算法实践进阶_task3.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "算法实践进阶-task1.ipynb",
      "version": "0.3.2",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "metadata": {
        "id": "tJ9trV2bV5jz",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "# 一、准备工作\n",
        "## 1. 指定文件目录"
      ]
    },
    {
      "metadata": {
        "id": "UjCxoGpu-FAG",
        "colab_type": "code",
        "outputId": "ea12e553-4bbe-4208-d586-df4db545cd34",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 87
        }
      },
      "cell_type": "code",
      "source": [
        "!mkdir -p drive\n",
        "!google-drive-ocamlfuse drive\n",
        "import os\n",
        "\n",
        "os.chdir(\"drive/Colab Notebooks\") \n",
        "!ls"
      ],
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            " com2018.ipynb\t     drive\t\t      python基础-task3.ipynb\n",
            " data_all.csv\t     homework.txt\t      授权代码.ipynb\n",
            " data.csv\t    'one week.ipynb'\t      网格搜索调参.ipynb\n",
            " data-原始文件.csv   Python基础-task2.ipynb\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "5LSHisEw-1FU",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## 2. 导入需要的包"
      ]
    },
    {
      "metadata": {
        "id": "zOkTDpAO_NrS",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "import pandas as pd\n",
        "from sklearn.model_selection import train_test_split\n",
        "import matplotlib.pyplot as plt\n",
        "from imblearn.over_sampling import SMOTE\n",
        "from sklearn.preprocessing import LabelBinarizer, Imputer\n",
        "from mlxtend.classifier import StackingClassifier\n",
        "from sklearn.model_selection import GridSearchCV\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "bVnooc-dGgXT",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "# 二、读取数据并拆分数据集\n"
      ]
    },
    {
      "metadata": {
        "id": "LM4H4Zj4D53E",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "跑原始文件的时候会有报错：“'utf-8' codec can't decode byte 0xbf in position 0”。  \n",
        "这是因为数据文件不是utf-8编码格式，所以需要将csv文件另存为utf8 csv格式的，就可以解决该问题。"
      ]
    },
    {
      "metadata": {
        "id": "kcSti7FZGfyN",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "data_all = pd.read_csv('./data-原始文件.csv',encoding='gbk')\n",
        "Y = data_all['status']\n",
        "X = data_all.drop('status', axis=1)\n",
        "\n",
        "# 按7:3拆分成training和testing集，随机种子为2018。随机种子的目的是使每次拆分的结果一致\n",
        "X_train, X_test, Y_train, Y_test = train_test_split(X, Y, test_size=0.3,random_state=2018)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "P3veUqhtGZt6",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "# 三、数据预处理"
      ]
    },
    {
      "metadata": {
        "id": "vDjh5DZTHCqd",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## 1. 概览\n",
        "\n"
      ]
    },
    {
      "metadata": {
        "id": "iPc8SkO-TtX7",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "* 该训练集，一共有89个变量，3327个样本。  \n",
        "* 很明显第一列没有列名，其实是ID，所以可以直接去掉。  \n",
        "  \n",
        "* 数据的缺失值情况还是比较多，在本次情况中数据缺失并不是空值，而是存在大量的NA值。  \n",
        "\n",
        "* 从数据类型角度来看，有一些object类型的特殊数据。    \n",
        "  \n",
        "|列名|情况说明|\n",
        "| :--- | :--- |\n",
        "|bank_card_no  | 值里面有很多个???               \n",
        "|reg_preference_for_trad| 值里面有很多个???\n",
        "|source| 值都为xs\n",
        "|id_name|值里面有很多个??? \n",
        "|latest_query_time|日期\n",
        "|loans_latest_time|日期\n",
        "\n",
        "     那些值里是??的字段直接去掉，source只有一个值，也可直接去掉。\n"
      ]
    },
    {
      "metadata": {
        "id": "VtT2dYdiGtqR",
        "colab_type": "code",
        "outputId": "ab10b49a-94e9-4d94-f844-5a95ec242087",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 361
        }
      },
      "cell_type": "code",
      "source": [
        "# 总体情况\n",
        "print(X_train.shape)\n",
        "X_train.describe()"
      ],
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(3327, 89)\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "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>Unnamed: 0</th>\n",
              "      <th>custid</th>\n",
              "      <th>low_volume_percent</th>\n",
              "      <th>middle_volume_percent</th>\n",
              "      <th>take_amount_in_later_12_month_highest</th>\n",
              "      <th>trans_amount_increase_rate_lately</th>\n",
              "      <th>trans_activity_month</th>\n",
              "      <th>trans_activity_day</th>\n",
              "      <th>transd_mcc</th>\n",
              "      <th>trans_days_interval_filter</th>\n",
              "      <th>...</th>\n",
              "      <th>loans_max_limit</th>\n",
              "      <th>loans_avg_limit</th>\n",
              "      <th>consfin_credit_limit</th>\n",
              "      <th>consfin_credibility</th>\n",
              "      <th>consfin_org_count_current</th>\n",
              "      <th>consfin_product_count</th>\n",
              "      <th>consfin_max_limit</th>\n",
              "      <th>consfin_avg_limit</th>\n",
              "      <th>latest_query_day</th>\n",
              "      <th>loans_latest_day</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>count</th>\n",
              "      <td>3327.000000</td>\n",
              "      <td>3.327000e+03</td>\n",
              "      <td>3325.000000</td>\n",
              "      <td>3325.000000</td>\n",
              "      <td>3327.000000</td>\n",
              "      <td>3325.000000</td>\n",
              "      <td>3325.000000</td>\n",
              "      <td>3325.000000</td>\n",
              "      <td>3325.000000</td>\n",
              "      <td>3321.000000</td>\n",
              "      <td>...</td>\n",
              "      <td>3112.000000</td>\n",
              "      <td>3112.000000</td>\n",
              "      <td>3112.000000</td>\n",
              "      <td>3112.000000</td>\n",
              "      <td>3112.000000</td>\n",
              "      <td>3112.000000</td>\n",
              "      <td>3112.000000</td>\n",
              "      <td>3112.000000</td>\n",
              "      <td>3107.000000</td>\n",
              "      <td>3112.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>mean</th>\n",
              "      <td>5998.984070</td>\n",
              "      <td>1.685829e+06</td>\n",
              "      <td>0.022063</td>\n",
              "      <td>0.902433</td>\n",
              "      <td>1946.735798</td>\n",
              "      <td>19.065489</td>\n",
              "      <td>0.806848</td>\n",
              "      <td>0.364448</td>\n",
              "      <td>17.446316</td>\n",
              "      <td>28.980729</td>\n",
              "      <td>...</td>\n",
              "      <td>3386.825193</td>\n",
              "      <td>1818.745501</td>\n",
              "      <td>9186.278920</td>\n",
              "      <td>76.240681</td>\n",
              "      <td>4.716260</td>\n",
              "      <td>5.200835</td>\n",
              "      <td>16082.776350</td>\n",
              "      <td>7993.420308</td>\n",
              "      <td>24.020277</td>\n",
              "      <td>55.572622</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>std</th>\n",
              "      <td>3449.939393</td>\n",
              "      <td>1.039677e+06</td>\n",
              "      <td>0.041444</td>\n",
              "      <td>0.143441</td>\n",
              "      <td>4113.834226</td>\n",
              "      <td>829.746170</td>\n",
              "      <td>0.196462</td>\n",
              "      <td>0.170165</td>\n",
              "      <td>4.486967</td>\n",
              "      <td>22.315568</td>\n",
              "      <td>...</td>\n",
              "      <td>1473.031960</td>\n",
              "      <td>590.992278</td>\n",
              "      <td>7266.313607</td>\n",
              "      <td>13.989810</td>\n",
              "      <td>2.970034</td>\n",
              "      <td>3.398880</td>\n",
              "      <td>14253.689421</td>\n",
              "      <td>5591.836724</td>\n",
              "      <td>37.305071</td>\n",
              "      <td>53.764925</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>min</th>\n",
              "      <td>5.000000</td>\n",
              "      <td>1.140000e+02</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.120000</td>\n",
              "      <td>0.033000</td>\n",
              "      <td>2.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>...</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>-2.000000</td>\n",
              "      <td>-2.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>25%</th>\n",
              "      <td>3102.500000</td>\n",
              "      <td>7.545715e+05</td>\n",
              "      <td>0.010000</td>\n",
              "      <td>0.880000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.600000</td>\n",
              "      <td>0.670000</td>\n",
              "      <td>0.230000</td>\n",
              "      <td>15.000000</td>\n",
              "      <td>16.000000</td>\n",
              "      <td>...</td>\n",
              "      <td>2300.000000</td>\n",
              "      <td>1533.000000</td>\n",
              "      <td>4800.000000</td>\n",
              "      <td>77.000000</td>\n",
              "      <td>2.000000</td>\n",
              "      <td>3.000000</td>\n",
              "      <td>7800.000000</td>\n",
              "      <td>4741.500000</td>\n",
              "      <td>5.000000</td>\n",
              "      <td>10.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>50%</th>\n",
              "      <td>5960.000000</td>\n",
              "      <td>1.623683e+06</td>\n",
              "      <td>0.010000</td>\n",
              "      <td>0.960000</td>\n",
              "      <td>500.000000</td>\n",
              "      <td>0.970000</td>\n",
              "      <td>0.860000</td>\n",
              "      <td>0.350000</td>\n",
              "      <td>17.000000</td>\n",
              "      <td>24.000000</td>\n",
              "      <td>...</td>\n",
              "      <td>3100.000000</td>\n",
              "      <td>1800.000000</td>\n",
              "      <td>7700.000000</td>\n",
              "      <td>79.000000</td>\n",
              "      <td>4.000000</td>\n",
              "      <td>5.000000</td>\n",
              "      <td>13800.000000</td>\n",
              "      <td>7028.000000</td>\n",
              "      <td>14.000000</td>\n",
              "      <td>36.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>75%</th>\n",
              "      <td>8998.000000</td>\n",
              "      <td>2.603724e+06</td>\n",
              "      <td>0.020000</td>\n",
              "      <td>0.990000</td>\n",
              "      <td>2000.000000</td>\n",
              "      <td>1.620000</td>\n",
              "      <td>1.000000</td>\n",
              "      <td>0.477000</td>\n",
              "      <td>20.000000</td>\n",
              "      <td>33.000000</td>\n",
              "      <td>...</td>\n",
              "      <td>4300.000000</td>\n",
              "      <td>2100.000000</td>\n",
              "      <td>11700.000000</td>\n",
              "      <td>80.000000</td>\n",
              "      <td>7.000000</td>\n",
              "      <td>7.000000</td>\n",
              "      <td>20400.000000</td>\n",
              "      <td>9944.000000</td>\n",
              "      <td>23.000000</td>\n",
              "      <td>92.250000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>max</th>\n",
              "      <td>11992.000000</td>\n",
              "      <td>4.004694e+06</td>\n",
              "      <td>1.000000</td>\n",
              "      <td>1.000000</td>\n",
              "      <td>68000.000000</td>\n",
              "      <td>47596.740000</td>\n",
              "      <td>1.000000</td>\n",
              "      <td>0.941000</td>\n",
              "      <td>41.000000</td>\n",
              "      <td>261.000000</td>\n",
              "      <td>...</td>\n",
              "      <td>9800.000000</td>\n",
              "      <td>5600.000000</td>\n",
              "      <td>87100.000000</td>\n",
              "      <td>87.000000</td>\n",
              "      <td>18.000000</td>\n",
              "      <td>20.000000</td>\n",
              "      <td>266400.000000</td>\n",
              "      <td>70200.000000</td>\n",
              "      <td>359.000000</td>\n",
              "      <td>323.000000</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>8 rows × 82 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "         Unnamed: 0        custid  low_volume_percent  middle_volume_percent  \\\n",
              "count   3327.000000  3.327000e+03         3325.000000            3325.000000   \n",
              "mean    5998.984070  1.685829e+06            0.022063               0.902433   \n",
              "std     3449.939393  1.039677e+06            0.041444               0.143441   \n",
              "min        5.000000  1.140000e+02            0.000000               0.000000   \n",
              "25%     3102.500000  7.545715e+05            0.010000               0.880000   \n",
              "50%     5960.000000  1.623683e+06            0.010000               0.960000   \n",
              "75%     8998.000000  2.603724e+06            0.020000               0.990000   \n",
              "max    11992.000000  4.004694e+06            1.000000               1.000000   \n",
              "\n",
              "       take_amount_in_later_12_month_highest  \\\n",
              "count                            3327.000000   \n",
              "mean                             1946.735798   \n",
              "std                              4113.834226   \n",
              "min                                 0.000000   \n",
              "25%                                 0.000000   \n",
              "50%                               500.000000   \n",
              "75%                              2000.000000   \n",
              "max                             68000.000000   \n",
              "\n",
              "       trans_amount_increase_rate_lately  trans_activity_month  \\\n",
              "count                        3325.000000           3325.000000   \n",
              "mean                           19.065489              0.806848   \n",
