3dqDG/o1Jy27kH99uoNFPxykuMTi6JKdSkFRCSdTcrHU7CmFInWGJgBXQBOznJd6Y7vyzng6fDKb\nBV/v4WxGPk1CfbhnVMQVz6WpqX0pLC7l4PFMEo6mk5CUzuFT2VisVq5qG8LdozpU67V67KWm9qa2\nU19sp7OZKklfMuel3lweq9XKltPbWXRgGfl/OOPJx+zHZ6sPsD72FC5mEzcPDmfoFZzCXVP6UlxS\nysETWSQkpZNwNJ1DJ7MotZz7q9BkGDRv4EtpqZWkM9m0bxbIAzd1wtO9Zh+Vrym9qWvUF9spzFSS\nvmTOS72pnPJGaWIPpvL+twnk5BcT0TyQqddHEOhb/pkD5XHWvhSXWDh86vfwcvBEFiWl5w6tGQY0\nq+9Lu2aBtGsaSOvG/ni6u1BUXMp/vopn58EUWjTw5aGbu+DrVf6tVZyds/amrlNfbKcwU0n6kjkv\n9abyyhulcbX48P63e9mVmIq3hwuTR7SlV/vLuwK3s/SlpNTCkVPZ7D2azr6j6Rw8nknRr/OCDKBJ\nqE9ZeGnTxB8vD9eL7qfUYuGDFQlsjDtNWJAXMyd0pZ6//a7TY0/O0hs5n/piO4WZStKXzHmpN1fu\nYqM0/Rr2Yn3saaJXH6CoxEKfDmHcMrwNXh62HWJxVF9KLRaSTueUzXk5cDyTwuLSsuWNQrxp1/Rc\neGnbNAAfz4uHl4uxWq0s+iGRlVuOEujrzswJXWkY7G2Pt2FX+s04J/XFdgozlaQvmfNSb6pGeaM0\nJfkeLPg6nsOnsqnn58Fdf2pv0ync1dUXi8XKsbM57P31sNGB4xnkF/4eXhrU86Jds0DaNw2kTdMA\n/Krg8NCKzUks+iERH09XHrq5Cy0b+l3xPquTfjPOSX2xncJMJelL5rzUm6p1sVGaq8N68u2mo3z9\n0xGwQuTV5+7C7WIu/8wee/XFYrVyIjm3bM7LvqMZ5BX+ftG/+oGeZYeN2jUNwN/n8uf72GJD7Ek+\nWJmAm4uZB27qRIcWQXZ5HXvQb8Y5qS+2U5ipJH3JnJd6U/XKG6XJSDPzzq+ncDcN9eHuCk7hrqq+\nWK1WTqbkknA0oyy85OQXly0P9vcoG3lp1yywUpOVK2v7/mT+81U8VquVu0dFXPa8IkfRb8Y5qS+2\nU5ipJH3JnJd6Yz8XG6W5KqQHn69JvOQp3JXti9Vq5XRa3rnwknRu0m5W3u/hpZ6f+6/zXQJp1yyA\nYH/PK36fVyIhKZ1Xv9hFYVEpt17bhmu6N3ZoPbbQb8Y5qS+2U5ipJH3JnJd6Y1/ljdIcPVbKByvO\nncLdoXkgd/7PKdy29sVqtZKckV8WXvYeTScz5/eb7QX4uP1+2KhZICH+HhiVvPaNvSSdzualz3eS\nlVfM6P4tGNWvudPV+Ef6zTgn9cV2CjOVpC+Z81JvqsfFRmk6+Xflw5X7y07hvi2yHT3bhQIV9yUl\nM5+EpHOHjRKOppOWVVi2zM/bjXZNA8rCS/1Az2oJBkWlRaQXZpJRkEl6YQYZhZmkF5z7d2FpEcOa\nDqJjcPtytz+TlscL0TtJySxgaI/GTBzWutIXHLQ3/Wack/piO4WZStKXzHmpN9XnYqM0k9qNY3dC\nPtFrDlJUYqFvxzAmDWtDsyaBZX1Jyyr49VTpcwEmJbOgbJ8+nq60axrw62GjQBrW86ry8FJUWkxG\nYQbpBZnnQkphxq/BJaMswOSW5FW4DwOD61oMI7L5UEzGxSc+p2cX8uLnOzmRnMvVEfW58/r2FU6S\ndhT9ZpyT+mI7hZlK0pfMeak31e9iozStPDrxzvK9HDl97hTum4e1YU9iCglH0zmbnl+2rbeHC22a\nBJRN2m0Y4n1FIxi/BZVzIymZpP8aVv44wpJbXH5QcTe7EegeQIC7P4Eev/3bnwD3AAJ//XNyfioL\n4j4irSCdjvXaMyUiCi/Xi8/VyS0o5pVFuzh4IpOOLYOYNroT7m7mSr8/e9BvxjmpL7ZTmKkkfcmc\nl3rjGBcbpYlqM5aN27P4ZtMRfvvbxNPdTJvGAWXzXpqE+mAy2RZezgWVzLJRlfTCzPMO/6QXZlQY\nVNx+DSqB7v4EePj/4c+/BxUPs21zcHKKc3l/96ckpB8gxLMe93SaQkOfsIuuW1hcyptf7ibuUCrh\njfyYMa7LZV2cz970m3FO6ovtFGYqSV8y56XeONbFRmkamyJIzi4i1M+dpvV9MJsuPNRSXFpMRmHW\nuUM+ZeEk87zDQTnFueW+7oVB5VxY+S20BLj74+lStZOFLVYLXx+K4bukH3AzuXJr+5vpUb/rRdct\nKbXw3rd7+Tn+DI2CvXlkQtdqPW28IvrNOCf1xXYKM5WkL5nzUm8c72KjNLdfNY6zqRnnTar94zyV\nCoOKyZVAj4A/HP7xP/8wkHtAlQeVy7HjbBwf7Y2msLSIoU0GcmP4SMymCw8lWaxWPlt1gFXbjlPP\nz4OZUV0JC/JyQMXn02/GOakvtlOYqSR9yZyXeuM8/jhKU57fgkrAeSMpv4eVQHd/PF2q5wymK3E6\n9wxvx/0fZ/KSaRMQzp0db8HX7cILCFqtVpb/dIQvNxzG18uVR8Z3pVlY+X8RVwf9ZpyT+mI7hZlK\n0pfMeak3zsVqtbLtzE6O5CfhbvH4w/yUmhNUbJVfUsBHe6KJTYkn0D2AuztNpplfk4uu+8OOE3wc\nsw93NzMPju1M+2aXvr+Vveg345zUF9spzFSSvmTOS71xTnWlLxarhe+T1vL1oRjMhokJbcfQt2Gv\ni667Ze8ZFny9B8OAP9/QkR5tQ6q52nPqSm9qGvXFdhWFGee7GIKIiJMzGSZGNB/C/V3uxM3sxicJ\ni/k04QuKLSUXrNurfX0eGt8Fs8nEm0vjWB970gEVi9RuCjMiIpUUUa8tj/ecQWOfhmw8uZmXtr9F\nekHGBet1aB7EYxO74e3hygcrEvj25yRq+KC4iFNRmBERuQLBnkHM7HE/Pet3JynrGM//8ioH0hMv\nWK9lQz+euKU7gb7uLF6byOc/HMSiQCNSJRRmRESukJvZjSkRE7i5zY3kluTx6s4FrDm6/oLRl4bB\n3vx1cg8a1PMiZssx3v9mL6UWi4OqFqk9FGZERKqAYRgMbtyPGd3+jI+rN18cXM4HexZSWFp03npB\nfh48cUt3WjTwZePu07yxZDdFxaUOqlqkdlCYERGpQq0CWvB4z+m09G/G1jM7mb/1dZLzUs9bx9fL\njUejuhHRPJCdB1N4MXoneQXFDqpYpOZTmBERqWIB7v7M6PZnBjbqw8nc0zy/9VV2/89FBT3dXZgx\nrgtXtQtl//FMnv90B5k5hQ6qWKRmU5gREbEDF5MLE9qO4db24ym2FPOfXR+w4vAqLNbf58i4upi4\n94YODO7WiGNnc/jHxxqeDg0AABocSURBVNs5m5FfwV5F5GIUZkRE7KhPg6uY2eN+Aj0CWH74O96O\n+5C84t8Di8lkMPnaNozq25yzGfn846NtHDub48CKRWoehRkRETtr6tuYx6+aTrvA1sSl7OXfW1/j\nZM7psuWGYTBmYEsmDmtNZm4R//xkO/uPXXi9GhG5OIUZEZFq4OPmzbSuUxnedDBn81P497bX2XYm\n9rx1hl/VhLtHRVBUXMoL0TvZeTDFQdWK1CwKMyIi1cRkmBjd6jru6jgZA3gv/hOWHFxOqeX3U7P7\ndAjjwbGdMYDXv4hjY9wph9UrUlMozIiIVLNuoZ147KoHCfUKZvXR9bwe+y7ZRb/Pk+kcXo9Ho7rh\n4Wbm3W/28t2Wow6sVsT5KcyIiDhAA+/6/OWq6XQO7sD+9IM8/8urJGUdK1veqrE/T9zaHX8fNz5b\nc5Av1iXqfk4i5bBrmJk3bx4TJkwgKiqKXbt2nbdsyJAhTJo0icmTJzN58mTOnDlTtqygoIBhw4ax\nZMmS/2/vzsOjqs89gH/PrMksWckQIKxBhCwEEpHKElFBKFpRBBMjwZXWgg/FRq/cAMbeemlj1Spg\nwbXFIBIERFBZ7BUUS5DVAClrwAAhKxkyk0xmJrPcP5JMNgg0MJlzku/neXwy68k7vs7k63vO/I43\nyyMi8il/hR9mxabiVwMm4bKtEm8e+Bt2X9zruT8iTIcFMxJgCPbHVzkFWLn1BFwuBhqilhTe2vDe\nvXtRUFCA7Oxs5OfnIz09HdnZ2c0e8/7770Or1bZ67vLlyxEYGOit0oiIREMmyDCp393oo++Fv+et\nxifH16HAdB7TBk2BUqZAtyB/pM9IwJtrf8L3uRdRXVOLXz8QBaVC7uvSiUTDa5OZnJwcjB8/HgAQ\nGRmJyspKVFVde+2E/Px8nD59GuPGjfNWaUREohMVeiteGjEXvXQ98MPFH/HWwRUwWuu+nh2gVeGl\nlHjc2jsIB06W4a3PDqPG5vBxxUTi4bXJTHl5OaKjoz3XQ0JCUFZWBp1O57ktIyMDhYWFSEhIQFpa\nGgRBQGZmJhYtWoSNGzde1+8JDtZA4cX/QwkL03tt23Rj2BtxYl/aLwx6/LnnfLy3/xPsKtiLvxxY\niudHPYMowyAAwOI5Y/CXVfux52gx/vpZLjKeuQNBevX1b5+9EQWny43SCgsKy6qQd+4yfhHTA35q\nr/057hI67N9eywPX5s6di7FjxyIwMBBz5szBtm3bYLVaMWzYMPTu3fu6t2s0Wm52qR5hYXqUlZm9\ntn1qP/ZGnNiXmyNpwMMIV/XA+tOb8T8738ZDA+/DXRFjIAgCnp48GEqZgF2Hi/DCku+RlhSHboH+\n19wme9PxrHYHiissKLpU90/xpWoUVVhQUlEDh7PxtBb9e+jxu2lxCNCqfFit+LUVxr0WZgwGA8rL\nGxd8Ki0tRVhYmOf6gw8+6LmcmJiIkydP4syZMzh//jx27tyJ4uJiqFQqhIeHY9SoUd4qk4hIdARB\nwLjeoxGh74kPjmZh/anNKDCdR8rgaVDLVXjil4Oh0yixZc85/GnVQfz+kTj0CtNde8N007ndbhjN\nNk9oKb5kQVFFNYouWWA0tz5xqFolR0SYFj1CNQgP1eKS2YbvDxVicdYBPJ8Uh+7BGh+8CunzWpgZ\nPXo0li5diuTkZOTl5cFgMHh2MZnNZsybNw/Lly+HSqXCvn37MHHiRMydO9fz/KVLl6JXr14MMkTU\nZQ0M6o/5I36HD46swv6Sn1BUXYJZMTMRpgnF9HEDofdXYe2O0/jzJwcxb3ocInvxixPeUutwosRY\nUxdW6icsRZcsKK6wwGZ3tnp8SIAa0f2CER5aH1xCNOgRqkWQTgVBEDyP69ZNhwB/Jb7c/TMWZx3A\nvOlx6N8joCNfWqfgtTATHx+P6OhoJCcnQxAEZGRkYMOGDdDr9ZgwYQISExORlJQEtVqNqKgoTJo0\nyVulEBFJVpA6EPPif4P1pzbj+8IcZO5fgiejH0V06GBMGtkHOn8l/rHlOP6y5hCeeygWMQNCfV2y\npJktdk9IKbpU7Zm2lFXWoOUyPwq5DOEh/nWBJUSDHqF1gaV7iD/8VNf351UQBExNHIAQvRpZ208g\nc/VB/HZKDOIGdvPCq+u8BLfEV2Hy5j5g7mMWL/ZGnNgX78op2o81JzbA6XLivv4TMLHf3ZAJMhw6\nVYblG/PgdrvxzP1RGBnVvdVz2ZtGTpcL5ZetdceyVDQGlqJL1ai2tv6WWIBG6Zmw9Aip2z0UHqpB\ntwA/yGTCFX7D9Wval0OnyvDuF3lwON2YOelWJMb1vKFtdzZtHTPDMNMGvvnFi70RJ/bF+86ZLuC9\nIx/DaLuM2G5D8HhUMvwV/jhxzogl6w/DanMiZcIg3JMQ0ex5XbE3NTZHqwlL3QG4FjhbLD4oEwSE\nBft7Jizh9VOW8BANdP5Kr9XYsi+nCyuxZN1hVNXUYsqY/nhgdL9mu6W6MoaZduqKb36pYG/EiX3p\nGFX2anyU9wlOGE/D4N8Ns2JnoqcuHOdKzHhzbS5M1XY8MLofpozp7/lD2Fl743K7YTTZWk1Yiios\nqKyyt3q8v1rRZMJSF1h6hGoQFuQPhbzjz/Bzpb4UXarGX9fmorzSisS4HkideCvkMp59iGGmnTrr\nm78zYG/EiX3pOE6XE5vPbMM353ZCJVchdcgjiDcMRanRgtfX/ITySivuju+FlAmD6qYOEuiN2+2G\n3eGCze6EtdYJe/1PW60TNnvjT1O1vf4A3GoUV1hgr3W12la3QD+ENw0s9ROXAK1KVJOOq/WlssqG\nv36Wi3MlVRgaGYrfTomBWtW1V31mmGknKbz5uyr2RpzYl453sPQwVh1bC5vTjvF97sQDAybBbHHg\nzeyfcKGsGrcPMeCZ+6PQIzzwpvXG5XbDXuuErdYFm90Bq90Je60L1loHbHYXbLWO+vucsNod9fc1\nCST1l+ue1xhY7HYn/pM/SCqFrHlgqf/WUPcQDdRKafzhb+s9U2Nz4G8bjyLvbAXXogHDTLvxg1m8\n2BtxYl98o6i6BO8dWYlSSzkGBQ/EU9EpkLnUeHvdYZy+UIno/iGYM30YSsvMjQGiRbCw1bYIF20E\njytNQv5TMkGAWiWHWimDWqWAWimDn1LuuVx3n9zz06/JZZ2/EuGhGoQE+EEmoilLe1zrPeNwuvCP\nLcex+2gxDEH+XXotGoaZduIHs3ixN+LEvvhOjaMGH/97LQ6X5yFYHYRZsakI9++J5RuP4nD+pXZv\nVy4TmoUKT7hQyaFqETKa3yeDn1LR4j4Z/FQKqJVyKOSCqHb3eJPb7YbVaUOlzQST3YTLNhNMdjMq\nbSbIVcCI0NvQS9ejzed/vusMvtxdAL1G2WXXomGYaSd+MIsXeyNO7ItvudwubC/YgS/PbIdcJkfS\noIdwe/cEbPnxHKqsDridrusKJZ7bVHKfHBQrFW63GzUOa6uAUmk31f20mWGqv2x31V51O3JBjl/2\nG497+46DXHb13WM7DhVi1fYTUCpkXXItGoaZduIHs3ixN+LEvohD3qUT+EfealgcNRjTcySmDZqC\nnt2D2ZvrVBdSalDZEE7qA4rJZsZluwkmz21m1LYRUgQI0Kt0CFQHIFClR6A6AAGqgGbXXWo73t/3\nKSrtJvTW90LqkEfanNIcOlmGFZvy4OyCa9EwzLQTP5jFi70RJ/ZFPMprLuG9Ix+jsKoI/QL6IGXY\nFNRUOaCSKaGQKaCSK6GUKaGUKaCsv62z7/Zxu92wOGqaTU9MNnPjJKU+vJjsJtS6Wi+e10CAgACV\nHoHqJgGlPpzUBZUABKj10Ct1bU5agLr3TMHFUqw/tRl7ivdDLsgxuf94TOhz9SlNV12LhmGmnfjB\nLF7sjTixL+Jid9qx+vh67Cs5dM3HChCgkCk84UYpV3qCj1KmhEpeH4JkSs/9nsfKlFDKGy+rZAoo\n5UoomlxuGpyaXpcJN74by+12o9ph8YQTz/TEXrerpyGgVNrNcLQRUmSCDAEqfX1QaRJQ6qcpAWo9\nAlWB0Ku0N6VuoPl75mj5Maw+vt4zpZk5JAk9deFXfF5XXIuGYaad+MEsXuyNOLEv4uN2u5FbdhQm\nXMZlczVqXbWodTlQ66xtvOyqhd1ZC0fDZVctauuv2121be5KuVEKQV4ffJoEpWYBqPl1lUwJp9vl\nORal0m6GyWaCw936ZI8NGkJK01ASpGoIJwGeiYpOefNCyvVq+Z6x1Fqw7tRm/Fh84JpTmq62Fg3D\nTDvxg1m82BtxYl/E60Z643a764OOoz4ANQaihrDjcDlg9wSkKwemhsd7QpKztvk2nc2373K3/RVw\nuSCvm6R4wknDFKVx90+QOgBapabDQ8r1ulpfmk5p+uh7IfUqU5pWa9FMj0OApnOuRcMw0078YBYv\n9kac2BfxkmJvnC6nJ9jYmwQjmVB3zIqYQ8r1aqsv1zulabYWTbA/fv9IHAydcC0ahpl2kuKbv6tg\nb8SJfREv9kacrqcvR8r/jU+Pr0el3XzVKU1XWIumrTAj7UhLRETUycV2i8LCkWkYGZ6Ac+ZCZO57\nG1t//hZOV+NxQoIgYGpiJFIn3oqqmlpkrj6Iw/nlPqy6YzHMEBERiZxGqcHMqCQ8O/QJaJUabD6z\nFa8feAcXq4qbPe6u4b3w3EOxcLuBJeuOYFfuRR9V3LEYZoiIiCSiYUpze3g8zpkvIHPf29jWYkoz\nfFAYXnx0ODR+Cvx9y3Fs+uEsJH5EyTUxzBAREUmIRqnB41HJeHboE9AoNdh0ZiveOPC3ZlOagb0C\n8d8z4tEt0A8bfziLlVtPwOm68ROEihXDDBERkQQ1ndIUmM+3mtL0CNViQWoC+nTX4fvci1i2/ghs\n9quvxyNlDDNEREQSpa2f0vwm9vFmU5qi6hIAQKBOjZdS4hHdPwS5+Zfw2qeHYLLYfVz1zccwQ0RE\nJHFDw6KxcGQaRnSvm9L8ee9b2P7zDjhdTvirFfjdtKEYFROOs0UmLM46gFKjxdcl31QMM0RERJ2A\nVqnBE9GNU5ovzmzBGwfrpjQKuQxP3zcE94/qi1JjDf436wDOFpl8XfJNwzBDRETUiTROaYajwFQ/\npSnYAZfb1WnXomGYISIi6mTqpjSP4texj8Nf6Y8v8hunNJ1xLRqGGSIiok4qLiwai0a+0Dil2fc2\nthfswNCBIZ1qLRqGGSIiok6scUozE/4KP8+URhdk7TRr0TDMEBERdQFxYTFYODINt3UfhgLTefxp\n39s4WrUP/z1juOTXomGYISIi6iJ0Si2ejE7xTGk25n+ND058iCcfjJD0WjQMM0RERF1M0ynNz6Zz\neDN3GWJGGnFHjEGSa9EwzBAREXVBDVOaWbEz4S/3w6YzW3A5fAfuuiNAcmvRMMwQERF1YcPCYrDw\nFw1TmvPY796A2++sQlWNXTJr0TDMEBERdXGeKU1MKvzlfjhS8wP6jD4Mt6pKEmvRKHxdABEREYnD\nMEMsBgYNwNqTG3GgNBfqmFK4Lg7C37e4YDTb8KvR/SAIgq/LbIWTGSIiIvLQqbR4KuYxzIpJhUbh\nD3ePY9DG7sUX+4+Idi0ahhkiIiJqZZghFgtHpiHBEAeXvxF+sbuxu+RfWLo+V3Rr0TDMEBER0RU1\nndJoVf5Q9jmB46ot+NO670S1Fo1Xj5lZvHgxcnNzIQgC0tPTMXToUM99d999