              "std                           829.746170              0.196462   \n",
              "min                             0.000000              0.120000   \n",
              "25%                             0.600000              0.670000   \n",
              "50%                             0.970000              0.860000   \n",
              "75%                             1.620000              1.000000   \n",
              "max                         47596.740000              1.000000   \n",
              "\n",
              "       trans_activity_day   transd_mcc  trans_days_interval_filter  \\\n",
              "count         3325.000000  3325.000000                 3321.000000   \n",
              "mean             0.364448    17.446316                   28.980729   \n",
              "std              0.170165     4.486967                   22.315568   \n",
              "min              0.033000     2.000000                    0.000000   \n",
              "25%              0.230000    15.000000                   16.000000   \n",
              "50%              0.350000    17.000000                   24.000000   \n",
              "75%              0.477000    20.000000                   33.000000   \n",
              "max              0.941000    41.000000                  261.000000   \n",
              "\n",
              "             ...         loans_max_limit  loans_avg_limit  \\\n",
              "count        ...             3112.000000      3112.000000   \n",
              "mean         ...             3386.825193      1818.745501   \n",
              "std          ...             1473.031960       590.992278   \n",
              "min          ...                0.000000         0.000000   \n",
              "25%          ...             2300.000000      1533.000000   \n",
              "50%          ...             3100.000000      1800.000000   \n",
              "75%          ...             4300.000000      2100.000000   \n",
              "max          ...             9800.000000      5600.000000   \n",
              "\n",
              "       consfin_credit_limit  consfin_credibility  consfin_org_count_current  \\\n",
              "count           3112.000000          3112.000000                3112.000000   \n",
              "mean            9186.278920            76.240681                   4.716260   \n",
              "std             7266.313607            13.989810                   2.970034   \n",
              "min                0.000000             0.000000                   0.000000   \n",
              "25%             4800.000000            77.000000                   2.000000   \n",
              "50%             7700.000000            79.000000                   4.000000   \n",
              "75%            11700.000000            80.000000                   7.000000   \n",
              "max            87100.000000            87.000000                  18.000000   \n",
              "\n",
              "       consfin_product_count  consfin_max_limit  consfin_avg_limit  \\\n",
              "count            3112.000000        3112.000000        3112.000000   \n",
              "mean                5.200835       16082.776350        7993.420308   \n",
              "std                 3.398880       14253.689421        5591.836724   \n",
              "min                 0.000000           0.000000           0.000000   \n",
              "25%                 3.000000        7800.000000        4741.500000   \n",
              "50%                 5.000000       13800.000000        7028.000000   \n",
              "75%                 7.000000       20400.000000        9944.000000   \n",
              "max                20.000000      266400.000000       70200.000000   \n",
              "\n",
              "       latest_query_day  loans_latest_day  \n",
              "count       3107.000000       3112.000000  \n",
              "mean          24.020277         55.572622  \n",
              "std           37.305071         53.764925  \n",
              "min           -2.000000         -2.000000  \n",
              "25%            5.000000         10.000000  \n",
              "50%           14.000000         36.000000  \n",
              "75%           23.000000         92.250000  \n",
              "max          359.000000        323.000000  \n",
              "\n",
              "[8 rows x 82 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 4
        }
      ]
    },
    {
      "metadata": {
        "id": "5-_51rwkouZH",
        "colab_type": "code",
        "outputId": "b8f84f30-4de7-4ad8-ffca-cd52b1e61ace",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1647
        }
      },
      "cell_type": "code",
      "source": [
        "X_train.info()"
      ],
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "<class 'pandas.core.frame.DataFrame'>\n",
            "Int64Index: 3327 entries, 110 to 1274\n",
            "Data columns (total 89 columns):\n",
            "Unnamed: 0                                    3327 non-null int64\n",
            "custid                                        3327 non-null int64\n",
            "trade_no                                      3327 non-null object\n",
            "bank_card_no                                  3327 non-null object\n",
            "low_volume_percent                            3325 non-null float64\n",
            "middle_volume_percent                         3325 non-null float64\n",
            "take_amount_in_later_12_month_highest         3327 non-null int64\n",
            "trans_amount_increase_rate_lately             3325 non-null float64\n",
            "trans_activity_month                          3325 non-null float64\n",
            "trans_activity_day                            3325 non-null float64\n",
            "transd_mcc                                    3325 non-null float64\n",
            "trans_days_interval_filter                    3321 non-null float64\n",
            "trans_days_interval                           3325 non-null float64\n",
            "regional_mobility                             3325 non-null float64\n",
            "student_feature                               1208 non-null float64\n",
            "repayment_capability                          3327 non-null int64\n",
            "is_high_user                                  3327 non-null int64\n",
            "number_of_trans_from_2011                     3325 non-null float64\n",
            "first_transaction_time                        3325 non-null float64\n",
            "historical_trans_amount                       3327 non-null int64\n",
            "historical_trans_day                          3325 non-null float64\n",
            "rank_trad_1_month                             3325 non-null float64\n",
            "trans_amount_3_month                          3327 non-null int64\n",
            "avg_consume_less_12_valid_month               3325 non-null float64\n",
            "abs                                           3327 non-null int64\n",
            "top_trans_count_last_1_month                  3325 non-null float64\n",
            "avg_price_last_12_month                       3327 non-null int64\n",
            "avg_price_top_last_12_valid_month             3249 non-null float64\n",
            "reg_preference_for_trad                       3325 non-null object\n",
            "trans_top_time_last_1_month                   3321 non-null float64\n",
            "trans_top_time_last_6_month                   3321 non-null float64\n",
            "consume_top_time_last_1_month                 3321 non-null float64\n",
            "consume_top_time_last_6_month                 3321 non-null float64\n",
            "cross_consume_count_last_1_month              3037 non-null float64\n",
            "trans_fail_top_count_enum_last_1_month        3314 non-null float64\n",
            "trans_fail_top_count_enum_last_6_month        3314 non-null float64\n",
            "trans_fail_top_count_enum_last_12_month       3314 non-null float64\n",
            "consume_mini_time_last_1_month                3312 non-null float64\n",
            "max_cumulative_consume_later_1_month          3327 non-null int64\n",
            "max_consume_count_later_6_month               3321 non-null float64\n",
            "railway_consume_count_last_12_month           3319 non-null float64\n",
            "pawns_auctions_trusts_consume_last_1_month    3327 non-null int64\n",
            "pawns_auctions_trusts_consume_last_6_month    3327 non-null int64\n",
            "jewelry_consume_count_last_6_month            3319 non-null float64\n",
            "source                                        3327 non-null object\n",
            "first_transaction_day                         3325 non-null float64\n",
            "trans_day_last_12_month                       3325 non-null float64\n",
            "id_name                                       3125 non-null object\n",
            "apply_score                                   3107 non-null float64\n",
            "apply_credibility                             3107 non-null float64\n",
            "query_org_count                               3107 non-null float64\n",
            "query_finance_count                           3107 non-null float64\n",
            "query_cash_count                              3107 non-null float64\n",
            "query_sum_count                               3107 non-null float64\n",
            "latest_query_time                             3107 non-null object\n",
            "latest_one_month_apply                        3107 non-null float64\n",
            "latest_three_month_apply                      3107 non-null float64\n",
            "latest_six_month_apply                        3107 non-null float64\n",
            "loans_score                                   3112 non-null float64\n",
            "loans_credibility_behavior                    3112 non-null float64\n",
            "loans_count                                   3112 non-null float64\n",
            "loans_settle_count                            3112 non-null float64\n",
            "loans_overdue_count                           3112 non-null float64\n",
            "loans_org_count_behavior                      3112 non-null float64\n",
            "consfin_org_count_behavior                    3112 non-null float64\n",
            "loans_cash_count                              3112 non-null float64\n",
            "latest_one_month_loan                         3112 non-null float64\n",
            "latest_three_month_loan                       3112 non-null float64\n",
            "latest_six_month_loan                         3112 non-null float64\n",
            "history_suc_fee                               3112 non-null float64\n",
            "history_fail_fee                              3112 non-null float64\n",
            "latest_one_month_suc                          3112 non-null float64\n",
            "latest_one_month_fail                         3112 non-null float64\n",
            "loans_long_time                               3112 non-null float64\n",
            "loans_latest_time                             3112 non-null object\n",
            "loans_credit_limit                            3112 non-null float64\n",
            "loans_credibility_limit                       3112 non-null float64\n",
            "loans_org_count_current                       3112 non-null float64\n",
            "loans_product_count                           3112 non-null float64\n",
            "loans_max_limit                               3112 non-null float64\n",
            "loans_avg_limit                               3112 non-null float64\n",
            "consfin_credit_limit                          3112 non-null float64\n",
            "consfin_credibility                           3112 non-null float64\n",
            "consfin_org_count_current                     3112 non-null float64\n",
            "consfin_product_count                         3112 non-null float64\n",
            "consfin_max_limit                             3112 non-null float64\n",
            "consfin_avg_limit                             3112 non-null float64\n",
            "latest_query_day                              3107 non-null float64\n",
            "loans_latest_day                              3112 non-null float64\n",
            "dtypes: float64(70), int64(12), object(7)\n",
            "memory usage: 2.3+ MB\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "WdmUEwdSJzbS",
        "colab_type": "code",
        "outputId": "5c3b1f85-d68e-4bad-b463-72b509bccad7",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 139
        }
      },
      "cell_type": "code",
      "source": [
        "# 找出object类型的列\n",
        "for i in X_train.columns:\n",
        "    if X_train[i].dtypes == 'object':\n",
        "        print(i)"
      ],
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "trade_no\n",
            "bank_card_no\n",
            "reg_preference_for_trad\n",
            "source\n",
            "id_name\n",
            "latest_query_time\n",
            "loans_latest_time\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "ImY1dvptN731",
        "colab_type": "code",
        "outputId": "380d5155-23cb-42c5-cb21-a8c66dd1b488",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 104
        }
      },
      "cell_type": "code",
      "source": [
        "# 清除第一列和包含???的列\n",
        "X_train.drop(['Unnamed: 0','id_name', 'bank_card_no', 'reg_preference_for_trad',\\\n",
        "              'latest_query_time', 'loans_latest_time','source'], axis=1, inplace=True)\n"
      ],
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:1: SettingWithCopyWarning: \n",
            "A value is trying to be set on a copy of a slice from a DataFrame\n",
            "\n",
            "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n",
            "  \"\"\"Entry point for launching an IPython kernel.\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "metadata": {
        "id": "G2h-3JXiXPVi",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "**遇到问题**：  \n",
        "剩下两个object是日期格式的字段。单纯的日期格式在很多地方都无法直接处理。  \n",
        "需要对日期做进一步的处理。比如转换成年月日三个单独的数据，或者根据字段具体含义做天数或月数计算，或者直接删除日期字段。  \n",
        "这些都需要视具体业务知识来定。  \n",
        "  \n",
        "在这里偷懒了，为了方便起见，直接删除了。其实转换成距离现在的天数差更好？"
      ]
    },
    {
      "metadata": {
        "id": "zzX9bkR1TP02",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## 2. 区分离散变量和连续变量"
      ]
    },
    {
      "metadata": {
        "id": "pLo5tJZhMQYN",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "两种思路：  \n",
        "一是看变量唯一值个数，先挑选出个数较少的，然后看是否是整数类型。  \n",
        "二是画每个变量的直方图的，稀疏离散的一般是离散变量 。  \n",
        "最后再结合字段含义去理解。  "
      ]
    },
    {
      "metadata": {
        "id": "ZkvGF94nQMfT",
        "colab_type": "code",
        "outputId": "08b9ac1a-7e2c-4045-d0ca-b9dda4248127",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1439
        }
      },
      "cell_type": "code",
      "source": [
        "# 查看每个字段唯一值个数\n",
        "count_value = {}\n",
        "for col in X_train.columns:\n",
        "    count_value[col] = len(X_train[col].unique())\n",
        "sorted(count_value.items(), key=lambda item:item[1])    "
      ],
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[('is_high_user', 2),\n",
              " ('student_feature', 3),\n",
              " ('regional_mobility', 6),\n",
              " ('railway_consume_count_last_12_month', 7),\n",
              " ('jewelry_consume_count_last_6_month', 7),\n",
              " ('top_trans_count_last_1_month', 9),\n",
              " ('avg_consume_less_12_valid_month', 13),\n",
              " ('latest_one_month_loan', 14),\n",
              " ('trans_fail_top_count_enum_last_1_month', 15),\n",
              " ('cross_consume_count_last_1_month', 18),\n",
              " ('query_cash_count', 18),\n",
              " ('consfin_org_count_behavior', 18),\n",
              " ('latest_one_month_suc', 18),\n",
              " ('consfin_org_count_current', 18),\n",
              " ('rank_trad_1_month', 21),\n",
              " ('avg_price_top_last_12_valid_month', 21),\n",
              " ('consfin_product_count', 21),\n",
              " ('consfin_credibility', 24),\n",
              " ('trans_fail_top_count_enum_last_6_month', 25),\n",
              " ('loans_credibility_behavior', 25),\n",
              " ('loans_overdue_count', 25),\n",
              " ('query_finance_count', 26),\n",
              " ('trans_fail_top_count_enum_last_12_month', 27),\n",
              " ('trans_top_time_last_1_month', 29),\n",
              " ('consume_top_time_last_1_month', 29),\n",
              " ('latest_three_month_loan', 29),\n",
              " ('max_consume_count_later_6_month', 30),\n",
              " ('loans_cash_count', 32),\n",
              " ('loans_org_count_current', 32),\n",
              " ('loans_product_count', 32),\n",
              " ('loans_credibility_limit', 33),\n",
              " ('latest_one_month_apply', 35),\n",
              " ('low_volume_percent', 38),\n",
              " ('latest_one_month_fail', 39),\n",
              " ('transd_mcc', 41),\n",
              " ('apply_credibility', 41),\n",
              " ('loans_org_count_behavior', 42),\n",
              " ('query_org_count', 47),\n",
              " ('loans_credit_limit', 53),\n",
              " ('latest_three_month_apply', 55),\n",
              " ('latest_six_month_apply', 62),\n",
              " ('latest_six_month_loan', 66),\n",
              " ('number_of_trans_from_2011', 68),\n",
              " ('query_sum_count', 74),\n",
              " ('middle_volume_percent', 81),\n",
              " ('trans_activity_month', 82),\n",
              " ('loans_max_limit', 89),\n",
              " ('consume_top_time_last_6_month', 91),\n",
              " ('trans_top_time_last_6_month', 92),\n",
              " ('trans_days_interval', 103),\n",
              " ('trans_day_last_12_month', 115),\n",
              " ('loans_settle_count', 117),\n",
              " ('loans_count', 128),\n",
              " ('history_fail_fee', 133),\n",
              " ('trans_days_interval_filter', 137),\n",
              " ('take_amount_in_later_12_month_highest', 145),\n",
              " ('consfin_max_limit', 158),\n",
              " ('history_suc_fee', 163),\n",
              " ('loans_long_time', 183),\n",
              " ('latest_query_day', 187),\n",
              " ('apply_score', 203),\n",
              " ('loans_latest_day', 225),\n",
              " ('loans_score', 237),\n",
              " ('avg_price_last_12_month', 301),\n",
              " ('consfin_credit_limit', 301),\n",
              " ('historical_trans_day', 450),\n",
              " ('trans_activity_day', 487),\n",
              " ('pawns_auctions_trusts_consume_last_1_month', 495),\n",
              " ('trans_amount_increase_rate_lately', 659),\n",
              " ('max_cumulative_consume_later_1_month', 729),\n",
              " ('loans_avg_limit', 844),\n",
              " ('consfin_avg_limit', 1319),\n",
              " ('abs', 1393),\n",
              " ('consume_mini_time_last_1_month', 1404),\n",
              " ('first_transaction_time', 1478),\n",
              " ('first_transaction_day', 1478),\n",
              " ('repayment_capability', 1965),\n",
              " ('pawns_auctions_trusts_consume_last_6_month', 2146),\n",
              " ('trans_amount_3_month', 2652),\n",
              " ('historical_trans_amount', 3200),\n",
              " ('custid', 3327),\n",
              " ('trade_no', 3327)]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 8
        }
      ]
    },
    {
      "metadata": {
        "id": "Wto5R4u4S43J",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "对属性值个数小于10个的进行分析：  \n",
        " ('is_high_user', 2)  \n",
        " ('student_feature', 3)  \n",
        " ('regional_mobility', 6)  \n",
        " ('railway_consume_count_last_12_month', 7)  \n",
        " ('jewelry_consume_count_last_6_month', 7)  \n",
        " ('top_trans_count_last_1_month', 9  \n",
        " \n",
        " 后三个是在过去几个月中的消费/交易计数，统计值，我们认为这是连续数值型属性，不作处理。  \n",
        " \n",
        " is_high_user的取值是0和1，是离散分类变量。  \n",
        " \n",
        " student_feature取值是NA、1和2，暂时当做是离散分类变量。"
      ]
    },
    {
      "metadata": {
        "id": "OyKL2HmAUKyW",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## 3. 变量筛选"
      ]
    },
    {
      "metadata": {
        "id": "1HqDtOH_aFYE",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "一是看变量唯一值的个数，如果只有一种值，肯定不要。  \n",
        "二是缺失值过多的变量也不要，这里判断依据是缺失值超过2/3。  \n",
        "三是采用机器学习方法，信息增益等。"
      ]
    },
    {
      "metadata": {
        "id": "Y8c1vI5CMEsS",
        "colab_type": "code",
        "outputId": "3b48757c-10c0-482b-d558-066221011ddf",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1092
        }
      },
      "cell_type": "code",
      "source": [
        "# 变量取值个数之前已完成\n",
        "# 查看缺失值个数\n",
        "\n",
        "data_isnull = X_train.isnull().sum(axis=0)\n",
        "data_isnull[data_isnull != 0].sort_values(ascending=False)"
      ],
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "student_feature                            2119\n",
              "cross_consume_count_last_1_month            290\n",
              "query_finance_count                         220\n",
              "query_cash_count                            220\n",
              "latest_six_month_apply                      220\n",
              "latest_three_month_apply                    220\n",
              "query_sum_count                             220\n",
              "latest_one_month_apply                      220\n",
              "latest_query_day                            220\n",
              "query_org_count                             220\n",
              "apply_credibility                           220\n",
              "apply_score                                 220\n",
              "latest_one_month_loan                       215\n",
              "loans_score                                 215\n",
              "loans_credibility_behavior                  215\n",
              "loans_count                                 215\n",
              "loans_settle_count                          215\n",
              "loans_overdue_count                         215\n",
              "loans_org_count_behavior                    215\n",
              "consfin_org_count_behavior                  215\n",
              "loans_cash_count                            215\n",
              "loans_latest_day                            215\n",
              "latest_three_month_loan                     215\n",
              "loans_product_count                         215\n",
              "latest_six_month_loan                       215\n",
              "consfin_max_limit                           215\n",
              "consfin_avg_limit                           215\n",
              "consfin_product_count                       215\n",
              "consfin_org_count_current                   215\n",
              "consfin_credibility                         215\n",
              "                                           ... \n",
              "history_suc_fee                             215\n",
              "avg_price_top_last_12_valid_month            78\n",
              "consume_mini_time_last_1_month               15\n",
              "trans_fail_top_count_enum_last_12_month      13\n",
              "trans_fail_top_count_enum_last_1_month       13\n",
              "trans_fail_top_count_enum_last_6_month       13\n",
              "jewelry_consume_count_last_6_month            8\n",
              "railway_consume_count_last_12_month           8\n",
              "max_consume_count_later_6_month               6\n",
              "consume_top_time_last_6_month                 6\n",
              "consume_top_time_last_1_month                 6\n",
              "trans_top_time_last_6_month                   6\n",
              "trans_top_time_last_1_month                   6\n",
              "trans_days_interval_filter                    6\n",
              "regional_mobility                             2\n",
              "middle_volume_percent                         2\n",
              "trans_amount_increase_rate_lately             2\n",
              "trans_activity_month                          2\n",
              "trans_activity_day                            2\n",
              "transd_mcc                                    2\n",
              "trans_days_interval                           2\n",
              "first_transaction_time                        2\n",
              "trans_day_last_12_month                       2\n",
              "number_of_trans_from_2011                     2\n",
              "historical_trans_day                          2\n",
              "rank_trad_1_month                             2\n",
              "avg_consume_less_12_valid_month               2\n",
              "top_trans_count_last_1_month                  2\n",
              "first_transaction_day                         2\n",
              "low_volume_percent                            2\n",
              "Length: 70, dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 9
        }
      ]
    },
    {
      "metadata": {
        "id": "lRgpfPQ5WPZz",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "student_feature 缺失2000多个，直接删除该属性。"
      ]
    },
    {
      "metadata": {
        "id": "pEvQrcMOWW56",
        "colab_type": "code",
        "outputId": "373f3401-fe71-4b32-b4cb-027250eec42d",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 104
        }
      },
      "cell_type": "code",
      "source": [
        "X_train.drop(['student_feature'], axis=1, inplace=True)"
      ],
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:1: SettingWithCopyWarning: \n",
            "A value is trying to be set on a copy of a slice from a DataFrame\n",
            "\n",
            "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n",
            "  \"\"\"Entry point for launching an IPython kernel.\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "metadata": {
        "id": "pMTJzUzMUGkI",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## 4. 缺失值处理"
      ]
    },
    {
      "metadata": {
        "id": "njgLtsaVbaLc",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "处理NA的情况。由于已经把object数据删除了，所以剩下的数值型数据，都采用均值填充。"
      ]
    },
    {
      "metadata": {
        "id": "BxM63m8abZhV",
        "colab_type": "code",
        "outputId": "849be6b8-2e2b-4734-c67b-f05357ea58c6",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1563
        }
      },
      "cell_type": "code",
      "source": [
        "imputer = Imputer(strategy='mean')\n",
        "num = imputer.fit_transform(X_train)\n",
        "X_train = pd.DataFrame(num, columns=X_train.columns)\n",
        "X_train.info()"
      ],
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/sklearn/utils/deprecation.py:58: DeprecationWarning: Class Imputer is deprecated; Imputer was deprecated in version 0.20 and will be removed in 0.22. Import impute.SimpleImputer from sklearn instead.\n",
            "  warnings.warn(msg, category=DeprecationWarning)\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "<class 'pandas.core.frame.DataFrame'>\n",
            "RangeIndex: 3327 entries, 0 to 3326\n",
            "Data columns (total 81 columns):\n",
            "custid                                        3327 non-null float64\n",
            "trade_no                                      3327 non-null float64\n",
            "low_volume_percent                            3327 non-null float64\n",
            "middle_volume_percent                         3327 non-null float64\n",
            "take_amount_in_later_12_month_highest         3327 non-null float64\n",
            "trans_amount_increase_rate_lately             3327 non-null float64\n",
            "trans_activity_month                          3327 non-null float64\n",
            "trans_activity_day                            3327 non-null float64\n",
            "transd_mcc                                    3327 non-null float64\n",
            "trans_days_interval_filter                    3327 non-null float64\n",
            "trans_days_interval                           3327 non-null float64\n",
            "regional_mobility                             3327 non-null float64\n",