N8LDwyGXywEAr7/+\nOgICAjB//nxcunQJNpsNs2fPxl133eXNEomIiOgaGo6lWXPicxzCYZRqt+OVzRfwXxOmIjxE5+vy\nvBdm9u7di4KCAmRnZyM/Px/p6enIzs5u9pj3338fWq3Wc/3rr79GTEwMZs2ahcLCQjz11FMMM0RE\nRCKgU2nxTOwMHCw5jI/z1sEWdhSv/qsQT8YmI6Fff5/W5rUwk5OTg/HjxwMAIiMjUVlZiaqqKuh0\nV09wkydP9lwuKipC9+7dvVUeERERtUN896G4JXgAlv34KS5oT+HD0+/i38Y78diwiZAJvjl6xWth\npry8HNHR0Z7rISEhKCsraxZmMjIyUFhYiISEBKSlpXm+7pWcnIzi4mKsWLHimr8nOFgDhUJ+819A\nvbAwvde2TTeGvREn9kW82BtxkmJfwqDHm1N/j5X/+j98eXYT9hh34Ozuk1g04bcwaEM7vJ4OW2fG\n7XY3uz537lyMHTsWgYGBmDNnDrZt24ZJkyYBANasWYNjx47hxRdfxKZNm9r8TrvRi+eOCAvTo6zM\n7LXtU/uxN+LEvogXeyNOUu/L5EG3I0IVgXcPZqMkqBBvfZON50eneuV3tRX6vDYPMhgMKC9vXAK5\ntLQUYWFhnusPPvggQkNDoVAokJiYiJMnT+Lo0aMoKioCAAwZMgROpxMVFRXeKpGIiIhu0NB+PbHw\nzl8juPhu3KIc6ZMavBZmRo8ejW3btgEA8vLyYDAYPLuYzGYznn76adjtdV/r2rdvH2655Rbs378f\nH330EYC63VQWiwXBwcHeKpGIiIhugh6hWryaMgn33z7IJ7/fa7uZ4uPjER0djeTkZAiCgIyMDGzY\nsAF6vR4TJkxAYmIikpKSoFarERUVhUmTJsFms2HBggVISUmB1WrFyy+/DJmMS+EQERHR1Qnulgez\nSIw39zVKfV9mZ8beiBP7Il7sjTixL9fPJ8fMEBEREXUEhhkiIiKSNIYZIiIikjSGGSIiIpI0hhki\nIiKSNIYZIiIikjSGGSIiIpI0hhkiIiKSNIYZIiIikjSGGSIiIpI0hhkiIiKSNMmfm4mIiIi6Nk5m\niIiISNIYZoiIiEjSGGaIiIhI0hhmiIiISNIYZoiIiEjSGGaIiIhI0hhmrmDx4sVISkpCcnIyDh8+\n7OtyqInXXnsNSUlJePjhh7F9+3Zfl0MtWK1WjB8/Hhs2bPB1KdTEpk2b8MADD2Dq1KnYuXOnr8sh\nANXV1XjuueeQmpqK5ORk7Nq1y9clSZrC1wWIzd69e1FQUIDs7Gzk5+cjPT0d2dnZvi6LAOzZswen\nTp1CdnY2jEYjHnroIdx7772+LouaWL58OQIDA31dBjVhNBrxzjvvYP369bBYLFi6dCnGjRvn67K6\nvM8//xz9+/dHWloaSkpK8Pjjj2Pr1q2+LkuyGGZayMnJwfjx4wEAkZGRqKysRFVVFXQ6nY8roxEj\nRmDo0KEAgICAANTU1MDpdEIul/u4MgKA/Px8nD59mn8oRSYnJwd33HEHdDoddDod/vjHP/q6JAIQ\nHByMEydOAABMJhOCg4N9XJG0cTdTC+Xl5c3+owoJCUFZWZkPK6IGcrkcGo0GALBu3TokJiYyyIhI\nZmYm5s+f7+syqIULFy7AarXi2WefRUpKCnJycnxdEgG47777cPHiRUyYMAEzZszASy+95OuSJI2T\nmWvg2R7E55///CfWrVuHjz76yNelUL2NGzdi2LBh6N27t69LoSu4fPkyli1bhosXL2LmzJnYsWMH\nBEHwdVld2hdffIGePXviww8/xPHjx5Gens5jzW4Aw0wLBoMB5eXlnuulpaUICwvzYUXU1K5du7Bi\nxQp88MEH0Ov1vi6H6u3cuRPnz5/Hzp07UVxcDJVKhfDwcIwaNcrXpXV5oaGhGD58OBQKBfr06QOt\nVouKigqEhob6urQu7eDBgxgzZgwAYPDgwSgtLeVu8xvA3UwtjB49Gtu2bQMA5OXlwWAw8HgZkTCb\nzXjttdfw7rvvIigoyNflUBNvvfUW1q9fj7Vr12L69OmYPXs2g4xIjBkzBnv27IHL5YLRaITFYuHx\nGSLQt29f5ObmAgAKCwuh1WoZZG4AJzMtxMfHIzo6GsnJyRAEARkZGb4uiep9/fXXMBqNmDdvnue2\nzMxM9OzZ04dVEYlb9+7dMXHiRDzyyCMAgIULF0Im4//H+lpSUhLS09MxY8YMOBwOvPLKK74uSdIE\nNw8KISIiIgljPCciIiJJY5ghIiIiSWOYISIiIkljmCEiIiJJY5ghIiIiSWOYIaIOc+HCBcTExCA1\nNdVztuC0tDSYTKbr3kZqaiqcTud1P/7RRx/Fjz/+2J5yiUgiGGaIqEOFhIQgKysLWVlZWLNmDQwG\nA5YvX37dz8/KyuLiYkTUDBfNIyKfGjFiBLKzs3H8+HFkZmbC4XCgtrYWL7/8MqKiopCamorBgwfj\n2LFjWLlyJaKiopCXlwe73Y5FixahuLgYDocDU6ZMQUpKCmpqavD888/DaDSib9++sNlsAICSkhK8\n8MILAACr1YqkpCRMmzbNly+diG4Shhki8hmn04lvvvkGCQkJePHFF/HOO++gT58+rU68p9FosGrV\nqmbPzcrKQkBAAN544w1YrVZMnjwZY8eOxe7du+Hn54fs7GyUlpbinnvuAQBs2bIFAwYMwB/+8AfY\nbDZ89tlnHf56icg7GGaIqENVVFQgNTUVAOByuXDbbbfh4YcfxpIlS7BgwQLP46qqquByuQDUnWak\npdzcXEydOhUA4Ofnh5iYGOTl5eHkyZNISEgAUHfi2AEDBgAAxo4di9WrV2P+/Pm48847kZSU5NXX\nSUQdh2GGiDpUwzEzTZnNZiiVyla3N1Aqla1uEwSh2XW32w1BEOB2u5ude6ghEEVGRuKrr77Cvn37\nsHXrVqxcuRJr1qy50ZdDRCLAA4CJyOf0ej0iIiLw3XffAQDOnj2LZcuWtfmcuLg47Nq1CwBgsViQ\nl5eH6OhoREZG4tChQwCAoqIinD17FgCwefNmHDlyBKNGjUJGRgaKiorgcDi8+KqIqKNwMkNEopCZ\nmYlXX30V7733HhwOB+bPn9/m41NTU7Fo0SI89thjsNvtmD17NiIiIjBlyhR8++23SElJQUREBGJj\nYwEAAwcOREZGBlQqFdxuN2bNmgWFgh+BRJ0Bz5pNREREksbdTERERCRpDDNEREQkaQwzREREJGkM\nM0RERCRpDDNEREQkaQwzREREJGkMM0RERCRpDDNEREQkaf8PzM3hlp6I2VEAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<matplotlib.figure.Figure at 0x7f68d6536e90>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "metadata": {