            "repayment_capability                          3327 non-null float64\n",
            "is_high_user                                  3327 non-null float64\n",
            "number_of_trans_from_2011                     3327 non-null float64\n",
            "first_transaction_time                        3327 non-null float64\n",
            "historical_trans_amount                       3327 non-null float64\n",
            "historical_trans_day                          3327 non-null float64\n",
            "rank_trad_1_month                             3327 non-null float64\n",
            "trans_amount_3_month                          3327 non-null float64\n",
            "avg_consume_less_12_valid_month               3327 non-null float64\n",
            "abs                                           3327 non-null float64\n",
            "top_trans_count_last_1_month                  3327 non-null float64\n",
            "avg_price_last_12_month                       3327 non-null float64\n",
            "avg_price_top_last_12_valid_month             3327 non-null float64\n",
            "trans_top_time_last_1_month                   3327 non-null float64\n",
            "trans_top_time_last_6_month                   3327 non-null float64\n",
            "consume_top_time_last_1_month                 3327 non-null float64\n",
            "consume_top_time_last_6_month                 3327 non-null float64\n",
            "cross_consume_count_last_1_month              3327 non-null float64\n",
            "trans_fail_top_count_enum_last_1_month        3327 non-null float64\n",
            "trans_fail_top_count_enum_last_6_month        3327 non-null float64\n",
            "trans_fail_top_count_enum_last_12_month       3327 non-null float64\n",
            "consume_mini_time_last_1_month                3327 non-null float64\n",
            "max_cumulative_consume_later_1_month          3327 non-null float64\n",
            "max_consume_count_later_6_month               3327 non-null float64\n",
            "railway_consume_count_last_12_month           3327 non-null float64\n",
            "pawns_auctions_trusts_consume_last_1_month    3327 non-null float64\n",
            "pawns_auctions_trusts_consume_last_6_month    3327 non-null float64\n",
            "jewelry_consume_count_last_6_month            3327 non-null float64\n",
            "first_transaction_day                         3327 non-null float64\n",
            "trans_day_last_12_month                       3327 non-null float64\n",
            "apply_score                                   3327 non-null float64\n",
            "apply_credibility                             3327 non-null float64\n",
            "query_org_count                               3327 non-null float64\n",
            "query_finance_count                           3327 non-null float64\n",
            "query_cash_count                              3327 non-null float64\n",
            "query_sum_count                               3327 non-null float64\n",
            "latest_one_month_apply                        3327 non-null float64\n",
            "latest_three_month_apply                      3327 non-null float64\n",
            "latest_six_month_apply                        3327 non-null float64\n",
            "loans_score                                   3327 non-null float64\n",
            "loans_credibility_behavior                    3327 non-null float64\n",
            "loans_count                                   3327 non-null float64\n",
            "loans_settle_count                            3327 non-null float64\n",
            "loans_overdue_count                           3327 non-null float64\n",
            "loans_org_count_behavior                      3327 non-null float64\n",
            "consfin_org_count_behavior                    3327 non-null float64\n",
            "loans_cash_count                              3327 non-null float64\n",
            "latest_one_month_loan                         3327 non-null float64\n",
            "latest_three_month_loan                       3327 non-null float64\n",
            "latest_six_month_loan                         3327 non-null float64\n",
            "history_suc_fee                               3327 non-null float64\n",
            "history_fail_fee                              3327 non-null float64\n",
            "latest_one_month_suc                          3327 non-null float64\n",
            "latest_one_month_fail                         3327 non-null float64\n",
            "loans_long_time                               3327 non-null float64\n",
            "loans_credit_limit                            3327 non-null float64\n",
            "loans_credibility_limit                       3327 non-null float64\n",
            "loans_org_count_current                       3327 non-null float64\n",
            "loans_product_count                           3327 non-null float64\n",
            "loans_max_limit                               3327 non-null float64\n",
            "loans_avg_limit                               3327 non-null float64\n",
            "consfin_credit_limit                          3327 non-null float64\n",
            "consfin_credibility                           3327 non-null float64\n",
            "consfin_org_count_current                     3327 non-null float64\n",
            "consfin_product_count                         3327 non-null float64\n",
            "consfin_max_limit                             3327 non-null float64\n",
            "consfin_avg_limit                             3327 non-null float64\n",
            "latest_query_day                              3327 non-null float64\n",
            "loans_latest_day                              3327 non-null float64\n",
            "dtypes: float64(81)\n",
            "memory usage: 2.1 MB\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "KjKcjC4iUPJK",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## 5. 数据不平衡处理"
      ]
    },
    {
      "metadata": {
        "id": "oUg5mnquU3E3",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "本次问题存在类别不平衡问题，原始数据类别间比例约为2.5:1。  \n",
        "采用SMOTE方法进行过采样处理。"
      ]
    },
    {
      "metadata": {
        "colab_type": "code",
        "id": "lTIN2KHqjY9c",
        "outputId": "0e69184d-599b-43fc-96fa-3e10e331e9d1",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 677
        }
      },
      "cell_type": "code",
      "source": [
        "# 查看label的分布，存在数据不平衡问题的。直接初步建模也验证了这一现象。\n",
        "plt.hist(Y_train, bins=3)\n",
        "plt.show()\n",
        "plt.hist(Y_test, bins=3)\n",
        "plt.show()"
      ],
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAeoAAAFKCAYAAADScRzUAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4yLCBo\ndHRwOi8vbWF0cGxvdGxpYi5vcmcvOIA7rQAAF1dJREFUeJzt3X9MVff9x/HXhcvdLeulct293Wzt\nuiw2Iw2jEqxDqh0ojZottS2iOO3Ssq0m6OrK4tC006SZopXGNpJoyXRG9oP1xj/4LkZMKyZ1Itu8\nCwObRWuyhqqRe1taFTSiPd8/ljGdeqFwuffN9fn4z3N/fD7nHdNn7z14cDmO4wgAAJiUluwNAACA\n2yPUAAAYRqgBADCMUAMAYBihBgDAMEINAIBh7mRv4FYikQtxfb/s7Ez19vbH9T3vNMxw9Jjh6DHD\n+GCOoxfvGQYCvts+dkd8ona705O9hXGPGY4eMxw9ZhgfzHH0EjnDOyLUAACMV4QaAADDCDUAAIYR\nagAADCPUAAAYRqgBADCMUAMAYBihBgDAsGHdmWzz5s06duyYrl69qhdeeEEHDx7U8ePHNWHCBElS\nZWWlvvvd76q5uVm7d+9WWlqaysvLtXDhQg0MDKimpkZnzpxRenq6Nm7cqMmTJ4/pSQEAkCqGDPXR\no0d18uRJNTU1qbe3V0899ZS+853v6KWXXlJxcfHg8/r7+1VfX69QKKSMjAyVlZWptLRUra2tysrK\nUl1dnQ4fPqy6ujpt3bp1TE8KAIBUMeRX39OmTdMbb7whScrKytKlS5d07dq1m57X0dGh3Nxc+Xw+\neb1e5efnKxwOq62tTaWlpZKkGTNmKBwOx/kUAABIXUN+ok5PT1dmZqYkKRQKadasWUpPT1djY6N2\n7dqliRMn6pVXXlE0GpXf7x98nd/vVyQSueF4WlqaXC6Xrly5Io/Hc9s1s7Mz434f1Vg3PMfwMMPR\nY4ajxwzjgzmOXqJmOOzfnvXOO+8oFApp586d6urq0oQJE5STk6O33npL27Zt09SpU294vuM4t3yf\n2x2/Xrx/q8vztQfj+n5IHTtrShK2ViDgi/tvhrvTMMP4YI6jF+8Zjvq3Z7333nvavn27Ghoa5PP5\nVFhYqJycHElSSUmJTpw4oWAwqGg0Ovianp4eBYNBBYNBRSIRSdLAwIAcx4n5aRoAAPzXkKG+cOGC\nNm/erB07dgz+lPfKlSvV3d0tSWpvb9eUKVOUl5enzs5OnT9/Xn19fQqHwyooKFBRUZH2798vSWpt\nbdX06dPH8HQAAEgtQ371vW/fPvX29mrVqlWDx55++mmtWrVKd911lzIzM7Vx40Z5vV5VV1ersrJS\nLpdLVVVV8vl8mj9/vo4cOaKKigp5PB7V1taO6QkBAJBKXM5wLhonWLyvnXCNGrfDNerxhRnGB3Mc\nPXPXqAEAQHIQagAADCPUAAAYRqgBADCMUAMAYBihBgDAMEINAIBhhBoAAMMINQAAhhFqAAAMI9QA\nABhGqAEAMIxQAwBgGKEGAMAwQg0AgGGEGgAAwwg1AACGEWoAAAwj1AAAGEaoAQAwjFADAGAYoQYA\nwDBCDQCAYYQaAADDCDUAAIYRagAADCPUAAAYRqgBADCMUAMAYBihBgDAMEINAIBhhBoAAMMINQAA\nhhFqAAAMI9QAABhGqAEAMIxQAwBgGKEGAMAwQg0AgGGEGgAAwwg1AACGEWoAAAwj1AAAGEaoAQAw\njFADAGAYoQYAwDBCDQCAYYQaAADD3MN50ubNm3Xs2DFdvXpVL7zwgnJzc7V69Wpdu3ZNgUBAr732\nmjwej5qbm7V7926lpaWpvLxcCxcu1MDAgGpqanTmzBmlp6dr48aNmjx58lifFwAAKWHIUB89elQn\nT55UU1OTent79dRTT6mwsFBLlizRvHnz9PrrrysUCmnBggWqr69XKBRSRkaGysrKVFpaqtbWVmVl\nZamurk6HDx9WXV2dtm7dmohzAwBg3Bvyq+9p06bpjTfekCRlZWXp0qVLam9v1+zZsyVJxcXFamtr\nU0dHh3Jzc+Xz+eT1epWfn69wOKy2tjaVlpZKkmbMmKFwODyGpwMAQGoZ8hN1enq6MjMzJUmhUEiz\nZs3S4cOH5fF4JEkTJ05UJBJRNBqV3+8ffJ3f77/peFpamlwul65cuTL4+lvJzs6U250+qhMDhiMQ\n8KX0eqmIGcYHcxy9RM1wWNeoJemdd95RKBTSzp079cQTTwwedxznls//osev19vbP9xtAaMSiVxI\n2FqBgC+h66UiZhgfzHH04j3DWNEf1k99v/fee9q+fbsaGhrk8/mUmZmpy5cvS5LOnTunYDCoYDCo\naDQ6+Jqenp7B45FIRJI0MDAgx3FifpoGAAD/NWSoL1y4oM2bN2vHjh2aMGGCpH9fa25paZEkHThw\nQDNnzlReXp46Ozt1/vx59fX1KRwOq6CgQEVFRdq/f78kqbW1VdOnTx/D0wEAILUM+dX3vn371Nvb\nq1WrVg0eq62t1csvv6ympiZNmjRJCxYsUEZGhqqrq1VZWSmXy6Wqqir5fD7Nnz9fR44cUUVFhTwe\nj2pra8f0hAAASCUuZzgXjRMs3tdOnq89GNf3Q+rYWVOSsLW4Ljh6zDA+mOPombtGDQAAkoNQAwBg\nGKEGAMAwQg0AgGGEGgAAwwg1AACGEWoAAAwj1AAAGEaoAQAwjFADAGAYoQYAwDBCDQCAYYQaAADD\nCDUAAIYRagAADCPUAAAYRqgBADCMUAMAYBihBgDAMEINAIBhhBoAAMMINQAAhhFqAAAMI9QAABhG\nqAEAMIxQAwBgGKEGAMAwQg0AgGGEGgAAwwg1AACGEWoAAAwj1AAAGEaoAQAwjFADAGAYoQYAwDBC\nDQCAYYQaAADDCDUAAIYRagAADCPUAAAYRqgBADCMUAMAYBihBgDAMEINAIBhhBoAAMMINQAAhhFq\nAAAMI9QAABg2rFCfOHFCc+bMUWNjoySppqZG3//+97Vs2TItW7ZMhw4dkiQ1NzfrmWee0cKFC/X2\n229LkgYGBlRdXa2KigotXbpU3d3dY3MmAACkIPdQT+jv79err76qwsLCG46/9NJLKi4uvuF59fX1\nCoVCysjIUFlZmUpLS9Xa2qqsrCzV1dXp8OHDqqur09atW+N/JgAApKAhP1F7PB41NDQoGAzGfF5H\nR4dyc3Pl8/nk9XqVn5+vcDistrY2lZaWSpJmzJihcDgcn50DAHAHGPITtdvtltt989MaGxu1a9cu\nTZw4Ua+88oqi0aj8fv/g436/X5FI5IbjaWlpcrlcunLlijwez23XzM7OlNudPpLzAb6QQMCX0uul\nImYYH8xx9BI1wyFDfStPPvmkJkyYoJycHL311lvatm2bpk6desNzHMe55Wtvd/x6vb39I9kW8IVF\nIhcStlYg4EvoeqmIGcYHcxy9eM8wVvRH9FPfhYWFysnJkSSVlJToxIkTCgaDikajg8/p6elRMBhU\nMBhUJBKR9O8fLHMcJ+anaQAA8F8jCvXKlSsHf3q7vb1dU6ZMUV5enjo7O3X+/Hn19fUpHA6roKBA\nRUVF2r9/vySptbVV06dPj9/uAQBIcUN+9d3V1aVNmzbp9OnTcrvdamlp0dKlS7Vq1SrdddddyszM\n1MaNG+X1elVdXa3Kykq5XC5VVVXJ5/Np/vz5OnLkiCoqKuTxeFRbW5uI8wIAICW4nOFcNE6weF87\neb72YFzfD6ljZ01JwtbiuuDoMcP4YI6jZ/4aNQAASAxCDQCAYYQaAADDCDUAAIYRagAADCPUAAAY\nRqgBADCMUAMAYBihBgDAMEINAIBhhBoAAMMINQAAhhFqAAAMI9QAABhGqAEAMIxQAwBgGKEGAMAw\nQg0AgGGEGgAAwwg1AACGEWoAAAwj1AAAGEaoAQAwjFADAGAYoQYAwDBCDQCAYYQaAADDCDUAAIYR\nagAADCPUAAAYRqgBADCMUAMAYBihBgDAMEINAIBhhBoAAMMINQAAhhFqAAAMI9QAABhGqAEAMIxQ\nAwBgGKEGAMAwQg0AgGGEGgAAwwg1AACGEWoAAAwj1AAAGEaoAQAwjFADAGDYsEJ94sQJzZkzR42N\njZKks2fPatmyZVqyZIlefPFFXblyRZLU3NysZ555RgsXLtTbb78tSRoYGFB1dbUqKiq0dOlSdXd3\nj9GpAACQeoYMdX9/v1599VUVFhYOHnvzzTe1ZMkS/e53v9PXv/51hUIh9ff3q76+Xr/5zW+0Z88e\n7d69W59++qn+9Kc/KSsrS7///e+1fPly1dXVjekJAQCQSoYMtcfjUUNDg4LB4OCx9vZ2zZ49W5JU\nXFystrY2dXR0KDc3Vz6fT16vV/n5+QqHw2pra1NpaakkacaMGQqHw2N0KgAApJ4hQ+12u+X1em84\ndunSJXk8HknSxIkTFYlEFI1G5ff7B5/j9/tvOp6WliaXyzX4VTkAAIjNPdo3cBwnLsevl52dKbc7\nfVT7AoYjEPCl9HqpiBnGB3McvUTNcEShzszM1OXLl+X1enXu3DkFg0EFg0FFo9HB5/T09OiRRx5R\nMBhUJBLRt771LQ0MDMhxnMFP47fT29s/km0BX1gkciFhawUCvoSul4qYYXwwx9GL9wxjRX9E/zxr\nxowZamlpkSQdOHBAM2fOVF5enjo7O3X+/Hn19fUpHA6roKBARUVF2r9/vySptbVV06dPH8mSAADc\nkYb8RN3V1aVNmzbp9OnTcrvdamlp0ZYtW1RTU6OmpiZNmjRJCxYsUEZGhqqrq1VZWSmXy6Wqqir5\nfD7Nnz9fR44cUUVFhTwej2praxNxXgAApASXM5yLxgkW769knq89GNf3Q+rYWVOSsLX4unH0mGF8\nMMfRM//VNwAASAxCDQCAYYQaAADDCDUAAIYRagAADCPUAAAYRqgBADCMUAMAYBihBgDAMEINAIBh\nhBoAAMMINQAAhhFqAAAMI9QAABg25O+jBoA7Fb8iF7fzf3VPJmwtPlEDAGAYoQYAwDBCDQCAYYQa\nAADDCDUAAIYRagAADCPUAAAYRqgBADCMUAMAYBihBgDAMEINAIBhhBoAAMMINQAAhhFqAAAMI9QA\nABhGqAEAMIxQAwBgGKEGAMAwQg0AgGGEGgAAwwg1AACGEWoAAAwj1AAAGEaoAQAwjFADAGAYoQYA\nwDBCDQCAYYQaAADDCDUAAIYRagAADCPUAAAYRqgBADCMUAMAYJh7JC9qb2/Xiy++qClTpkiSHnro\nIf3oRz/S6tWrde3aNQUCAb322mvyeDxqbm7W7t27lZaWpvLyci1cuDCuJwAAQCobUagl6dFHH9Wb\nb745+Oc1a9ZoyZIlmjdvnl5//XWFQiEtWLBA9fX1CoVCysjIUFlZmUpLSzVhwoS4bB4AgFQXt6++\n29vbNXv2bElScXGx2tra1NHRodzcXPl8Pnm9XuXn5yscDsdrSQAAUt6IP1F/8MEHWr58uT777DOt\nWLFCly5dksfjkSRNnDhRkUhE0WhUfr9/8DV+v1+RSGT0uwYA4A4xolA/+OCDWrFihebNm6fu7m49\n++yzunbt2uDjjuPc8nW3O/6/srMz5Xanj2RrwBcSCPhSer1UxAxhRaL+Lo4o1Pfee6/mz58vSXrg\ngQf0la98RZ2dnbp8+bK8Xq/OnTunYDCoYDCoaDQ6+Lqenh498sgjQ75/b2//SLYFfGGRyIWErRUI\n+BK6XipihrAknn8XY0V/RNeom5ub9etf/1qSFIlE9PHHH+vpp59WS0uLJOnAgQOaOXOm8vLy1NnZ\nqfPnz6uvr0/hcFgFBQUjWRIAgDvSiD5Rl5SU6Oc//7neffddDQwMaP369crJydEvfvELNTU1adKk\nSVqwYIEyMjJUXV2tyspKuVwuVVVVyefjaysAAIZrRKG+++67tX379puO79q166Zjc+fO1dy5c0ey\nDAAAdzzuTAYAgGGEGgAAwwg1AACGEWoAAAwj1AAAGEaoAQAwjFADAGAYoQYAwDBCDQCAYYQaAADD\nCDUAAIYRagAADCPUAAAYRqgBADCMUAMAYBihBgDAMEINAIBhhBoAAMMINQAAhhFqAAAMI9QAABhG\nqAEAMIxQAwBgGKEGAMAwQg0AgGGEGgAAwwg1AACGEWoAAAwj1AAAGEaoAQAwjFADAGAYoQYAwDBC\nDQCAYYQaAADDCDUAAIYRagAADCPUAAAYRqgBADCMUAMAYBihBgDAMEINAIBhhBoAAMMINQAAhhFq\nAAAMI9QAABhGqAEAMIxQAwBgGKEGAMAwdyIW2bBhgzo6OuRyubR27Vp9+9vfTsSyAACMe2Me6r/8\n5S/68MMP1dTUpFOnTmnt2rVqamoa62UBAEgJY/7Vd1tbm+bMmSNJ+uY3v6nPPvtMFy9eHOtlAQBI\nCWMe6mg0quzs7ME/+/1+RSKRsV4WAICUkJBr1NdzHGfI5wQCvriu+X91T8b1/YCRivff7TtRImfI\nfzsQS6L+Lo75J+pgMKhoNDr4556eHgUCgbFeFgCAlDDmoS4qKlJLS4sk6fjx4woGg7r77rvHelkA\nAFLCmH/1nZ+fr4cffliLFy+Wy+XSunXrxnpJAABShssZzkVjAACQFNyZDAAAwwg1AACGpVSoN2zY\noEWLFmnx4sX6xz/+ccNjR44cUVlZmRYtWqT6+vok7XB8iDXHo0ePqry8XIsXL9aaNWv0+eefJ2mX\ntsWa4X/U1dVp2bJlCd7Z+BFrhmfPnlVFRYXKysr0y1/+Mkk7tC/WDH/7299q0aJFqqio0K9+9ask\n7XB8OHHihObMmaPGxsabHktIW5wU0d7e7vzkJz9xHMdxPvjgA6e8vPyGx+fNm+ecOXPGuXbtmlNR\nUeGcPHkyGds0b6g5lpaWOmfPnnUcx3FWrlzpHDp0KOF7tG6oGTqO45w8edJZtGiRs3Tp0kRvb1wY\naoY//elPnQMHDjiO4zjr1693Tp8+nfA9WhdrhhcuXHCKi4udgYEBx3Ec57nnnnP+/ve/J2Wf1vX1\n9TlLly51Xn75ZWfPnj03PZ6ItqTMJ+pYtyrt7u7WPffco6997WtKS0vT448/rra2tmRu16yhbvm6\nd+9effWrX5X077vM9fb2JmWflg3ntrm1tbX62c9+loztjQuxZvj555/r2LFjKikpkSStW7dOkyZN\nStperYo1w4yMDGVkZKi/v19Xr17VpUuXdM899yRzu2Z5PB41NDQoGAze9Fii2pIyoY51q9JIJCK/\n33/Lx3CjoW75+p9/A9/T06M///nPevzxxxO+R+uGmuHevXv16KOP6r777kvG9saFWDP85JNP9OUv\nf1kbN25URUWF6urqkrVN02LN8Etf+pKqqqo0Z84cFRcXKy8vT9/4xjeStVXT3G63vF7vLR9LVFtS\nJtT/y+FfncXFreb48ccfa/ny5Vq3bt0N/yHArV0/w08//VR79+7Vc889l8QdjT/Xz9BxHJ07d07P\nPvusGhsb9f777+vQoUPJ29w4cf0ML168qB07dmj//v1699131dHRoX/+859J3B1iSZlQx7pV6f8+\ndu7cuVt+jYGhb/l68eJF/fjHP9aqVav02GOPJWOL5sWa4dGjR/XJJ5/oBz/4gVasWKHjx49rw4YN\nydqqWbFmmJ2drUmTJumBBx5Qenq6CgsLdfLkyWRt1axYMzx16pQmT54sv98vj8ejgoICdXV1JWur\n41ai2pIyoY51q9L7779fFy9e1EcffaSrV6+qtbVVRUVFydyuWUPd8rW2tlY//OEPNWvWrGRt0bxY\nM5w7d6727dunP/7xj9q2bZsefvhhrV27NpnbNSnWDN1utyZPnqx//etfg4/zte3NYs3wvvvu06lT\np3T58mVJUldXlx588MFkbXXcSlRbUurOZFu2bNHf/va3wVuVvv/++/L5fCotLdVf//pXbdmyRZL0\nxBNPqLKyMsm7tet2c3zsscc0bdo0TZ06dfC53/ve97Ro0aIk7tamWH8X/+Ojjz7SmjVrtGfPniTu\n1K5YM/zwww9VU1Mjx3H00EMPaf369UpLS5nPHXETa4Z/+MMftHfvXqWnp2vq1KlavXp1srdrUldX\nlzZt2qTTp0/L7Xbr3nvvVUlJie6///6EtSWlQg0AQKrhf0EBADCMUAMAYBihBgDAMEINAIBhhBoA\nAMMINQAAhhFqAAAMI9QAABj2/5gt/Ocv7ZqSAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 576x396 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAeoAAAFKCAYAAADScRzUAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nBgDAMEINAIBhhBoAAMMINQAAhv0fLNJ+o6BzlXIAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 576x396 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "metadata": {
        "id": "ELqntdhkU0Sq",
        "colab_type": "code",
        "outputId": "c0106874-e765-4b02-901e-98af90f946ee",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 121
        }
      },
      "cell_type": "code",
      "source": [
        "over_samples = SMOTE(random_state=2018) \n",
        "over_samples_X,over_samples_y = over_samples.fit_sample(X_train, Y_train)\n",
        "\n",
        "# 重抽样前的类别比例\n",
        "print(Y_train.value_counts()/len(Y_train))\n",
        "# 重抽样后的类别比例\n",
        "print(pd.Series(over_samples_y).value_counts()/len(over_samples_y))\n",
        "\n",
        "df = pd.DataFrame(over_samples_X)\n",
        "df.columns = list(X_train[1:0])\n",
        "over_samples_X = df\n",
        "over_samples_y = pd.DataFrame(over_samples_y)"
      ],
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "0    0.749324\n",
            "1    0.250676\n",
            "Name: status, dtype: float64\n",
            "1    0.5\n",
            "0    0.5\n",
            "dtype: float64\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "oZLU5jMvaJ8T",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "# 四、本次遇到问题或后续待处理\n",
        "1. 日期数据的处理\n",
        "2. 如何自动区分连续变量和离散变量？\n",
        "3. 貌似问号是数据编码问题，其实是中文。如果是中文的话，是直接删除还是需要另做处理？怎么做？\n",
        "4. 缺失值填充，连续变量和离散变量应该分开来处理？\n",
        "5. 时间原因，变量筛选尚为完成，后续补充"
      ]
    },
    {
      "metadata": {
        "id": "4vqr2-y1mLye",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "# 特征工程"
      ]
    },
    {
      "metadata": {
        "id": "nExsi8nomUA3",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## IV方法筛选变量\n",
        "通过信息增益的方法来筛选变量。  \n",
        "在此之前，需要对连续变量进行离散化处理。  "
      ]
    },
    {
      "metadata": {
        "id": "3BYuaSt0mKe9",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "from sklearn.tree import DecisionTreeClassifier\n",
        "\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "1Qs6-TwLw5Hp",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## 随机森林筛选变量\n"
      ]
    },
    {
      "metadata": {
        "id": "etgS6IkUw-bw",
        "colab_type": "code",
        "outputId": "0b91d921-c06a-44c2-8cbb-947b677a0486",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1476
        }
      },
      "cell_type": "code",
      "source": [
        "from sklearn.ensemble import RandomForestClassifier\n",
        "import numpy as np\n",
        "feat_labels = data_all.columns[1:]\n",
        "\n",
        "clf = RandomForestClassifier(n_estimators=100, random_state=2018, n_jobs=-1)\n",
        "clf.fit(over_samples_X, over_samples_y)\n",
        "importances = clf.feature_importances_\n",
        "indices = np.argsort(importances)[::-1]\n",
        "for f in range(over_samples_X.shape[1]):\n",
        "    print(\"%2d) %-*s %f\" % (f + 1, 30, feat_labels[indices[f]], importances[indices[f]]))"
      ],
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:6: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples,), for example using ravel().\n",
            "  \n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            " 1) consume_top_time_last_1_month  0.099910\n",
            " 2) loans_org_count_behavior       0.052795\n",
            " 3) consume_top_time_last_6_month  0.041965\n",
            " 4) latest_one_month_apply         0.036360\n",
            " 5) loans_cash_count               0.031163\n",
            " 6) pawns_auctions_trusts_consume_last_6_month 0.029163\n",
            " 7) query_finance_count            0.023588\n",
            " 8) jewelry_consume_count_last_6_month 0.023496\n",
            " 9) consfin_org_count_behavior     0.022374\n",
            "10) avg_consume_less_12_valid_month 0.022077\n",
            "11) cross_consume_count_last_1_month 0.021784\n",
            "12) historical_trans_amount        0.015792\n",
            "13) custid                         0.015296\n",
            "14) loans_max_limit                0.014784\n",
            "15) loans_avg_limit                0.013825\n",
            "16) trans_days_interval_filter     0.013755\n",
            "17) avg_price_top_last_12_valid_month 0.012181\n",
            "18) transd_mcc                     0.011632\n",
            "19) bank_card_no                   0.011481\n",
            "20) latest_one_month_loan          0.010807\n",
            "21) historical_trans_day           0.010402\n",
            "22) max_consume_count_later_6_month 0.010320\n",
            "23) trans_top_time_last_1_month    0.010269\n",
            "24) regional_mobility              0.010202\n",
            "25) take_amount_in_later_12_month_highest 0.010131\n",
            "26) loans_count                    0.010072\n",
            "27) is_high_user                   0.010023\n",
            "28) avg_price_last_12_month        0.009769\n",
            "29) reg_preference_for_trad        0.009550\n",
            "30) top_trans_count_last_1_month   0.009549\n",
            "31) trans_amount_3_month           0.009535\n",
            "32) status                         0.009462\n",
            "33) low_volume_percent             0.009370\n",
            "34) abs                            0.009185\n",
            "35) loans_org_count_current        0.009102\n",
            "36) loans_product_count            0.009071\n",
            "37) loans_overdue_count            0.008972\n",
            "38) trans_fail_top_count_enum_last_6_month 0.008940\n",
            "39) middle_volume_percent          0.008878\n",
            "40) number_of_trans_from_2011      0.008859\n",
            "41) latest_one_month_suc           0.008804\n",
            "42) loans_latest_time              0.008789\n",
            "43) pawns_auctions_trusts_consume_last_1_month 0.008771\n",
            "44) first_transaction_time         0.008684\n",
            "45) latest_three_month_apply       0.008617\n",
            "46) latest_one_month_fail          0.008598\n",
            "47) trans_amount_increase_rate_lately 0.008582\n",
            "48) repayment_capability           0.008530\n",
            "49) trans_activity_month           0.008509\n",
            "50) loans_long_time                0.008474\n",
            "51) latest_query_time              0.008441\n",
            "52) rank_trad_1_month              0.008271\n",
            "53) trans_days_interval            0.008259\n",
            "54) loans_settle_count             0.008249\n",
            "55) trans_activity_day             0.008181\n",
            "56) trans_fail_top_count_enum_last_1_month 0.007950\n",
            "57) loans_credibility_behavior     0.007893\n",
            "58) history_fail_fee               0.007762\n",
            "59) latest_six_month_apply         0.007728\n",
            "60) source                         0.007627\n",
            "61) apply_credibility              0.007511\n",
            "62) loans_score                    0.007466\n",
            "63) trans_fail_top_count_enum_last_12_month 0.007315\n",
            "64) trans_day_last_12_month        0.007299\n",
            "65) loans_credibility_limit        0.007284\n",
            "66) latest_three_month_loan        0.007145\n",
            "67) query_cash_count               0.007037\n",
            "68) id_name                        0.006912\n",
            "69) query_sum_count                0.006880\n",
            "70) history_suc_fee                0.006870\n",
            "71) latest_six_month_loan          0.006756\n",
            "72) query_org_count                0.006555\n",
            "73) loans_credit_limit             0.006193\n",
            "74) max_cumulative_consume_later_1_month 0.006162\n",
            "75) first_transaction_day          0.006111\n",
            "76) apply_score                    0.005857\n",
            "77) trans_top_time_last_6_month    0.005716\n",
            "78) trade_no                       0.001441\n",
            "79) consume_mini_time_last_1_month 0.000574\n",
            "80) student_feature                0.000199\n",
            "81) railway_consume_count_last_12_month 0.000108\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "D0FKNpL4RKVD",
        "colab_type": "code",
        "outputId": "ce1d7bff-e2cb-4f01-e4d5-ffedddc621bb",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "cell_type": "code",
      "source": [
        "type(over_samples_X)"
      ],
      "execution_count": 16,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "pandas.core.frame.DataFrame"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 16
        }
      ]
    },
    {
      "metadata": {
        "id": "sc9tyWMU2sc1",
        "colab_type": "code",
        "outputId": "19b4628b-8696-4d23-8ac6-52b11def45e5",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "cell_type": "code",
      "source": [
        "\n",
        "# 筛选 重要度>0.01的变量\n",
        "threshold = 0.01\n",
        "X_selected = over_samples_X.loc[:, importances > 0.01]\n",
        "X_selected.shape"
      ],
      "execution_count": 17,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(4986, 27)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 17
        }
      ]
    },
    {
      "metadata": {
        "id": "160XvNSUSQD2",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "over_samples_X已经改为了DataFrame格式，所以调用应该用over_samples_X.loc  \n",
        "如果直接跑over_samples_X[:, importances > 0.01]，会提示报错 `unhashable type: 'slice'``\n"
      ]
    },
    {
      "metadata": {
        "id": "Vd9PiSGu2b4B",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## 各种模型\n"
      ]
    },
    {
      "metadata": {
        "id": "R36VlfLb3tbC",
        "colab_type": "code",
        "outputId": "57196ac7-9581-4f81-ea6f-9ded019d9522",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 3679
        }
      },
      "cell_type": "code",
      "source": [
        "from __future__ import print_function\n",
        "import pandas as pd\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "# 引入要用到的评价函数\n",
        "from sklearn.metrics import precision_score,roc_auc_score\n",
        "from sklearn.metrics import accuracy_score\n",
        "from sklearn.metrics import precision_score\n",
        "from sklearn.metrics import recall_score\n",
        "from sklearn.metrics import f1_score\n",
        "from sklearn.metrics import roc_curve, auc\n",
        "\n",
        "# 引入用到的分类算法\n",
        "from sklearn.ensemble import RandomForestClassifier\n",
        "from sklearn.ensemble import GradientBoostingClassifier\n",
        "from sklearn.model_selection import train_test_split\n",
        "from sklearn.preprocessing import scale\n",
        "from sklearn.model_selection import train_test_split\n",
        "from sklearn.linear_model import LogisticRegression\n",
        "from sklearn.tree import DecisionTreeClassifier\n",
        "from sklearn.svm import LinearSVC\n",
        "from xgboost import XGBClassifier\n",
        "from lightgbm import LGBMClassifier\n",
        "\n",
        "# 定义一个包含多个评价指标的函数\n",
        "def multi_score(model,x_test,y_test):\n",
        "    test_predict=model.predict(x_test)\n",
        "    if hasattr(model,'decision_function'):\n",
        "        pre_test=model.decision_function(x_test)\n",
        "    else:\n",
        "        pre_test=model.predict_proba(x_test)[:,1]\n",
        "    print(\"准确率\",accuracy_score(y_test,test_predict))\n",
        "    print(\"精确率\",precision_score(y_test,test_predict))\n",
        "    print(\"召回率\",recall_score(y_test,test_predict))\n",
        "    print(\"F1-score\",f1_score(y_test,test_predict))\n",
        "    print(\"AUC\",roc_auc_score(y_test,pre_test))\n",
        "    false_positive_rate,true_positive_rate,thresholds=roc_curve(y_test,pre_test)\n",
        "    roc_auc=auc(false_positive_rate, true_positive_rate)\n",
        "    plt.title('ROC')\n",
        "    plt.plot(false_positive_rate, true_positive_rate,'b',label='AUC = %0.4f'% roc_auc)\n",
        "    plt.legend(loc='lower right')\n",
        "    plt.plot([0,1],[0,1],'r--')\n",
        "    plt.ylabel('True Positive Rate')\n",
        "    plt.xlabel('False Positive Rate')\n",
        "    plt.show()\n",
        "\n",
        "# 依次训练不同分类器\n",
        "# LogisticRegression\n",
        "lr = LogisticRegression(random_state =2018)\n",
        "lr.fit(X_selected, over_samples_y)\n",
        "\n",
        "# 决策树\n",
        "dt = DecisionTreeClassifier(random_state=2018)\n",
        "dt.fit(X_selected, over_samples_y)\n",
        "\n",
        "# SVM\n",
        "svm_linearSVC = LinearSVC(random_state=2018)\n",
        "svm_linearSVC.fit(X_selected, over_samples_y)\n",
        "\n",
        "# 随机森林\n",
        "rfc = RandomForestClassifier(n_estimators=100, random_state=2018)\n",
        "rfc.fit(X_selected, over_samples_y)\n",
        "\n",
        "# GBDT\n",
        "gbc = GradientBoostingClassifier(random_state=2018)\n",
        "gbc.fit(X_selected, over_samples_y)\n",
        "\n",
        "# xgboost\n",
        "xgb = XGBClassifier(random_state=2018)\n",
        "xgb.fit(X_selected, over_samples_y)\n",
        "\n",
        "# lightgbm\n",
        "lgbm = LGBMClassifier(random_state=2018)\n",
        "lgbm.fit(X_selected, over_samples_y)\n",
        "\n",
        "# 依次对训练好的多个模型用多个指标评价，并画出ROC曲线\n",
        "X_test.drop(['Unnamed: 0','id_name', 'bank_card_no', 'reg_preference_for_trad', 'latest_query_time', 'loans_latest_time','source','student_feature'], axis=1, inplace=True)\n",