        "id": "i2e3TlyL57Qs",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "### Solution\n",
        "\n",
        "Click below to see the solution.\n",
        "\n"
      ]
    },
    {
      "metadata": {
        "id": "5YxXd2hn6MuF",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "def train_linear_classifier_model(\n",
        "    learning_rate,\n",
        "    steps,\n",
        "    batch_size,\n",
        "    training_examples,\n",
        "    training_targets,\n",
        "    validation_examples,\n",
        "    validation_targets):\n",
        "  \"\"\"Trains a linear classification model.\n",
        "  \n",
        "  In addition to training, this function also prints training progress information,\n",
        "  as well as a plot of the training and validation loss over time.\n",
        "  \n",
        "  Args:\n",
        "    learning_rate: A `float`, the learning rate.\n",
        "    steps: A non-zero `int`, the total number of training steps. A training step\n",
        "      consists of a forward and backward pass using a single batch.\n",
        "    batch_size: A non-zero `int`, the batch size.\n",
        "    training_examples: A `DataFrame` containing one or more columns from\n",
        "      `california_housing_dataframe` to use as input features for training.\n",
        "    training_targets: A `DataFrame` containing exactly one column from\n",
        "      `california_housing_dataframe` to use as target for training.\n",
        "    validation_examples: A `DataFrame` containing one or more columns from\n",
        "      `california_housing_dataframe` to use as input features for validation.\n",
        "    validation_targets: A `DataFrame` containing exactly one column from\n",
        "      `california_housing_dataframe` to use as target for validation.\n",
        "      \n",
        "  Returns:\n",
        "    A `LinearClassifier` object trained on the training data.\n",
        "  \"\"\"\n",
        "\n",
        "  periods = 10\n",
        "  steps_per_period = steps / periods\n",
        "  \n",
        "  # Create a linear classifier object.\n",
        "  my_optimizer = tf.train.GradientDescentOptimizer(learning_rate=learning_rate)\n",
        "  my_optimizer = tf.contrib.estimator.clip_gradients_by_norm(my_optimizer, 5.0)  \n",
        "  linear_classifier = tf.estimator.LinearClassifier(\n",
        "      feature_columns=construct_feature_columns(training_examples),\n",
        "      optimizer=my_optimizer\n",
        "  )\n",
        "  \n",
        "  # Create input functions.\n",
        "  training_input_fn = lambda: my_input_fn(training_examples, \n",
        "                                          training_targets[\"median_house_value_is_high\"], \n",
        "                                          batch_size=batch_size)\n",
        "  predict_training_input_fn = lambda: my_input_fn(training_examples, \n",
        "                                                  training_targets[\"median_house_value_is_high\"], \n",
        "                                                  num_epochs=1, \n",
        "                                                  shuffle=False)\n",
        "  predict_validation_input_fn = lambda: my_input_fn(validation_examples, \n",
        "                                                    validation_targets[\"median_house_value_is_high\"], \n",
        "                                                    num_epochs=1, \n",
        "                                                    shuffle=False)\n",
        "  \n",
        "  # Train the model, but do so inside a loop so that we can periodically assess\n",
        "  # loss metrics.\n",
        "  print(\"Training model...\")\n",
        "  print(\"LogLoss (on training data):\")\n",
        "  training_log_losses = []\n",
        "  validation_log_losses = []\n",
        "  for period in range (0, periods):\n",
        "    # Train the model, starting from the prior state.\n",
        "    linear_classifier.train(\n",
        "        input_fn=training_input_fn,\n",
        "        steps=steps_per_period\n",
        "    )\n",
        "    # Take a break and compute predictions.    \n",
        "    training_probabilities = linear_classifier.predict(input_fn=predict_training_input_fn)\n",
        "    training_probabilities = np.array([item['probabilities'] for item in training_probabilities])\n",
        "    \n",
        "    validation_probabilities = linear_classifier.predict(input_fn=predict_validation_input_fn)\n",
        "    validation_probabilities = np.array([item['probabilities'] for item in validation_probabilities])\n",
        "    \n",
        "    training_log_loss = metrics.log_loss(training_targets, training_probabilities)\n",
        "    validation_log_loss = metrics.log_loss(validation_targets, validation_probabilities)\n",
        "    # Occasionally print the current loss.\n",
        "    print(\"  period %02d : %0.2f\" % (period, training_log_loss))\n",
        "    # Add the loss metrics from this period to our list.\n",
        "    training_log_losses.append(training_log_loss)\n",
        "    validation_log_losses.append(validation_log_loss)\n",
        "  print(\"Model training finished.\")\n",
        "  \n",
        "  # Output a graph of loss metrics over periods.\n",
        "  plt.ylabel(\"LogLoss\")\n",
        "  plt.xlabel(\"Periods\")\n",
        "  plt.title(\"LogLoss vs. Periods\")\n",
        "  plt.tight_layout()\n",
        "  plt.plot(training_log_losses, label=\"training\")\n",
        "  plt.plot(validation_log_losses, label=\"validation\")\n",
        "  plt.legend()\n",
        "\n",