        "num = imputer.fit_transform(X_test)\n",
        "X_test = pd.DataFrame(num, columns=X_test.columns)\n",
        "X_test = X_test.iloc[:, importances > 0.01]\n",
        "\n",
        "model_name = [\"lr\", \"dt\", \"svm_linearSVC\", \"rfc\", \"gbc\", \"xgb\", \"lgbm\"]\n",
        "for name in model_name:\n",
        "    model = eval(name)\n",
        "    print(name)\n",
        "    multi_score(model, X_test, Y_test)\n",
        "\n",
        "    "
      ],
      "execution_count": 19,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/sklearn/linear_model/logistic.py:433: FutureWarning: Default solver will be changed to 'lbfgs' in 0.22. Specify a solver to silence this warning.\n",
            "  FutureWarning)\n",
            "/usr/local/lib/python3.6/dist-packages/sklearn/utils/validation.py:761: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples, ), for example using ravel().\n",
            "  y = column_or_1d(y, warn=True)\n",
            "/usr/local/lib/python3.6/dist-packages/sklearn/utils/validation.py:761: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples, ), for example using ravel().\n",
            "  y = column_or_1d(y, warn=True)\n",
            "/usr/local/lib/python3.6/dist-packages/sklearn/svm/base.py:931: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
            "  \"the number of iterations.\", ConvergenceWarning)\n",
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:62: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples,), for example using ravel().\n",
            "/usr/local/lib/python3.6/dist-packages/sklearn/utils/validation.py:761: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples, ), for example using ravel().\n",
            "  y = column_or_1d(y, warn=True)\n",
            "/usr/local/lib/python3.6/dist-packages/sklearn/preprocessing/label.py:219: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples, ), for example using ravel().\n",
            "  y = column_or_1d(y, warn=True)\n",
            "/usr/local/lib/python3.6/dist-packages/sklearn/preprocessing/label.py:252: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples, ), for example using ravel().\n",
            "  y = column_or_1d(y, warn=True)\n",
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:77: SettingWithCopyWarning: \n",
            "A value is trying to be set on a copy of a slice from a DataFrame\n",
            "\n",
            "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "lr\n",
            "准确率 0.680448493342677\n",
            "精确率 0.4079696394686907\n",
            "召回率 0.5988857938718662\n",
            "F1-score 0.4853273137697516\n",
            "AUC 0.7162582287461008\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAe8AAAFnCAYAAACPasF4AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nXg4sRm3NAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 576x396 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "dt\n",
            "准确率 0.7021723896285914\n",
            "精确率 0.417910447761194\n",
            "召回率 0.467966573816156\n",
            "F1-score 0.44152431011826543\n",
            "AUC 0.6244327251103252\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAe8AAAFnCAYAAACPasF4AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4yLCBo\ndHRwOi8vbWF0cGxvdGxpYi5vcmcvOIA7rQAAIABJREFUeJzs3Wl4U+X29/HvztQpKbTQgoDIpIgg\noiCKIMikiIqiKKAyCDK0zIXDJIrIICAyWsogoqIPg1iOiB7QI6KoKIo4gCgKMk8tlDZpm3k/L/ax\nf1BKgTbZTbo+1+V1SNImq/dp88vOXlm3oqqqihBCCCFChkHvAoQQQghxeSS8hRBCiBAj4S2EEEKE\nGAlvIYQQIsRIeAshhBAhRsJbCCGECDEmvQsQQgRX3bp1qV69OkajEQCfz8ett97KhAkTiI6OBuDU\nqVPMnj2bHTt2YDQaiYiIoFu3bnTv3r3gftxuN6mpqWzatIm/PnHaoUMHBg0ahMViCf4PJkQZosjn\nvIUoW+rWrctnn31G5cqVAS2ER4wYQZ06dRgxYgR5eXl07tyZjh07MmjQIEwmE0eOHGHIkCG0bduW\nwYMHAzB8+HDy8/N56aWXiI2N5ezZs4wZMwar1crLL7+s548oRNiTI28hyjiLxcKdd97J5s2bAVi3\nbh3x8fEMGzas4GuqVavG9OnT6dq1K7169eLEiRN89tlnfPrpp8TGxgJQvnx5pk2bxt69e3X5OYQo\nSyS8hSjjsrOz2bBhA82bNwdg+/bttG7d+h9fV7duXeLj4/npp584cOAAjRo1onz58ud9TYUKFWjW\nrFlQ6haiLJPwFqIM6tGjB0ajEY/HQ3Z2Nr1796Zfv36AFuZxcXEX/L6KFSuSnZ1NdnY2FSpUCGbJ\nQohzSHgLUQatWLGCypUrc+bMGTp06EDHjh0xmbSng7i4OE6dOnXB78vMzCQ+Pp7s7GxOnjwZzJKF\nEOeQj4oJUYbFx8fTo0cPXnrppYLrWrZsySeffPKPr927dy/Z2dk0bNiQpk2b8uOPP/4jwHNycpg3\nbx7SBytEYEl4C1HGPfXUU+zcuZPt27cD0KlTJ7xeL9OnT8fj8QBw7Ngxxo4dS3JyMtHR0dSuXZuO\nHTuSkpJCZmYmAGfPniUlJYWsrCwURdHt5xGiLJCPiglRxvz9o2IAr732Gh988AFr165FURROnz7N\nrFmz+PbbbzGZTERERPDkk0/y6KOPFnyPx+MhLS2NDz74AEVRMJvNdOrUib59+2IwyHGBEIEk4S2E\nEEKEGHl5LIQQQoQYCW8hhBAixEh4CyGEECFGwlsIIYQIMRLeQgghRIgJmQlrGRn2Er2/uLhosrLy\nSvQ+yyJZx+KTNSw+WcPikzUsvkCsYUKC7YLXl9kjb5PJqHcJYUHWsfhkDYtP1rD4ZA2LL5hrWGbD\nWwghhAhVEt5CCCFEiJHwFkIIIUKMhLcQQggRYiS8hRBCiBAj4S2EEEKEGAlvIYQQIsRIeAshhBAh\nJqDhvXfvXtq1a8dbb731j9u++uorunTpQteuXUlNTQ1kGUIIIURYCVh45+XlMXnyZJo1a3bB26dM\nmcKCBQtYuXIlX375JX/88UegShFCCCHCSsDC22KxsHTpUhITE/9x2+HDhylXrhxXXXUVBoOBVq1a\nsW3btkCVIoQQQgTU3s3HWbNaRVWD83gB25jEZDJhMl347jMyMoiPjy+4HB8fz+HDhy96f3Fx0SU+\nN7awge/i8sg6Fp+sYfHJGhafrOHlUVX4eJOfP4YtoPfecSxhES2P96Ry5cA/dsjsKhaInVpKeqey\nskjWsfhkDYtP1rD4ZA0vnccD69aZ2DD3EOP+6EcyWzlrqsjICTaMRjsZGSX3WIW9oNIlvBMTE8nM\nzCy4fPLkyQu+vS6EEEKUFnY7rFhhZuliE52Pp/EuY4gmn5N3Pohx0Wwa3lAraC+AdPmoWLVq1XA4\nHBw5cgSv18unn35K8+bN9ShFCCGEuKjjxxUmTYqgUSMrzz8fiXL2LNMiX8BSLpKcxa9hWPsmakJC\nUGsK2JH3rl27mDFjBkePHsVkMrFp0ybatGlDtWrVaN++Pc8//zwjR44EoGPHjtSsWTNQpQghhBCX\nbc8eAwsXWkhPN+H1qNwSv5+7x1Wjd+8IvL+9zdmatVArVdKlNkVVg9UbVzwl/VaEnN8pGbKOxSdr\nWHyyhsUna6hRVfjiCyOpqRY2b9aOb++65g/eMD9N1exfyfriG9S4+At+byDWsFSd8xZCCCFKE68X\n3n/fRGqqhZ9+0j7ZdHtTN3PrpXHrO89gyMvF1aEjeH06V6qR8BZCCFFmORywcqWZxYstHDpkQFFU\n7r/fQ8ojf9D8tWQsb2zBX648OalLcHXpCoqid8mAhLcQQogy6ORJhWXLzLz+uoWzZxUiI1V693Yz\ncKCbWrVUynUegOXLrbju7oBj1jz8la/Su+TzSHgLIYQoM37/3UBampk1a8y43Qrx8X5GjXLTp4+H\nitG5EB0NgGPqTEw//YCr6+Ol5mj7XBLeQgghwpqqwjffaE1omzZpsVejhp+kJBddu3qIjlKJfPtN\nYqZM5Oza9/E1uBHfDfXx3VBf58oLJ+EthBAiLPl88OGHJhYutLBjh9aE1rixj+RkNx07ejEawXDs\nKLanBmP59BP8tliMhw/ha3CjzpUXTcJbCCFEWMnLg9WrzSxaZOHPP7VZZB06eEhO9nDbbT7tXXBV\nJWLl21gnjMVgz8Hdui322QvwV62mb/GXSMJbCCFEWMjMVHjtNTPLl5s5fdqAxaLy5JNukpI8XHut\n/7yvjVqcivW58fitNuyzF+B8omepPLddGAlvIYQQIW3/foVFiyysWmXG6VQoX15lxAgXffp4qFTp\nnDlkf80kUxSc3Z7A9OMP5D4zEX+1q/UpvBgkvIUQQoSk777Txpd+8IEJVVWoXt3PgAEuunf3YLWe\n/7WGE8exjhqG65HHcHXuglo+Dnvaq/oUXgIkvIUQQoQMvx8++kjrHP/mGy3CGjb0MWiQmwce8GL6\ne6qpKhHvrML6zBgM2WdRo6Jxde4S/MJLmIS3EEKIUs/phLVrzaSlmfn9d61zvG1bL4MGuWne3HfB\n09XKyZPY/jWMiI0fokbHYJ85B2evPkGuPDAkvIUQQpRaWVnw+usWXn3VTEaGAbNZpWtXD8nJburV\n8xf6fca9v1H+gbsxZGXhbtES+5xX8F9TI3iFB5iEtxBCiFLn0CGFxYstvP22mbw8BZtNZfBgF/36\nebjqqqI3w/TVroO30S247r4X51NPg8EQhKqDR8JbCCFEqfHjj1oT2vr1Jnw+hSpV/Iwe7aJHDw+2\nC++OqVFVIv79LobDh8gfmgJGI9mr0kPq41+XQ8JbCCGErlQVNm/WmtC++EKLpRtu0Cahde7sxWy+\n+PcrGRnYxqQQseE9/FYbzid7ocZXCNvgBglvIYQQOnG7IT3dRFqahT17tCa0li21JrS77rpwE9rf\nRbyXjnXsSAynT+O5rRk58xZqwR3mJLyFEEIEVU4OvPGGhaVLzZw4YcBoVHn4YQ+DBrm58cbCm9DO\n4/NhG9iXyPfSUaOicEx+kfx+SWF3brswEt5CCCGC4uhRhSVLLKxYYcbhUIiJURkwwM2AAW6qVSu6\nCe08RiNq+Tg8t96Gff5CfLWvDUzRpZSEtxBCiIDavVtrQlu3zoTXq5CY6Gf4cDc9e7opX/7S70c5\nfZrI/7eC/MHDQFFwvDANLBYwGgNXfCkl4S2EEKLEqSp8/rnWhLZlixY1112nNaE98oiXiIjLuz/L\nB+9j+9dwDJkZ+GrVxn3fAxAVFYDKQ4OEtxBCiBLj8cD69SZSUy3s2qUdEd9xh9aE1rat77JPSStn\nTmMdP5rI9HdQIyJwTJyCu0PHAFQeWiS8hRBCFJvDAW+9ZWbJEgtHjhgwGFQ6ddKa0G6++RKb0P7G\n8vFGrCOGYDx1Ek/jJtjnL8J37XUlXHlokvAWQghxxU6eVFi61Mwbb1jIzlaIilLp21drQqtR4zKb\n0P7GcOwYhrNZOJ59gfykwfxz15GyS1ZCCCHEZfvtNwNpaWbWrjXjditUrOhnzBg3Tz3lJj7+yu/X\nvPljPLfdATExOHs+hbtVa/w1apZc4WFCwlsIIcQlUVXYtk1rQvv4Yy0+atXyk5Tk4rHHPMXqH1PO\nZmGdMJbINSvJG5BM7uTpoCgS3IWQ8BZCCHFRXi98+KHWhLZzp9aEduut2h7aHTp4iz0XxfLfTVhT\nhmI8cRzPTTfjfLxnCVQd3iS8hRBCXFBuLqxaZSYtzcKhQwYUReXee7UmtKZNr6wJ7VxK9llinhtP\n1Mq3UM1mcsc9S97g4RQ5zFxIeAshhDhfRobCsmVmli+3kJWlEBGh0rOnm6QkN7VrF68J7VzGvb8R\nueptPDfehH1+Gr76DUrsvsOdhLcQQggA9u1TSEuzsGaNGadTIS5OJSXFRd++HhISSia0FXsOSm4u\n/spX4b31NrJXr8PT/E452r5MEt5CCFHGbd9u4NVX4b33YlBVherVtSa0bt08xMSU3OOYP/0EW8oQ\nfNfUIDt9AxgMeO5qU3IPUIZIeAshRBnk98PGjVoT2rffak1oN9/sZ9AgN/fd5y3RceGKPYeY5ycQ\nteJ1VJMJZ/cntQLKyA5ggSDhLYQQZUh+PqxZozWh7d+vhWf79l6eecZEvXp5l7SH9uUwf/YpthGD\nMR45jLdefeyvLMJ7400l+yBlkIS3EEKUAWfOwPLlFpYtM5OZacBiUXn8cTdJSR7q1vWTkGAjI6Nk\nH1Nx2Int1wvFbic3ZTR5KaO1XcBEsUl4CyFEGDtwQGHxYgsrV5rJy1OIjVUZOtRFv34eKlUquc7x\ncyn2HFRbLKrVhn3BYvyVK+O96eaAPFZZJeEthBBhaOdOA6mpFjZsMOH3K1St6mfsWBdPPunBag3Q\ngzocWCc/h+XjTWRt+Qo1thzue+4N0IOVbRLeQggRJvx++OQTbXzpV19pT+8NGmiT0Dp18gb001jm\nr77ANjQZ46EDeOtejyHjFL7YcoF7wDJOwlsIIUKcywXvvmsiLc3Cb79pbeJ33aXtod2ypa/Em9DO\nk5uLdcpEopYtQTUYyBuaQu6osRAZGcAHFRLeQggRorKz4Y03LCxdaubkSQMmk8qjj3pISnLToEHx\nx5deitikp4nY+AHea6/DPj8Nb+Nbg/K4ZZ2EtxBChJjDhxWWLLHw1ltmcnMVrFaVpCQ3/fu7qVo1\nME1o51FV/jqczxs1Bl/tOuSOeUaOtoNIwlsIIULEzz9rTWjvvWfC51OoXNlPSoqbXr3cxMYGpwbT\n19uwjRlBztI38F1XF2/DRngbNgrOg4sCEt5CCFGKqSp8+qnWhLZ1q/aUXa+ej6QkNw8/7A3ex6bz\n84l5cTJRi1MBMH/xOb7r6gbpwcXfSXgLIUQp5PHAunUmFi608MsvWhNaixZaE1qbNgFuQvsb07ff\nYBuahGnfH3hr1cY+Lw3vbbcHrwDxDxLeQghRitjtsGKFmSVLLBw7ZsBgUOnc2UNyspubbgpOE9q5\nIt5dg21Qf1BV8gYMInfcsxAdHfQ6xPkkvIUQohQ4flxrQnvzTTN2u0J0tEq/fm4GDHBTvXoQmtAK\n4W7VBm/jW8l9dhKe2+/QrQ5xvoCG97Rp0/jxxx9RFIXx48fTsGHDgtvefvtt1q9fj8FgoEGDBjzz\nzDOBLEUIIUqlPXsMLFxoIT3dhMejkJDgZ/BgN717u4mL06Egp5OYl17E06Qp7nvvQ61YkbMffKxD\nIeJiAhbe27dv5+DBg6xevZp9+/Yxfvx4Vq9eDYDD4WDZsmV89NFHmEwm+vTpww8//ECjRtKxKIQI\nf6oKX3yhNaFt3qw9Ddep4yM52UOXLh7dPnFl2rkD25CBmPb+hufW23B36EhQT66LSxaw8N62bRvt\n2rUDoHbt2mRnZ+NwOLBarZjNZsxmM3l5eURHR5Ofn0+5cjJGTwgR3rxeeP99bQ/tn37SmtBuu01r\nQrv7bp9+21u7XDD+RcrPmIHi95Pftz+OCZMkuEuxgIV3ZmYm9evXL7gcHx9PRkYGVquViIgIBg0a\nRLt27YiIiOC+++6jZs2agSpFCCF05XDAypVmFi+2cOiQAUVRuf9+rQmtSZPgN6Gdy3DsKOW6PQy/\n7sFf/Rrs8xbiaX6nrjWJogWtYU1V/6/hwuFwsHjxYjZu3IjVaqVXr178+uuvXH/99YV+f1xcNCaT\nsURrSkiwlej9lVWyjsUna1h8pXENT5yABQsgLQ2ysrQBZElJkJKiUKeOGQjgTiGXKq4OxNogORnj\njBmUD9iWY2VDsH4PAxbeiYmJZGZmFlw+deoUCQkJAOzbt4+rr76a+Ph4AJo0acKuXbsuGt5ZWXkl\nWp+28by9RO+zLJJ1LD5Zw+IrbWv4++8G0tLMrFljxu1WiI/3M2qUhz59PFSsqB3IZGToV5/ppx8w\n/fQjzid7aVes3UDC1QnaGuaXnnUMNYH4PSzsxUDAzrA0b96cTZs2AbB7924SExOx/u8VXdWqVdm3\nbx9OpxOAXbt2UaNGjUCVIoQQAaeq8PXXRnr0iKJ58xjeestClSoqM2Y4+f77XEaPdhcEt27cbqKn\nT6H8Pa2xjknBcPyYdr3MJA85ATvyvuWWW6hfvz7dunVDURQmTpxIeno6NpuN9u3b07dvX3r27InR\naOTmm2+mSZMmgSpFCCECxueDDz/UJqHt2KGd2mvc2EdyspuOHb0YS/Zs3xUz/vwTsUMGYvplF75q\nV2Of8wr+q6roXZa4Qop67snoUiwQb0WUprfZQpWsY/HJGhafHmuYlwerV5tZtMjCn39qb2J26OAh\nOdnDbbcFd3zpRakq0bOmEz3nJRSvl/wevcl9fgqq7fydTOT3sPiC+ba5TFgTQojLkJmp8NprZpYv\nN3P6tAGLReXJJ90kJXm49lp9O8cvSFEwHj6EP7ES9tkL8LRpp3dFogRIeAshxCXYv19h0SILq1aZ\ncToVypdXGT7cRd++HipVKmVvYHo8RHywHteDD4Oi4JgyHQA1VuZphAsJbyGEuIjvvtPGl37wgQlV\nVbj6aj8DB7ro3t1DafxUlfGX3diGJmH+6QdyANdDj0hohyEJbyGE+Bu/Hz76SBtf+s032tNkw4Y+\nBg1y88ADXkyl8ZnT6yV6wRyiZ01H8XhwdnsCd+u2elclAqQ0/goKIYQunE5Yu9ZMWpqZ33/X2sTb\ntvWSnOymRYtS1IT2N8Zf92AbOhDzDzvxVaqMY/Z83O076F2WCCAJbyFEmZeVBa+/buHVV81kZBgw\nm1W6dtXGl9arVwqb0P7G8vmnmH/YifOx7jimTEctr8d2ZCKYJLyFEGXWoUMKixdbePttM3l5Cjab\nyqBBbvr3d3PVVaWsCe1vjH/8jq/a1RAZSf7TA/HWv1FmkpchEt5CiDLnxx+1JrT16034fApVqvgZ\nPdpFjx4ebKVvRPr5fD6iFi4gZuZU8vslkfvcC2AwSHCXMRLeQogyQVVh82atCe2LL7Snvnr1tCa0\nhx7yYrHoXOAlMP6+V+sk3/Et/ooJeJo01bskoRMJbyFEWHO7IT3dRFqahT17tCa0li21JrTWrUtv\nE9p5fD6iFi8k5sUXUFwunA93wTHtJdT4CnpXJnQi4S2ECEs5OfDGGxaWLjVz4oQBo1Hl4Yc9DBrk\n5sYbS38T2rlMP3yP9fln8FesSE7aMtz3d9K7JKEzCW8hRFg5elRhyRILK1aYcTgUoqNVBgzQmtCu\nvrp0N6Gdx+9HseegliuPt/Gt5MxPw93uHtSKFfWuTJQCEt5CiLCwe7fWhLZunQmvVyEx0c/w4W56\n9nRTvrze1V0ew/59xA5LRo2KInv1OlAUXN2e0LssUYpIeAshQpaqwuefG1m6FD76KAaA667TtuN8\n5BEvERE6F3i5/H6iXl1EzNRJKPn5uB54CPLzITpa78pEKSPhLYQIOR4PrF9vIjXVwq5dWhPaHXdo\nTWjt2vkwGHQu8AoY/tyPbfggLNu+xB8fj31+mraxiBAXIOEthAgZDge89ZaZJUssHDliwGBQ6dTJ\nwzPPmKlZM1/v8q6c00n5B+7BeOokrvs6YZ8xGzUxUe+qRCkm4S2EKPVOnlRYutTMG29YyM5WiIpS\n6dPHzYABbmrWVElIMJORoXeVV8DnA6MRIiPJfX4KGI24HnqE0Pj8mtCThLcQotT67TcDaWlm1q41\n43YrVKzoZ8wYN717e6hQIYQ6x//O7yfy9WVEvbmcrA0fgdWKq0tXvasSIUTCWwhRqqgqbNumTUL7\n+GPtKapWLT9JSS4ee8xDVJTOBRaT4dBBbCMGY9n6Gf7y5TH9tgdv41v1LkuEGAlvIUSp4PXChx9q\nTWg7d2pNaLfeqo0vveceL0ajzgUWl6oS+eZyYp6fgCHXgeuee3HMmoe/UmW9KxMhSMJbCKGr3FxY\ntcpMWpqFQ4cMKIrKvfdqk9CaNg2tSWgXYx2dQtQby/CXK0/OK4txPdpNzm2LKybhLYTQRUaGwrJl\nZpYvt5CVpRARodKzp5ukJDe1a4fw+exCOLt2x3DiGI6X5uKvfJXe5YgQJ+EthAiqffsU0tIsrFlj\nxulUiItTSUlx0bevh4SE8Altw5HDWMePxvH8FPy1auNt0pScFav1LkuECQlvIURQbN9uIDXVwsaN\nJlRVoXp1rQmtWzcPMTF6V1eCVJXI/7eCmGfHYXDY8dW9ntxnJupdlQgzEt5CiIDx+2HjRq0J7dtv\ntY6zm2/WmtDuuy8MmtD+xnDsKLaUIVg2/xe/LRb73FSc3Z/UuywRhiS8hRAlLj8f1qzRmtD279dm\nlbZv72XQIDfNmoXIHtqXyfz5FmL79MCQk437rjbY57yCv2o1vcsSYUrCWwhRYs6cgeXLLSxbZiYz\n04DFovL4426SkjzUrRs+neMX4qt7PWq5ctifn4LziZ7SSS4CSsJbCFFsBw4oLF5sYeVKM3l5CrGx\nKkOHuujXz0OlSuHThHYeVSVizUr8iZXwtG6Lv1Jlzny9E8xmvSsTZYCEtxDiiu3cqTWhbdhgwu9X\nqFrVz9ixLp580oPVqnd1gWM4cRzrqGFEfLQR77XXkbV1OxgMEtwiaCS8hRCXxe+H//5XG1+6bZv2\nFNKggdaE1qmTN7zzS1WJWLsa6zOjMZw9i/vOVtjnvEJI7kEqQpqEtxDikrhc8O67JhYutLB3r9Ym\nftddWhNay5bh2YR2LuVsFrahyURs/AA1Ogb7jNk4e/WR4Ba6kPAWQlzU2bPw5psWli41c/KkAZNJ\n5dFHPSQluWnQILyb0M6lRsdgPHwId/M7sc9NxX9NDb1LEmXYJYV3VlYWR44c4cYbb8Tv92OQV5pC\nhL3DhxWWLLHw1ltmcnMVrFaVpCQ3/fu7qVo1TJvQ/kY5dQrzd9txd7wfLBbOrvk3aoUKcrQtdFdk\neG/YsIH58+djsVjYsGEDkydP5oYbbuDRRx8NRn1CiCD7+WetCe2990z4fAqVK/tJSXHTq5eb2Fi9\nqwsSVSXivXSsY0eiOBxkff41vlp1UBMS9K5MCACKfPm4fPly3nvvPeLi4gAYM2YMa9asCXhhQojg\nUVXYvNnII49E0bZtDOnpZq67zs/8+fl8910uQ4aUneBWMjKI7duT2P5PoeTnkztxMr4atfQuS4jz\nFHnkbbPZiIqKKrgcGRmJOazbSYUoOzweWLdOa0L75RetCa1FC60JrU2b8G9C+zvL+nXYxqRgOH0a\nz23NyJm3EH+t2nqXJcQ/FBnecXFxrFu3DpfLxe7du/nwww+Jj48PRm1CiACx22HFCjNLllg4dsyA\nwaDSubOH5GQ3N91UdprQ/i7iw/dRcnNxTH6R/KcHEnbD10XYUFRVvWjnSU5ODnPnzuWbb77BYrHQ\nuHFjhgwZQrly5YJVIwAZGfYSvb+EBFuJ32dZJOtYfMFcw+PHtSa0N980Y7crREerPPGEhwED3FSv\nHrpNaMVZQ9OOb/E2vhUA5cxpDGfO4KtzbUmWFxLkb7n4ArGGCQm2C15f5JH31q1bee655867buXK\nlXTv3r1kKhNCBNyePQYWLrSQnm7C41FISPAzeLCb3r3d/K+dpcxRzpzGOm4UkeveJXvZCtwPPIga\nXwFffAW9SxOiSIWG9y+//MLu3bt57bXXyM/PL7je6/WSmpoq4S1EKaeq8MUX2iS0zZu1P/U6dXwk\nJ3vo0sVDZKTOBerI8uEGbP8ajiHjFJ7GTfBdX0/vkoS4LIWGd0REBKdPn8Zut7Njx46C6xVFYfTo\n0UEpTghx+bxeeP99bQ/tn37SztnedpvWhHb33b4y/RFlJesM1vGjiXx3DWpEBI5nXyA/eYic2xYh\np9Dwrl27NrVr1+b222+nUaNG5922adOmgBcmhLg8DgesXGlm8WILhw4ZUBSV++/XmtCaNCm7TWjn\nilyzksh31+C5pTH2+YvwXVdX75KEuCJFnvNOTExk5syZZGVlAeB2u/nmm2+45557Al6cEKJoJ08q\nLFtm5vXXLZw9qxAZqdK7t5uBA93UqhW6TWglRTmbhRodAxYL+X0HoNpicT7WHUwyHVqEriLfQBs9\nejTly5fnhx9+oEGDBmRlZTFz5sxg1CaEuIjffzeQkhJB48YxzJ0bgcGgMmqUi++/z2XmTJcEN2D5\neCNxLW8nevb/nrNMJpyP95DgFiGvyN9go9FI//792bp1K0888QRdunQhJSWFO+64Ixj1CSHOoarw\nzTdaE9qmTdqfb40afpKSXHTt6iE6WucCSwkl+yzWZ8cRueptVLMZNSaMNxcXZVKR4e1yuThx4gSK\nonD48GGqVKnC0aNHg1GbEOJ/fD748ENtEtqOHVpzVePGPpKT3XTs6JV+q3NYPvkIa8pQjMeP4WnY\nCPv8NHw31Ne7LCFKVJHh/fTTT7Nt2zb69u3Lgw8+iNFo5P7777+kO582bRo//