        "  return linear_classifier"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "UPM_T1FXsTaL",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 0
        },
        "outputId": "9cfa9028-3088-4d55-9bfd-cd1a9396ef6c"
      },
      "cell_type": "code",
      "source": [
        "linear_classifier = train_linear_classifier_model(\n",
        "    learning_rate=0.000005,\n",
        "    steps=500,\n",
        "    batch_size=20,\n",
        "    training_examples=training_examples,\n",
        "    training_targets=training_targets,\n",
        "    validation_examples=validation_examples,\n",
        "    validation_targets=validation_targets)"
      ],
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Training model...\n",
            "LogLoss (on training data):\n",
            "  period 00 : 0.60\n",
            "  period 01 : 0.58\n",
            "  period 02 : 0.57\n",
            "  period 03 : 0.56\n",
            "  period 04 : 0.55\n",
            "  period 05 : 0.55\n",
            "  period 06 : 0.54\n",
            "  period 07 : 0.53\n",
            "  period 08 : 0.53\n",
            "  period 09 : 0.53\n",
            "Model training finished.\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAjMAAAGACAYAAABY5OOEAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<matplotlib.figure.Figure at 0x7f68d415af10>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "metadata": {
        "id": "i-Xo83_aR6s_",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## Task 3: Calculate Accuracy and plot a ROC Curve for the Validation Set\n",
        "\n",
        "A few of the metrics useful for classification are the model [accuracy](https://en.wikipedia.org/wiki/Accuracy_and_precision#In_binary_classification), the [ROC curve](https://en.wikipedia.org/wiki/Receiver_operating_characteristic) and the area under the ROC curve (AUC). We'll examine these metrics.\n",
        "\n",
        "`LinearClassifier.evaluate` calculates useful metrics like accuracy and AUC."
      ]
    },
    {
      "metadata": {
        "id": "DKSQ87VVIYIA",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 53
        },
        "outputId": "c391df63-3c08-407b-bc86-4e7f672a3c14"
      },
      "cell_type": "code",
      "source": [
        "evaluation_metrics = linear_classifier.evaluate(input_fn=predict_validation_input_fn)\n",
        "\n",
        "print(\"AUC on the validation set: %0.2f\" % evaluation_metrics['auc'])\n",
        "print(\"Accuracy on the validation set: %0.2f\" % evaluation_metrics['accuracy'])"
      ],
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "AUC on the validation set: 0.76\n",
            "Accuracy on the validation set: 0.77\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "47xGS2uNIYIE",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "You may use class probabilities, such as those calculated by `LinearClassifier.predict`,\n",
        "and Sklearn's [roc_curve](http://scikit-learn.org/stable/modules/model_evaluation.html#roc-metrics) to\n",
        "obtain the true positive and false positive rates needed to plot a ROC curve."