vgjiqIwfvx4GjZs\nWHDb8ePHSUlJwePxcMMNN/DCCy9c+U8hRJjKy4PVq80sWmThzz+1s1wdOnhITvZw221lb3xpUYx7\nfqFc9y6oZjO5YyeQN2QEyDhnEYaKDO927doV/Hv79u3k5uZe0nS17du3c/DgQVavXs2+ffsYP348\nq1evLrh9+vTp9OnTh/bt2zOyv4C9AAAgAElEQVRp0iSOHTtGlSpVrvDHECK8ZGYqvPaameXLzZw+\nbcBiUXnySTdJSR6uvVY6x//B4wHAV+8GHBOex932bnz1G+hclBCBU2jDmt/vZ9WqVUyePJkNGzYA\nYDKZsFgsTJo0qcg73rZtW0Hw165dm+zsbBwOR8F979ixgzZt2gAwceJECW4hgP37FUaPjuCWW2KY\nNSsCn09h+HAXO3bkMnu2S4L7bxR7DtaUIdCli9YQAOQPTZHgFmGv0CPvyZMnk52dTaNGjVi1ahVZ\nWVnUqVOH55577ryj8cJkZmZSv/7/nWeKj48nIyMDq9XKmTNniImJ4cUXX2T37t00adKEkSNHXvT+\n4uKiMZlK9sReYTNjxeWRdSy+fftszJoF6elaBl1zDaSkQJ8+ClZrBBChd4mlz8cfQ9++cPgw3HQT\nCSYvyKZJxSJ/y8UXrDUsNLz37NnDqlWrAOjSpQutW7ematWqzJkzhwYNLv9V7bn7n6iqysmTJ+nZ\nsydVq1alf//+bNmyhbvuuqvQ78/Kyrvsx7wYGcJfMmQdr5zfDx99ZGTJkmi++EK7rmFDH4MGuXng\nAS8mE+Tna/+J/6M47MRMnEDUiuWoJhN5o8YSM3USGdkukN/FKyZ/y8VXKjYmOXfP7ujoaGrWrMnb\nb7+N8RLbWhMTE8nMzCy4fOrUKRISEgBtm9EqVapQvXp1AJo1a8bvv/9+0fAWIlw4nbB2rZm0NDO/\n/679PbVt6yU52U2LFtKEdlE+H+XvbYvpt1/x1quPfUEa3oaNiLFYAJfe1QkRNIWe81b+9gxisVgu\nObgBmjdvXjBGdffu3SQmJmK1ap+1NJlMXH311Rw4cKDg9po1a15u7UKElKwsmDPHQuPGMaSkRHLg\ngIGuXT38/DOsXJnPnXdKcBfJaCS/T39yR4wi66MteBs2Kvp7hAhDhR55nzp1irVr1xZczsjIOO9y\nly5dLnrHt9xyC/Xr16dbt24oisLEiRNJT0/HZrPRvn17xo8fz9ixY1FVleuuu66geU2IcHPokMLi\nxRbefttMXp6CzaYyaJCb/v3dXHWVSkKCmYwMvassvcxffE70gjlkv/7/ICoK51NP612SELpT1HNP\nRp9j3LhxF/3GF198MSAFFSYQ5xHk/E7xyToW7scftT2016834fMpVKnip39/Nz16eLCdcxpL1rAQ\nDgfWKROJem0pqsFAzlurcbe78J4KsobFJ2tYfKXinHeww1mIcKCqsHmzNr70iy+0P6969bQmtIce\n8mKx6FxgiDBv+xLb0CSMBw/grXs99vlpeG9urHdZQpQaMp1fiBLgdkN6uom0NAt79mi9IS1bak1o\nrVvLuezLETV/NtYpz6MaDOQNGUHuv8ZBZKTeZQlRqkh4C1EMOTnwxhsWli41c+KEAaNR5eGHPQwa\n5ObGG2WgypXwNmmK97q62OctxNv4Vr3LEaJUkvAW4gocPaqwZImFFSvMOBwK0dEqAwZoTWhXXy1b\ncV6WvDxiZk0n/6mn8V9dHc8dLcj67GtktxUhCldkeP/666+MHz+evLw8Nm7cSGpqKi1atOCmm24K\nRn1ClCq7d2tNaOvWmfB6FRIT/Qwf7qZnTzfly+tdXegxffM1tmFJmPbvQ8k6g2POK9oNEtxCXFSh\nn/P+ywsvvMC0adMKBqx07NhRmtlEmaKq8NlnRh57LIrWrWN45x0ztWr5mTs3nx07chk6VIL7suXn\nE/PceMp3ugfjn/vJGzgYx7SX9K5KiJBR5JG3yWTi+uuvL7hcs2ZNTCZ5t12EP48H1q83kZpqYdcu\n7Ujwjju0JrR27XwYinzpKy7EuHsXsU/3xLTvD7w1a2Gfl4b39mZ6lyVESLmk8D58+HDBxLXPPvuM\nQj4aLkRYcDjgrbfMLFli4cgRAwaDSqdOHpKT3dxyizShFZdavjyGzEzyBiSTO+45iI7WuyQhQk6R\n4T1mzBiSk5P5888/ady4MVWrVmXmzJnBqE2IoDp5UmHpUjNvvGEhO1shKkqlTx83Awa4qVlTXrAW\nh+n778DjxXvb7firVuPMNztR4yvoXZYQIavI8Dabzbz//vucOXMGi8VSMJ9ciHDx228G0tLMrF1r\nxu1WqFjRz5gxbnr39lChgoR2sTidxLz0IlGp8/BXq86ZbTvAbJbgFqKYigzvpKQkbDYbnTp14v77\n7w9GTUIEnKrCtm3aJLSPP9b+DGrV8pOU5OKxxzxERelcYBgw7dyBbWgSpt9+xXdNDezzFsI5uxUK\nIa5ckeG9adMmdu3axX/+8x+6detGzZo1efDBB+nYsWMw6hOiRHm98OGHWhPazp1aE1qTJtr40g4d\nvPIJpZLgchEzazpRr8xF8fnI79MPx4RJIO/aCVFiLqltvEGDBjRo0IDevXuzcOFCRo8eLeEtQkpu\nLqxaZSYtzcKhQwYUReXee7VJaE2bShNaiVJVLP/ZgL9qNexzU/G0aKl3RUKEnSLD+9SpU3z00Uds\n3LiRM2fO0LFjRz744INg1CZEsWVkKCxbZmb5cgtZWQoRESo9erhJSnJTp46czy4xbjemH3bibXob\nREaS8+ZKfImV5WhbiAApMrwfeeQROnbsyJgxY7jxxhuDUZMQxbZvn0JamoU1a8w4nQpxcSopKS76\n9PGQmCihXZJMP/+IbfBAjAf2k7X5C3y1r8VXq47eZQkR1goN71OnTpGYmMibb75ZMJTl8OHDBbdf\nffXVga9OiMu0fbuB1FQLGzeaUFWF6tW1JrRu3TzExOhdXZhxu4meO4voubNQvF7yezyFP7GS3lUJ\nUSYUGt4zZszg5Zdfpm/fviiKct5gFkVR+OSTT4JSoBBF8fth40atCe3bb7WOs0aNtCa0++7zIgMB\nS55x18/YhiZh3vUTvqrVsM9egKd1W73LEqLMKPRp7eWXXwZg6dKl1K5d+7zbdu7cGdiqhLgE+fmw\nZo3WhLZ/vzartH17L4MGuWnWTPbQDqToeS9j3vUT+U/2Ivf5Kaix5fQuSYgypdDwzsnJ4ezZs4wf\nP55Zs2YVXO/xeBg7diybNm0KSoFC/N2ZM7B8uYVly8xkZhowm1W6d/eQlOTm+uulczxQDMeO4q9S\nFQDH1Jk4uz+Bp017nasSomwqNLx37tzJG2+8wZ49e+jVq1fB9QaDgRYtWgSlOCHOdeCAwqJFFlau\nNJOfrxAbqzJ0qIunn/ZQubI0oQWMx0P0gjlEvzyDnNffxt2+A2piogS3EDoqNLxbtWpFq1atWLly\nJd27dw9mTUKcZ+dOrQltwwYTfr9C1ap+Bgxw8eSTHvkkUoAZ9/yindv+cSe+ylehmi16lySE4CLh\n/e677/LII49w8uRJ5s2b94/bhw0bFtDCRNnm98N//6uNL922Tfs1rV9fa0J78EGvTNkMNK+XqNR5\nxLz0IorbjbPr4zgmv4haPk7vyoQQXCS8Df/brFj27hbB5HLBu++aWLjQwt69Wuf4XXdpTWgtW0oT\nWrBEvvUG1qmT8FWqjOPlebjvvlfvkoQQ5yg0mTt37gzA4MGDcTgcWK1WMjMzOXDgALfcckvQChRl\nw9mz8OabFpYuNXPypAGTSaVLF20P7QYNpAktKLxe7X9NJpxP9MRw4hj5AwfL0bYQpVCRh9WTJ0/m\n+uuvp3379nTr1o0GDRqwfv16XnjhhWDUJ8Lc4cMKS5ZYeOstM7m5ClarSlKSm/793VStKk1owWLc\n+xu2oQNxd7iPvOGjwGwmb+yzepclhCiEoagv+OWXX3j00Uf5z3/+Q+fOnZk7dy4HDx4MRm0ijP38\ns4GBAyNp2jSGxYst2Gwqzz7rYudOB5MmuSS4g8XnI+qVecS1bYH5+x0Y/tyv7ZcqhCjVijzy/muy\n2pYtWxg+fDgAbrc7sFWJsKSq8OmnWhPa1q3ar9711/tITnbz8MNeLNLIHFTGP37XOsm/246/YgI5\ni+fh7ni/3mUJIS5BkeFds2ZNOnbsSHx8PPXq1ePf//435crJNCVx6TweWLdOa0L75RetCa1FC60J\nrU0baULTg+HgAeLaNEdxOnF2fgTHtFmoFSroXZYQ4hIVGd5Tpkxh7969BSNS69Spw8yZMwNemAh9\ndjusWGFmyRILx44ZMBhUHnpIa0Jr1Eia0PTkv6YG+T2fwnPbHbgfeFDvcoQQl6nI8HY6nWzevJl5\n8+ahKAqNGjWiTh3Z7k9c3ObNRvr1i8JuV4iOVunXT2tCu+YaOZ+qC5+PqKVpmHbvwr5gEQC5U2bo\nXJQQ4koVGd7PPvsslSpVolu3bqiqyldffcWECRPOm3cuxLlOnVIYPDgStxvGjXPRu7ebOPm0kW4M\n+/cROywZ8zfb8FeogOH4MfxXVdG7LCFEMRQZ3pmZmcyePbvgcuvWrenRo0dAixKhS1VhxIhIMjMN\nTJ7sZMAAj94llV1+P1GvLiJm6iSU/Hxc9z+IfcZs1IQEvSsTQhRTkeGdn59Pfn4+UVFRAOTl5eFy\nuQJemAhNb75p5uOPTdx5p5d+/SS4daOqlOv+CJZPP8EfF4d9biquhx5BugOFCA9FhnfXrl259957\nadCgAQC7d++WuebigvbtU5g4MYJy5VQWLHBiKHKKgAgYRcHdsjVqZBT2mXNQK1XSuyIhRAlSVLXo\niQzHjx9n9+7dKIpCgwYNqKTDE0FGhr1E7y8hwVbi91kW/bWOHg888EA0339vZPHifDp39updWsgo\nqd9Fw8EDRM+fjWPaSxARoe3uoihl4mhb/p6LT9aw+AKxhgkJtgtef9Ej788++4z9+/fTuHFj2rVr\nV6IFifAyZ46F77838vDDHgnuYPP7iXzjNayTnkXJy8XT9HZcXR9H3voQInwV+te9YMEC0tLSOHXq\nFBMmTGD9+vXBrEuEkB07DMyZY6FqVT8zZjj1LqdMMRw6SLlHH8Q2JgXVbCYndQmux7rrXZYQIsAK\nPfL+4osvePvttzGZTNjtdoYMGUKnTp2CWZsIAQ4HJCdH4ffDggVOZPhe8ESkv4N15DAMuQ5cd3fA\nMWse/spX6V2WECIICg1vi8VSsJe3zWbD5/MFrSgROkaOhD//NJCU5KZFC/kdCSZ/XDyYTOQsWKQd\nbZeBc9tCCE2h4a387Yng75eF2LTJyJIlUK+ej/Hj5eODAaeqRK58C3ebdvgrX4WndVvO7PgZNVbe\n7hCirCk0vPft28fo0aMLvSzzzcu2jAyFESMisVggLc1JRITeFYU3w7Gj2EYMxvLpJzgfehj7ktcB\nJLiFKKMKDe9Ro0add7lZs2YBL0aEBlWFlBRtitqsWXDDDbLJSMD872g75tlxGOw5uNu0I/f5qXpX\nJYTQWaHh3blz52DWIULIW2+Z2bTJRIsWXkaMMHH6tN4VhSfDieNYRwwm4pOP8Vtt2Oe8gvPxHnJu\nWwhR9IQ1Ic61f7/Cs89GEBv71xQ1q94lhS+XC8u2r3C3ao19ziv4q12td0VCiFJCwltcMq8XBg2K\nIi9PIS0tn6pVZXvPkmY4cRwlMxNfgxvxX1ODrI+24Lv2OjnaFkKc55JGMGVlZfHzzz8D4Pdf+vnN\nadOm0bVrV7p168ZPP/10wa95+eWXZZeyEDF3roUdO4x07uzhkUdkilqJUlUi1qwk7s7biO3bA/Lz\nAfBdV1eCWwjxD0WG94YNG+jatSvjxo0DYPLkybzzzjtF3vH27ds5ePAgq1evZurUqUyd+s8mmz/+\n+INvv/32CsoWwbZzp4GXX7ZQpYpMUStxx48T27MbsYMHoHg85CcNgchIvasSQpRiRYb38uXLee+9\n94iLiwNgzJgxrFmzpsg73rZtW8E89Nq1a5OdnY3D4Tjva6ZPn86IESOupG4RRLm52hQ1n09h/nwn\n5cvrXVGYUFUi1q6G+vWJ2PQf3C1acuazbTh795WjbSHERRV5zttmsxXs5Q0QGRmJ2Wwu8o4zMzOp\nX79+weX4+HgyMjKwWrUGp/T0dJo2bUrVqlUvqdC4uGhMJuMlfe2lKmy3FnG+iRNh3z4YMQIeeST6\nH7fLOl4hpxNeng4uF6SmYhk4kAqymcgVk9/D4pM1LL5grWGR4R0XF8e6detwuVzs3r2bDz/8kPj4\n+Mt+oHN3Hj179izp6eksX76ckydPXtL3Z2XlXfZjXoxsf3dp/vtfI2lp0dSr52PEiDwyMs6/Xdbx\nMqkqhoMH8NeoCYBp8XLialYlw5YAp3N1Li50ye9h8ckaFl8wtwQt8mX+pEmT+Pnnn8nNzWXChAm4\nXC6mTJlS5AMmJiaSmZlZcPnUqVMkJCQA8PXXX3PmzBmeeOIJBg8ezO7du5k2bdql/iwiSDIzFYYN\ni8RiUUlNdcpp2GJSMjKI7dOD+LvuwHDwAADeho2gVi19CxNChJwij7xjY2N57rnnLvuOmzdvzoIF\nC+jWrRu7d+8mMTGx4C3zDh060KFDBwCOHDnCuHHjGD9+/GU/hggcVYWRIyPIyDDw3HNOGjSQKWrF\nEfFeOtaxIzGcPo379jv0LkcIEeKKDO9WrVpdcFOSLVu2XPT7brnlFurXr0+3bt1QFIWJEyeSnp6O\nzWajffv2V1ywCI6VK0385z9m7rjDS1KSR+9yQpaSmYl17Egi169DjYrCMWU6+U8PBDm3LYQoBkU9\n92T0BRw9erTg3x6Ph23btuFyuejdu3egaztPIM4jyPmdCztwQKF16xgMBtiyJZerry78V0TW8eJs\nSU8T+e4aPE1vxz5/Ib5adf7xNbKGxSdrWHyyhsUXzHPeRR55/70bvEaNGvTt2zfo4S2C468parm5\nCqmp+RcNblGIvDyI1rryc597Ae/Nt5DfdwAYS/bTEkKIsqvI8N62bdt5l0+cOMGhQ4cCVpDQ14IF\nFr791siDD3ro0kWmqF0uywfvYxs9gpyFS/G0ao3/qirk90/WuywhRJgpMrwXLlxY8G9FUbBarUya\nNCmgRQl9/PCDgZdesnDVVX5mznTKnJDLoJw5jXX8v4hMX4saEYHxyGGkU0AIEShFhvfYsWPPG7Yi\nwlNeHiQnR+L1Ksyfn8//BuqJS2D5zwfYRg3DkHEKT+Mm2Ocv0jYTEUKIACmy5XXGjBnBqEPo7IUX\nIvjjDyP9+7tp1cqndzkhI+K9dMr16o6SfRbHsy9w9v2PJLiFEAFX5JF3lSpV6NGjBzfddNN5Y1GH\nDRsW0MJE8GzebOS11yzUrevjmWdcepcTGlQVFAVXh/twPtqNvKEp+Oper3dVQogyosjwrlatGtWq\nVQtGLUIHp08rDB0aidmssnChk3PG2IsLUM5mYZ0wFu8NDchPHgIREdhTl+hdlhCijCk0vNevX0+n\nTp0YPHhwMOsRQaSqMGpUBKdOGZgwwcWNN8oUtYuxfLwR68hhGE8cx3NbM/IHDpJhK0IIXRT6zLN2\n7dpg1iF0sHq1iQ8+MHP77V4GDXLrXU6ppWSfxTY0iXJPPIbhdCa545/j7LoPJLiFELop8m1zEZ4O\nHlQYPz4Sq1XllVecMj+kEMqpU8S1b4nx+DE8DRthn5+G7wb59IUQQl+FhvfOnTu56667/nG9qqoo\nilLkbHNRevl8MGhQJA6H9rGw6tVlilph1IQEPM2a47z2OvKGpsAl7GUvhBCBVmh433DDDcyePTuY\ntYggeeUVC9u3m7j/fg9du8oUtb8zf/oJlq2fkfvcC6Ao2NNeRSbWCCFKk0LD22Kx/GOuuQh9P/1k\nYMYMC5Uq+Zk1S6aonUux5xDz/ASiVryOajLhfLKntpGILJIQopQptOOmYcOGwaxDBEF+/v9NUZs3\nz0l8vN4VlR7mzz4lrlUzola8jveGBmRt2nLBHcCEEKI0KDS8//WvfwWzDhEEkydHsHevkb593bRp\nI1PU/hIzYQzlH30Qw/Fj5KaMJuujLfhulBevQojSS7rNy4hPPzXy6qsWrr3Wx7PPyhS1c6nxFfDW\nuwH7/DS8N92sdzlCCFEk+aBqGXDmDAwdGonJpJKW5vxrq+myy+Egat7L4NH2/cobMoKsjz6T4BZC\nhAw58g5zqgqjR0dy8qSB8eNdNGxYtqeomb/cim3YIIyHDqBarTj7DpCPfwkhQo6Ed5h75x0T69eb\nadrUy5AhZXiKWm4u1ikTiVq2BNVgIG/YSJxP9ta7KiGEuCIS3mHs8GGFceMiiYlRSU0tu1PUTN98\nTezg/hgPHsB77XXYFyzCe0sTvcsSQogrJuEdpnw+GDw4ErtdYd68fK65puxOUTOczcJw+BB5g4eT\nO3o8REbqXZIQQhSLhHeYWrjQwrZtJjp29NCtW9mbomb65mt8tWqjJiTgvudezmz7Hn/NWnqXJYQQ\nJUK6zcPQzz8bmD7dQmKin5dfdpWtAWF5ecQ8O47yne7BNnZkwdUS3EKIcCJH3mHG6dQ2HfF4tLfL\nK1QoO2+Xm7Z/g23oQEz79+GtVZu8/sl6lySEEAEhR95hZurUCH791chTT7lp27aMTFHLzydm4jOU\nf+BujH/uJ2/AILI2f4n3ttv1rkwIIQJCjrzDyGefGVm82EKdOj4mTiw7U9SMx48StXwpvho1sc9L\nw3t7M71LEkKIgJLwDhNZWf83RW3hwjIwRc3pxHDyBP5rauCrVYfsle/iubkx4f+DCyGEvG0eFlQV\nxoyJ5PhxA6NGuWnUKLynqJm+/464dndS7snHtJP8gKf5nRLcQogyQ8I7DLz7rol//9tMkyY+hg4N\n4ylqLhcxU56nfMd2mPb+hqdFS/CH9wsVIYS4EHnbPMQdOaIwduxfU9TyMYXp/6OmnTuwDU3C9Nuv\n+Kpfg33eQu1oWwghyqAwfaovG/x+GDIkkpwchTlznNSsGaYfC/N6iR3QB+OBP8l/6mkcz74AVqve\nVQkhhG4kvENYWpqZL780ce+9Hh5/3KN3OSVOseeg2mLBZMI+byF4PHha3qV3WUIIoTs55x2idu82\n8OKLESQkhOEUNbeb6OlTiL+1IYZjRwHwNGsuwS2EEP8j4R2CnE5ITo7E7VaYO9dJxYrh83a58eef\niLv7LmJmz0SNisZw8oTeJQkhRKkj4R2Cpk2LYM8eI716uWnfPkymqLndRM+cRtw9d2H6ZRf5PXqT\n9fnXeG9urHdlQghR6sg57xCzdauRRYss1Krl5/nnw2eKmvXZsUQtfxVflarYZy/A06ad3iUJIUSp\nJeEdQs6e1brLjUaVhQvziYnRu6JiUlX+OlmfN2gYqCq5E55HjS2nc2FCCFG6ydvmIWTs2EiOHTMw\ncqSbW24J7eEkxl92U/6euzB//RUA/urX4Jg5R4JbCCEugYR3iEhPN5GebqZxYx/Dh4fwFDWvl+g5\nLxHXviXmH3Zi3vKJ3hUJIUTIkbfNQ8DRowqjR0cSHR3aU9SMv+7BNnQg5h924qt8FY6X5+Fu30Hv\nsoQQIuSEaAyUHX6/tltYTo7CrFlOatUKzY+FmT/fQrnHu6C43Tgf645jynTU8nF6lyWEECFJwruU\nW7LEzNatJu6+20uPHqE7Rc3TpCmeJk3JTxqC+5579S5HCCFCmoR3KbZnj4GpUyOoWNHP7NnO0Jqi\n5vUSlfYKarlyOHs+BdHRZP/7Q72rEkKIsCDhXUq5XJCUFInLpfDqq/kkJobO2+XG3/dq57Z3fIev\nRk2c3Z8Es1nvsoQQImxIt3kpNX16BL/8YqRHDzf33BMiU9R8PqJS5xPXpjnmHd/hfPhRsjZuluAW\nQogSFtAj72nTpvHjjz+iKArjx4+nYcOGBbd9/fXXzJ49G4PBQM2aNZk6dSoGg7yWAPjqKyMLF5qp\nWdPPpEmhMUVNycmmXLdHMH+3HX/FBHIWzcV93wN6lyWEEGEpYGm5fft2Dh48yOrVq5k6dSpTp049\n7/bnnnuO+fPns2rVKnJzc9m6dWugSgkpOTkweHAkBgOkpuaHzLbVqi0Wf3w8zoce5szW7RLcQggR\nQAE78t62bRvt2mnzqWvXrk12djYOhwPr/9IoPT294N/x8fFkZWUFqpSQMnZsJEeOGBg50kWTJqV7\nippx/x+w4lPo0Q8UhZxX34TISL3LEkKIsBewI+/MzEzi4v7vc7zx8fFkZGQUXP4ruE+dOsWXX35J\nq1atAlVKyHjvPRNr15q5+WYfKSmleIqa30/UkoXEtW4OI0di+ukH7XoJbiGECIqgdZur6j+7pU+f\nPs3AgQOZOHHieUF/IXFx0ZhMxhKtKSHBVqL3VxxHj8Lo0RAdDatWGalSpfTUdp4//oA+fWDrVqhQ\nAV5/nbi2d+pdVcgrTb+LoUrWsPhkDYsvWGsYsPBOTEwkMzOz4PKpU6dISEgouOxwOOjXrx/Dhw+n\nRYsWRd5fVlZeidaXkGAjI8Neovd5pfx+ePzxKLKyTMyc6SQuzsM5b1KUGpHLX8U6aQJKXh6u+zph\nnzGbivVrl5p1DFWl6XcxVMkaFp+sYfEFYg0LezEQsLfNmzdvzqZNmwDYvXs3iYmJBW+VA0yfPp1e\nvXrRsmXLQJUQMl591cznn5to395Lr16ld4qa4cQx1IgIcha/Rs5rK1ATE/UuSQghyiRFvdD72SVk\n1qxZfPfddyiKwsSJE/nll1+w2Wy0aNGCW2+9lZtvvrnga++//366du1a6H0F4tVMaXiV+euvBtq3\nj8ZqVdmyJY9KlUrRMBa/n4j163A98BAYjeByoWRnnxfapWUdQ5msYfHJGhafrGHxBfPIO6DnvEeN\nGnXe5euvv77g37t27QrkQ4cEtxuSk7UpakuWOEtVcBsOHcQ2fBCWLz7HMek4+UmDISJCjraFEKIU\nkKkoOpoxw8KuXUaeeMLNvfd69S5Ho6pEvr6MuFbNsHzxOa577sX1cBe9qxJCCHEOmW2uk23bjLzy\nioVrrvEzeXLpmKJmOHwI2/DBWLZuwV+uPDmvLMb1aDdCa0cUIYQIfxLeOvhripqiwMKFpWeKmumH\n77Fs3YKr/T04Xp6Pv/JVepckhBDiAiS8dTB+fCSHDxtISXFx6636TlEzHD2CGh2NGheP+4GHOJu+\nAU/zO+VoWwghSjE551cjum8AABlaSURBVB1k779vYs0aM40a+Rg5UscpaqpK5NtvEtfydqzj/lVw\ntadFSwluIYQo5eTIO4hOnFAYNSqSqCiVhQvzddsp03DsKLaUIVg2/xe/LRZPy7tAVSW0hRAiREh4\nB4nfD0OHRpKVpTB9upM6dXT4WJiqErHqbawTxmKw5+Bu3Rb77AX4q1YLfi1CCCGumIR3kLz2mpkt\nW0y0bevlqaf0maJmOHQQ27+Go1oisM9egPOJnnK0LYQQIUjCOwh++83ACy9EEB/vZ+5cZ3DzUlVR\nzmahxsXjv6YG9tQleBrfir/a1UEsQgghREmS8A6wv6aoOZ0KaWnBnaJmOHEc66hhGI4d4+zGzWCx\n4Hrw4aA9vhBCiMCQbvMAe+klCz//bKR7dw/33RekKWqqSsQ7q4hreRsRH21EjYtHscvMYiGECBdy\n5B1AX39tZMECC9Wr+5k61RmUx1ROnsT2r+FEbPwANToG+8w5OHv1kXPbQggRRiS8A8Ru16aoAaSm\nOoMzRU1VKd+1M6ZfduFu0RL7nFfwX1Pj/7d37/Ex3fkfx1+TmUxSRC42iUtYl1JEq24lLlE2cbet\nNhVZQUVdIrQSiyAkqWtIuohbutVqUeXXTX8tJYhlu9tESllW2J+IIm6RC5rJTSY5vz+yncqKEXKZ\njHyej0cfj858z5zz8dHOe77nzHxPDRxYCCFETZLwriYhIdZcvWrBe+8V0rNncfUerLi49JadKhW6\nxeGof/qJgonvgIVcFRFCiGeRvLtXg717NezcaclLLxUzZ041rqKmKFh99SX2fXugSk8HoGigJwWT\npkhwCyHEM0ze4atYerqKP/7RCmtrhY0bC9Bqq+c4qowMGk4aT8OpfqhvXMfyzKnqOZAQQohaR06b\nVyFFgVmzrMnOtmD58gLatauem45ov/kKm3lBWGRlUdTTjZ/XbqSkdZtqOZYQQojaR2beVeiTTyw5\nfFjDq6/q8fOrnlXU6n2wCtt3JqDKy0O3ZAV3v94vwS2EEHWMzLyryMWLKsLDrbC3V1i3rqDaLjkX\nvP4mlgnfo4uIpLhN2+o5iBBCiFpNZt5VoKgIpk9/jvx8FZGRBTRuXHWrqKmys7CZNgnNj8cBKGnd\nhntffi3BLYQQdZjMvKtAVJSWf/5TzejRRYwcWXWrqGm/3YPNnFlYZGaAWk1Otx5Vtm8hhBDmS8K7\nkn74wYI1a7Q0b17C8uVVs4qaKjuLBgvmYh37PyhWVugWLyHff0aV7FsIIYT5k/CuBJ0OAgKeQ1FK\nV1Fr2LDy+9T86zQNfbxQ306nqFt3ctZuorjdC5XfsRBCiGeGhHclLFpkxZUrFsycWUivXlWzipq+\nVRsUW1t0UwNKZ9sa+SsSQghRliTDU9q3T8OOHVo6dSpm3rzKraKmPbgfVW4uhaO8oEED7hxNBEvL\nKqpUCCHEs0bC+ymkp6uYPdsKK6vKraKmunuHBiHBWO/eSUmjRhQOHgb16klwCyGEMEp+KvaEFAUC\nA63JyrJg0aJC2rd/ulXUtPEHsHfvhfXunRR17sLd2G9Lg1sIIYR4DJl5P6FPP7UkPl6Du7ued955\nilXUCgpoMC+I53ZuR7G0JDc4hLyZgTLbFkIIUWES3k8gNVVFWJgVdnYK0dFPuYqalRXqG9cp6vQS\nOdGbKXbtVOV1CiGEeLZJeFfQL6uo5eWpWLs2nyZNKr6KmirnZ7TxB0u/kKZS8XPMxyg2DWW2LYQQ\n4qnINe8K+uADLadOqfHyKuK11yq+iprl0b9i796LhlP90PyQBIDi0EiCWwghxFOT8K6AEydKV1Fz\ncSlh5cqKraKm0uXQYPZ72I1+HYv0W+TOnof+5S7VXKkQQoi6QE6bP4ZOV3q6vKQE1q+v2Cpqlt8d\nxSZwBuq0q+g7uJITvQn9Sy9Xf7FCCCHqBJl5P0ZoqBWXL1swfXoRvXtXbBU17V/jsbhxndygOdw5\n9DcJbiGEEFVKZt5GxMWp2bZNi6trMcHBhUa31Zz6EX3nLmBhQe68hRS++Rb6FzvXUKVCCCHqEpl5\nP0JGhoqgIGvDKmpWVo/YUKejQfBs7AcPwPrjD0ufe+45CW4hhBDVRmbe5VAUCAqyJjPTgvffL6BD\nh/JXUbNM+Ac2705HffUy+hfao5f7bQshhKgBMvMux/btlhw4oKFfPz1TppSzilpuLvUXzsXu9WFY\nXLtK3sxA7hz6Dn2XbjVfrBBCiDpHZt7/5dIlFYsWWWFr++hV1LSHD1Lvz5vRt21HzrpNMuMWQghR\noyS8H6DXQ0BA6SpqMTH5NG36wCpqeXmoSopRGthwf+Tr/LxuE4WvvwnW1qYrWAghRJ0kp80fsGaN\nlh9/VPPGG0WMGvXrKmqapGPYD+xD/ZDg0idUKgrHjJXgFkIIYRIS3v9x8qQFUVFamjUrISLiP6uo\n5edTf/EC7H4/GPVPl1Ds7KHk6W4BKoQQQlQVOW0O5OaWrqJWXKwiOjofW1vQHE/C5l1/NKkX0bdu\nQ87aTeh79jJ1qUIIIYSEN0BYmBWXLlkwbdp9+vYtRpWejt0bI+D+ffKmBpA7fxHUq2fqMoUQQghA\nTptz6JCaTz/V0qFDMQvm6ABQnJ3RhS/n3tf7yV2yQoJbCCGq0aFDcfTv35O7d+8anlu2LIzvv/97\nme28vEaSl5cHwPnzycycOZWpUyfi5+fLxx9/iKJU/FbNv0hJucC0aX74+/sRGbmi3G0+/3wbb7/9\nB955ZzznzycDcPFiCtOnv8OMGVOYP382BQW/3rRKURT8/f3YsiWmzH4uXbrIq6/24ubNG09c53+r\n0+Gdmali1ixrbCzzies8h8beQ0u/cg4U+E2mqFdvE1cohBDPvkOHDtCsmQtHj8ZXaPvcXB3h4YsI\nDJxDTMwnfPjhVlJSLrBnz/8+8bHXrYvivfdms2nTx+h0OhITvy8zfulSKocPH+Sjjz5jzpwFJCT8\nA4A1a1YzY8Ys1q//EBeXFuzbt8fwmj17/peiorK3jlYUhfXr1+Li0vyJayxPtZ42X758OadPn0al\nUrFgwQJeeuklw1hCQgIffPABarUad3d3AgICqrOUh5SuombFbzN+ZN9vxuP4xb8p/m1LLG5cp6TF\nb2u0FiGEqKt+/vke588nM3/+Yj7//DNef93rsa85dCgOd/f+tG79PAAajYZFi8Kxsir7C6BPP93C\n8eNJZZ6bPTuYVq1aA1BUVMTNmzfo0MEVgD59+nHixA+4ufUxbJ+Q8HcGDvRAo9HwwgvteeGF9gBE\nRHxA/foNALCzs+Pnn+8BcPfuXQ4diuO1197g9u10w36+/fYbunfvYQj/yqq28P7hhx+4cuUKu3bt\nIjU1lQULFrBr1y7D+NKlS9myZQvOzs74+voyePBgnn/++eoq5yFbYwrpFxfOPCJQZ5aQP2kKupBw\nqF+/xmoQQojaYs4c2LWrat//Ro7UExZm/KZOf/1rPL1796VnTzciIpaSkXEbR0cno6+5cuWKIXB/\nUa/ew7VPmDCJCRMmPXI/9+7dxcbGxvDY3t6BrKzMMtvcunUTCwsLgoJmUlysZ8aMQNq2bWcI7vz8\nfOLivmXJkggANm5cy5Qp00lLu1rmOHFx37JmzcYqC+9qO22emJiIh4cHAG3atOHevXvodKXXlNPS\n0rC1taVJkyZYWFjQv39/EhMTq6uUh1y/rqJlwHAWsAJ90+bcjd2LbkWkBLcQQtSw+PgDeHgMRq1W\nM2DA7zh8+KDR7VUqFSoVlJRU7BbNT6K8a+aKolBSUkJU1Dr8/KYSEbHUMJafn09wcBA+PuNo2bIV\nx48fR61W8+J/3Zhq06ZoJk/2R6Opuvlytc28MzMzcXX99ZORg4MDGRkZNGjQgIyMDBwcHMqMpaWl\nGd2fvX09NBp1ldR24QJstphO89+14/nYVVg1aFAl+62rHB1tHr+RMEp6WHnSw8pZvRpWr67q+Zz2\nP/+U79atW5w7d5bNm9cRE6OioKAAGxsbZs70p0kTJ1SqojJ/ryUlxbRo4YSra3tSUlLKjGVnZ5Of\nn0+zZs0Mz23atImEhIQyxwwNDTWc5bWzs0anyzHsp7AwhxYtmpXZr4tLE1q3bo2TU0M8PPoREjIH\nR0cb9Ho9c+bMZNSo13jrrbcA2LLlMBcv/h8BAZPIzs7m/v37tG//PKdOnSAt7TIAFy9eZPHieWzd\nuhU7O7unays1+FOxp/kW4IPu3MmrokqgXTvYkf8Gd+54kpGvQH5Ole27rnF0tCEjQ/pXGdLDypMe\nVp4perhr118YNeotZs4MBEpzYsyYUfzzn+fp2PFl/vKXXbzyijsajYZDh+Lo1KkzGRk5uLkNYPPm\nD3F396Bjx04UFRURHr6QXr16M2LE64b9e3n54uXl+9BxH/xzuri0ID7+73Tu/DJ79+7jzTe9y4y/\n+GJ3vv76L/Ts2Z8rVy7j6OhERkYOn366BVfXzrz66hDD9sHBwYZ/37dvDzdv3qBPn9/Rp8/vDPub\nMWMKCxeGUVSkrlC/H/WhtNrC28nJiczMX68d3L59G0dHx3LH0tPTcXIyfo2jqlXh2QshhBBPIT7+\nACEh4YbHKpWKoUNHEB9/gAkTJnH58iUCAiZjaWlJo0aNCAycC0C9evWIilrLqlXLKSwsRK1W4+k5\npExwV9S7785m9erlKEoJHTt2okePngAEBwexcuUHdOr0IklJCUydOhGAoKB5AMTG/g9NmjTlxIkf\nAOjWrQdz5wZVqh9PQqVUdkr8CCdPniQ6OppPPvmE5ORkli5dys6dOw3jw4cPJyYmhsaNG+Pt7U1k\nZCStWrV65P6q+hOhfFKvGtLHypMeVp70sPKkh5VXHT2s8Zl3165dcXV1ZcyYMahUKkJDQ4mNjcXG\nxgZPT0/CwsKYPXs2AMOGDTMa3EIIIYT4VbXNvKuazLxrJ+lj5UkPK096WHnSw8qryZl3nV5hTQgh\nhDBHEt5CCCGEmZHwFkIIIcyMhLcQQghhZiS8hRBCCDMj4S2EEEKYGQlvIYQQwsxIeAshhBBmxmwW\naRFCCCFEKZl5CyGEEGZGwlsIIYQwMxLeQgghhJmR8BZCCCHMjIS3EEIIYWYkvIUQQggzUyfCe/ny\n5Xh7ezNmzBjOnDlTZiwhIQEvLy+8vb3ZsGGDiSqs/Yz18NixY4wePZoxY8Ywf/58SkpKTFRl7Was\nh7+Iiopi3LhxNVyZ+TDWw5s3b+Lj44OXlxeLFy82UYXmwVgfd+zYgbe3Nz4+PixbtsxEFdZ+Fy5c\nwMPDg+3btz80ViO5ojzjkpKSlClTpiiKoigXL15URo8eXWZ86NChyo0bN5Ti4mLFx8dHSUlJMUWZ\ntdrjeujp6ancvHlTURRFmTlzpnL06NEar7G2e1wPFUVRUlJSFG9vb8XX17emyzMLj+vhu+++qxw8\neFBRFEUJCwtTrl+/XuM1mgNjfczJyVEGDBigFBUVKYqiKBMnTlROnTplkjprs9zcXMXX11cJCQlR\ntm3b9tB4TeTKMz/zTkxMxMPDA4A2bdpw7949dDodAGlpadja2tKkSRMsLCzo378/iYmJpiy3VjLW\nQ4DY2FgaN24MgIODA3fu3DFJnbXZ43oIsHLlSgIDA01Rnlkw1sOSkhJ+/PFHBg4cCEBoaChNmzY1\nWa21mbE+WlpaYmlpSV5eHnq9nvz8fGxtbU1Zbq2k1Wr585//jJOT00NjNZUrz3x4Z2ZmYm9vb3js\n4OBARkYGABkZGTg4OJQ7Jn5lrIcADRo0AOD27dt8//339O/fv8ZrrO0e18PY2FheeeUVmjVrZory\nzIKxHmZnZ1O/fn1WrFiBj48PUVFRpiqz1jPWRysrKwICAvDw8GDAgAF07tyZVq1amarUWkuj0WBt\nbV3uWE3lyjMf3v9NkdVgK628HmZlZTFt2jRCQ0PLvDGI8j3Yw7t37xIbG8vEiRNNWJH5ebCHiqKQ\nnp7O+PHj2b59O+fOnePo0aOmK86MPNhHnU5HTEwMcXFxHD58mNOnT/Pvf//bhNWJR3nmw9vJyYnM\nzEzD49u3b+Po6FjuWHp6ermnQeo6Yz2E0v/hJ0+ezKxZs+jbt68pSqz1jPXw2LFjZGdnM3bsWGbM\nmEFycjLLly83Vam1lrEe2tvb07RpU1q0aIFarcbNzY2UlBRTlVqrGetjamoqzZs3x8HBAa1WS/fu\n3Tl79qypSjVLNZUrz3x49+nThwMHDgCQnJyMk5OT4TSvi4sLOp2Oa9euodfrOXLkCH369DFlubWS\nsR5C6bXaCRMm4O7ubqoSaz1jPRwyZAj79u1j9+7drF+/HldXVxYsWGDKcmslYz3UaDQ0b96cy5cv\nG8bldG/5jPWxWbNmpKamUlBQAMDZs2dp2bKlqUo1SzWVK3XirmKRkZGcOHEClUpFaGgo586dw8bG\nBk9PT44fP05kZCQAgwYNYtKkSSautnZ6VA/79u1Ljx496NKli2HbESNG4O3tbcJqaydj/x3+4tq1\na8yfP59t27aZsNLay1gPr1y5QnBwMIqi0K5dO8LCwrCweObnJ0/FWB+/+OILYmNjUavVdOnShblz\n55q63Frn7NmzREREcP36dTQaDc7OzgwcOBAXF5cay5U6Ed5CCCHEs0Q+lgohhBBmRsJbCCGEMDMS\n3kIIIYSZkfAWQgghzIyEtxBCCGFmNKYuQIi64Nq1awwZMqTMT+oAFixYQIcOHcp9TXR0NHq9vlLr\nnSclJTF9+nQ6duwIQGFhIR07dmThwoVYWlo+0b6+++47kpOT8ff35+TJkzg6OtK8eXOWLVvGa6+9\nRqdOnZ66zujoaGJjY3FxcQFAr9fTuHFj3n//fWxsbB75uvT0dC5duoSbm9tTH1sIcyThLUQNcXBw\nMMnvt9u1a2c4rqIoBAYGsmvXLnx9fZ9oP+7u7oaFeGJjYxk2bBjNmzdn4cKFVVLn73//+zIfVFav\nXs3mzZuZM2fOI1+TlJREamqqhLeocyS8hTCx1NRUQkNDUavV6HQ6Zs2aRb9+/Qzjer2ekJAQfvrp\nJ1QqFR06dCA0NJT79+/z/vvvc+XKFXJzcxkxYgR+fn5Gj6VSqejWrRuXLl0C4OjRo2zYsAFra2ue\ne+45lixZgrOzM5GRkRw7dgytVouzszMRERHs3buXhIQEBg8eTFxcHGfOnGH+/Pls3LgRf39/oqKi\nWLhwIV27dgXg7bffZuLEibRt25bw8HDy8/PJy8sjKCiI3r17P7YvXbp0Yffu3QCcOHGCyMhItFot\nBQUFhIaG0rBhQ9asWYOiKNjZ2TF27Ngn7ocQ5krCWwgTy8zM5L333qNHjx6cOnWKJUuWlAnvCxcu\ncPr0afbv3w/A7t27ycnJYdeuXTg5ObF06VKKi4sZPXo0vXv3pn379o88VmFhIUeOHMHLy4v8/HxC\nQkL48ssvady4Mdu3b2fNmjUEBwezY8cOTpw4gVqtZt++fWXWavb09OSzzz7D398fNzc3Nm7cCMDI\nkSM5cOAAXbt2JSsri9TUVPr27Yu/vz9+fn706tWLjIwMvL29OXjwIBrNo99+9Ho9e/fu5eWXXwZK\nb94SFhZG+/bt2bt3LzExMaxbt45Ro0ah1+uZOHEiH3300RP3QwhzJeEtRA3Jzs5m3LhxZZ5bu3Yt\njo6OrFq1ij/96U8UFRVx9+7dMtu0adMGe3t7Jk+ezIABAxg6dCg2NjYkJSVx69Ytjh8/DsD9+/e5\nevXqQ2F14cKFMscdMGAAw4YN4/z58zRq1MhwL/ZXXnmFL774AltbW/r164evry+enp4MGzbMsI0x\nw4cPx8fHh/nz5xMXF8eQIUNQq9UkJSWRm5vLhg0bgNJ1yLOysnB2di7z+m+++YaTJ0+iKArnzp1j\n/PjxTJkyBYDf/OY3rFq1isLCQnJycsq9x3RF+yHEs0DCW4ga8qhr3rNnz2b48OF4eXlx4cIFpk2b\nVmbcysqKzz//nOTkZMOseefOnWi1WgICAhgyZIjR4z54zftBKpWqzGNFUQzPrVu3jtTUVP72t7/h\n6+tLdHT0Y/98v3yB7cyZM+zfv5/g4GAAtFot0dHRZe5xXJ4Hr3lPmzaNZs2aGWbnc+fOJTw8HDc3\nN44cOcLHH3/80Osr2g8hngXyUzEhTCwzM5O2bdsCsG/fPu7fv19m/F//+hdfffUVrq6uzJgxA1dX\nVy5fvky3bt0Mp9JLSkpYsWLFQ7N2Y1q2bElWVhY3btwAIDExkc6dO5OWlsbWrVtp06YNfn5+eHp6\nPnRPZ5VKRVFR0UP7HDlyJF9++SX37t0zfPv8wTqzs7NZtmzZY2sLDQ0lOjqaW7dulelRcXExcXFx\nhh6pVCr0ev1Dx3mafghhTiS8hTAxPz8/5s6dy6RJk+jWrRu2trasXLnSMN6iRQsOHDjAmDFjGD9+\nPA0bNqRr166MHTuWevXq4e3tzejRo7GxscHOzq7Cx7W2tmbZsmUEBgYybtw4EhMTmTVrFs7Ozpw7\ndw4vLy8mTJjA9evXGTRoUJnX9unTh9DQUA4ePFjm+UGDBrFnzx6GDx9ueG7hwoXEx8fzhz/8gSlT\nptCrV6/H1takSRMmT57MokWLAJg8eTITJkxg2rRpjBo1ips3b7J161a6d+9ObGwsa9asqXQ/hDAn\nclcxIYQQwszIzFsIIYQwMxLeQgghhJmR8BZCCCHMjIS3EEIIYWYkvIUQQggzI+EthBBCmBkJbyGE\nEMLMSHgLIYQQZub/AYxoCVFDGzl5AAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 576x396 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "svm_linearSVC\n",
            "准确率 0.7484232655921513\n",
            "精确率 0.0\n",
            "召回率 0.0\n",
            "F1-score 0.0\n",
            "AUC 0.5437023358684652\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/sklearn/metrics/classification.py:1143: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
            "  'precision', 'predicted', average, warn_for)\n",
            "/usr/local/lib/python3.6/dist-packages/sklearn/metrics/classification.py:1143: UndefinedMetricWarning: F-score is ill-defined and being set to 0.0 due to no predicted samples.\n",
            "  'precision', 'predicted', average, warn_for)\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAe8AAAFnCAYAAACPasF4AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nEA5GircQQgjhYKR4CyGEEA7mL1I/zWEYqd9nAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 576x396 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "rfc\n",
            "准确率 0.7764540995094604\n",
            "精确率 0.5632911392405063\n",
            "召回率 0.4958217270194986\n",
            "F1-score 0.5274074074074074\n",
            "AUC 0.7661771149572786\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAe8AAAFnCAYAAACPasF4AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<Figure size 576x396 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "gbc\n",
            "准确率 0.7694463910301331\n",
            "精确率 0.5451807228915663\n",
            "召回率 0.5041782729805014\n",
            "F1-score 0.5238784370477569\n",
            "AUC 0.7661419047917123\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAe8AAAFnCAYAAACPasF4AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nb7ycxLTdtQAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 576x396 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "xgb\n",
            "准确率 0.7736510161177295\n",
            "精确率 0.5566037735849056\n",
            "召回率 0.49303621169916434\n",
            "F1-score 0.5228951255539144\n",
            "AUC 0.7646396043942287\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAe8AAAFnCAYAAACPasF4AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<Figure size 576x396 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "lgbm\n",
            "准确率 0.7680448493342676\n",
            "精确率 0.547945205479452\n",
            "召回率 0.4456824512534819\n",
            "F1-score 0.4915514592933948\n",
            "AUC 0.7515675044077911\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAe8AAAFnCAYAAACPasF4AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nEgAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 576x396 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "metadata": {
        "id": "ZToRrNL-CdGX",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "# 任务二思考及遇到的问题\n",
        "1. X_train和X_test的数据要严格一致，属性数、缺失值等。所以应该在原始全部数据上做预处理，然后在拆分成训练集、测试集。\n",
        "2. 对数据不平衡处理后，数据标签丢失。导致在xgb中特征名称不匹配，跑不出结果  \n",
        "3. IV没有找到很好的结果办法，待改进  \n",
        "  \n",
        "针对对题2，已解决。  \n",
        "smote方法后，丢失了列名并且数据格式转成了array。  \n",
        "所以将其重新传换为Dataframe格式，并且添加列名。  \n",
        "代码如下：  \n",
        "  \n",
        "```\n",
        "df = pd.DataFrame(over_samples_X)\n",
        "df.columns = list(X_train[1:0])\n",
        "over_samples_X = df\n",
        "over_samples_y = pd.DataFrame(over_samples_y)\n",
        "```"
      ]
    },
    {
      "metadata": {
        "id": "JijuK0MB0-V1",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "# 任务三 模型融合\n",
        "首先介绍一款功能强大的stacking利器，mlxtend库，它可以很快地完成对sklearn模型地stacking。  \n",
        "\n",
        "**StackingClassifier 使用API及参数解析：**\n",
        "  \n",
        "```\n",
        "StackingClassifier(classifiers, meta_classifier, use_probas=False, average_probas=False, verbose=0, use_features_in_secondary=False)\n",
        "```\n",
        "  \n",
        "**参数：**\n",
        "\n",
        "* classifiers : 基分类器，数组形式，[cl1, cl2, cl3]. 每个基分类器的属性被存储在类属性 self.clfs_.  \n",
        "* meta_classifier : 目标分类器，即将前面分类器合起来的分类器  \n",
        "* use_probas : bool (default: False) ，如果设置为True， 那么目标分类器的输入就是前面分类输出的类别概率值而不是类别标签  \n",
        "* average_probas : bool (default: False)，用来设置上一个参数当使用概率值输出的时候是否使用平均值。  \n",
        "* verbose : int, optional (default=0)。用来控制使用过程中的日志输出，当 verbose = 0时，什么也不输出， verbose = 1，输出回归器的序号和名字。verbose = 2，输出详细的参数信息。verbose > 2, 自动将verbose设置为小于2的，verbose -2.  \n",
        "* use_features_in_secondary : bool (default: False). 如果设置为True，那么最终的目标分类器就被基分类器产生的数据和最初的数据集同时训练。如果设置为False，最终的分类器只会使用基分类器产生的数据训练。  \n",
        "  \n",
        "**属性：**  \n",
        "* clfs_ : 每个基分类器的属性，list, shape 为 [n_classifiers]。  \n",
        "* meta_clf_ : 最终目标分类器的属性  \n",
        " "
      ]
    },
    {
      "metadata": {
        "id": "O5YwTre3TL6i",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 691
        },
        "outputId": "c23c7ea4-2d2d-4f05-f354-d7ef28b07769"
      },
      "cell_type": "code",
      "source": [
        "# stacking模型\n",
        "stack = StackingClassifier(classifiers=[lr, dt, gbc, xgb, lgbm], use_probas=True,\n",
        "                          average_probas=False,\n",
        "                          meta_classifier=rfc)\n",
        "\n",
        "stack.fit(X_selected, over_samples_y)\n",
        "multi_score(stack, X_test, Y_test)"
      ],
      "execution_count": 25,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/sklearn/linear_model/logistic.py:433: FutureWarning: Default solver will be changed to 'lbfgs' in 0.22. Specify a solver to silence this warning.\n",
            "  FutureWarning)\n",
            "/usr/local/lib/python3.6/dist-packages/sklearn/utils/validation.py:761: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples, ), for example using ravel().\n",
            "  y = column_or_1d(y, warn=True)\n",
            "/usr/local/lib/python3.6/dist-packages/sklearn/utils/validation.py:761: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples, ), for example using ravel().\n",
            "  y = column_or_1d(y, warn=True)\n",
            "/usr/local/lib/python3.6/dist-packages/sklearn/preprocessing/label.py:219: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples, ), for example using ravel().\n",
            "  y = column_or_1d(y, warn=True)\n",
            "/usr/local/lib/python3.6/dist-packages/sklearn/preprocessing/label.py:252: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples, ), for example using ravel().\n",
            "  y = column_or_1d(y, warn=True)\n",
            "/usr/local/lib/python3.6/dist-packages/mlxtend/classifier/stacking_classification.py:171: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples,), for example using ravel().\n",
            "  self.meta_clf_.fit(meta_features, y)\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "准确率 0.7021723896285914\n",
            "精确率 0.417910447761194\n",
            "召回率 0.467966573816156\n",
            "F1-score 0.44152431011826543\n",
            "AUC 0.6886247691778036\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAe8AAAFnCAYAAACPasF4AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nWwghhLAy0ryFEEIIKyPNWwghhLAy/w+tZZmMNhhVpgAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 576x396 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "metadata": {
        "id": "DmOudaiz9yTv",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "报错：AttributeError: 'LinearSVC' object has no attribute 'predict_proba'    \n",
        "暂时不用这个模型。。。  \n",
        "\n",
        "rfc  \n",
        "准确率 0.7764540995094604  \n",
        "精确率 0.5632911392405063  \n",
        "召回率 0.4958217270194986  \n",
        "F1-score 0.5274074074074074  \n",
        "AUC 0.7661771149572786  \n",
        "\n",
        "stack\n",
        "准确率 0.7021723896285914  \n",
        "精确率 0.417910447761194  \n",
        "召回率 0.467966573816156  \n",
        "F1-score 0.44152431011826543  \n",
        "AUC 0.6886247691778036  \n",
        "\n",
        "除了F1有略微提升，其他指标都不如初始的rfc。  "
      ]
    },
    {
      "metadata": {
        "id": "hj8holEs-vnR",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## 参考资料\n",
        "[stacking 的基本思想及代码实现] https://goo.gl/JkmffM  \n",
        "[StackingClassifier官方文档] https://goo.gl/69gAM8  \n",
        "[如何在 Kaggle 首战中进入前 10%] https://goo.gl/VgjxDq"
      ]
    },
    {
      "metadata": {
        "id": "CyxL2bo7_cK-",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    }
  ]
}