      ]
    },
    {
      "metadata": {
        "id": "xaU7ttj8IYIF",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 347
        },
        "outputId": "0cc1c835-37f2-44dc-9a56-1a26f7a933a8"
      },
      "cell_type": "code",
      "source": [
        "validation_probabilities = linear_classifier.predict(input_fn=predict_validation_input_fn)\n",
        "# Get just the probabilities for the positive class.\n",
        "validation_probabilities = np.array([item['probabilities'][1] for item in validation_probabilities])\n",
        "\n",
        "false_positive_rate, true_positive_rate, thresholds = metrics.roc_curve(\n",
        "    validation_targets, validation_probabilities)\n",
        "plt.plot(false_positive_rate, true_positive_rate, label=\"our model\")\n",
        "plt.plot([0, 1], [0, 1], label=\"random classifier\")\n",
        "_ = plt.legend(loc=2)"
      ],
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAeEAAAFKCAYAAAAqkecjAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nPZ6iAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<matplotlib.figure.Figure at 0x7f68d4094b10>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "metadata": {
        "id": "PIdhwfgzIYII",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "**See if you can tune the learning settings of the model trained at Task 2 to improve AUC.**\n",
        "\n",
        "Often times, certain metrics improve at the detriment of others, and you'll need to find the settings that achieve a good compromise.\n",
        "\n",
        "**Verify if all metrics improve at the same time.**"
      ]
    },
    {
      "metadata": {
        "id": "XKIqjsqcCaxO",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 671
        },
        "outputId": "ba6f5af9-050f-4472-ae05-8233e8bbdace"
      },
      "cell_type": "code",
      "source": [
        "linear_classifier = train_linear_classifier_model(\n",
        "    learning_rate=0.000003,\n",
        "    steps=30000,\n",
        "    batch_size=500,\n",
        "    training_examples=training_examples,\n",
        "    training_targets=training_targets,\n",
        "    validation_examples=validation_examples,\n",
        "    validation_targets=validation_targets)\n",
        "\n",
        "evaluation_metrics = linear_classifier.evaluate(input_fn=predict_validation_input_fn)\n",
        "\n",
        "print(\"AUC on the validation set: %0.2f\" % evaluation_metrics['auc'])\n",
        "print(\"Accuracy on the validation set: %0.2f\" % evaluation_metrics['accuracy'])"
      ],
      "execution_count": 23,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Training model...\n",
            "LogLoss (on training data):\n",
            "  period 00 : 0.49\n",
            "  period 01 : 0.48\n",
            "  period 02 : 0.47\n",
            "  period 03 : 0.47\n",
            "  period 04 : 0.47\n",
            "  period 05 : 0.47\n",
            "  period 06 : 0.47\n",
            "  period 07 : 0.47\n",
            "  period 08 : 0.47\n",
            "  period 09 : 0.47\n",
            "Model training finished.\n",
            "AUC on the validation set: 0.82\n",
            "Accuracy on the validation set: 0.78\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAjgAAAGACAYAAACgBBhzAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nxH5UAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<matplotlib.figure.Figure at 0x7f68d63534d0>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "metadata": {
        "id": "wCugvl0JdWYL",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "### Solution\n",
        "\n",
        "Click below for a possible solution."
      ]
    },
    {
      "metadata": {
        "id": "VHosS1g2aetf",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "One possible solution that works is to just train for longer, as long as we don't overfit. \n",
        "\n",
        "We can do this by increasing the number the steps, the batch size, or both.\n",
        "\n",
        "All metrics improve at the same time, so our loss metric is a good proxy\n",
        "for both AUC and accuracy.\n",
        "\n",
        "Notice how it takes many, many more iterations just to squeeze a few more \n",
        "units of AUC. This commonly happens. But often even this small gain is worth \n",
        "the costs."
      ]
    },
    {
      "metadata": {
        "id": "dWgTEYMddaA-",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 671
        },
        "outputId": "0e08261a-286f-46ba-b6af-2a6173b547fb"
      },
      "cell_type": "code",
      "source": [
        "linear_classifier = train_linear_classifier_model(\n",
        "    learning_rate=0.000003,\n",
        "    steps=20000,\n",
        "    batch_size=500,\n",
        "    training_examples=training_examples,\n",
        "    training_targets=training_targets,\n",
        "    validation_examples=validation_examples,\n",
        "    validation_targets=validation_targets)\n",
        "\n",
        "evaluation_metrics = linear_classifier.evaluate(input_fn=predict_validation_input_fn)\n",
        "\n",
        "print(\"AUC on the validation set: %0.2f\" % evaluation_metrics['auc'])\n",
        "print(\"Accuracy on the validation set: %0.2f\" % evaluation_metrics['accuracy'])"
      ],
      "execution_count": 17,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Training model...\n",
            "LogLoss (on training data):\n",
            "  period 00 : 0.50\n",
            "  period 01 : 0.48\n",
            "  period 02 : 0.48\n",
            "  period 03 : 0.48\n",
            "  period 04 : 0.47\n",
            "  period 05 : 0.47\n",
            "  period 06 : 0.47\n",
            "  period 07 : 0.47\n",
            "  period 08 : 0.47\n",
            "  period 09 : 0.47\n",
            "Model training finished.\n",
            "AUC on the validation set: 0.81\n",
            "Accuracy on the validation set: 0.78\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAjgAAAGACAYAAACgBBhzAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAABJRU5ErkJggg==\n",
            "text/plain": [
              "<matplotlib.figure.Figure at 0x7f68fa1e5690>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    }
  ]
}