Telecom Churn Assignment.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "from sklearn.impute import KNNImputer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Suppress warnings\n",
    "import warnings\n",
    "warnings.simplefilter(action='ignore')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_csv('telecom_churn_data.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "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>mobile_number</th>\n",
       "      <th>circle_id</th>\n",
       "      <th>loc_og_t2o_mou</th>\n",
       "      <th>std_og_t2o_mou</th>\n",
       "      <th>loc_ic_t2o_mou</th>\n",
       "      <th>last_date_of_month_6</th>\n",
       "      <th>last_date_of_month_7</th>\n",
       "      <th>last_date_of_month_8</th>\n",
       "      <th>last_date_of_month_9</th>\n",
       "      <th>arpu_6</th>\n",
       "      <th>...</th>\n",
       "      <th>sachet_3g_9</th>\n",
       "      <th>fb_user_6</th>\n",
       "      <th>fb_user_7</th>\n",
       "      <th>fb_user_8</th>\n",
       "      <th>fb_user_9</th>\n",
       "      <th>aon</th>\n",
       "      <th>aug_vbc_3g</th>\n",
       "      <th>jul_vbc_3g</th>\n",
       "      <th>jun_vbc_3g</th>\n",
       "      <th>sep_vbc_3g</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>7000842753</td>\n",
       "      <td>109</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>6/30/2014</td>\n",
       "      <td>7/31/2014</td>\n",
       "      <td>8/31/2014</td>\n",
       "      <td>9/30/2014</td>\n",
       "      <td>197.385</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>968</td>\n",
       "      <td>30.4</td>\n",
       "      <td>0.0</td>\n",
       "      <td>101.20</td>\n",
       "      <td>3.58</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>7001865778</td>\n",
       "      <td>109</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>6/30/2014</td>\n",
       "      <td>7/31/2014</td>\n",
       "      <td>8/31/2014</td>\n",
       "      <td>9/30/2014</td>\n",
       "      <td>34.047</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1006</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>7001625959</td>\n",
       "      <td>109</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>6/30/2014</td>\n",
       "      <td>7/31/2014</td>\n",
       "      <td>8/31/2014</td>\n",
       "      <td>9/30/2014</td>\n",
       "      <td>167.690</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1103</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4.17</td>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>7001204172</td>\n",
       "      <td>109</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>6/30/2014</td>\n",
       "      <td>7/31/2014</td>\n",
       "      <td>8/31/2014</td>\n",
       "      <td>9/30/2014</td>\n",
       "      <td>221.338</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2491</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>7000142493</td>\n",
       "      <td>109</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>6/30/2014</td>\n",
       "      <td>7/31/2014</td>\n",
       "      <td>8/31/2014</td>\n",
       "      <td>9/30/2014</td>\n",
       "      <td>261.636</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1526</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 226 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   mobile_number  circle_id  loc_og_t2o_mou  std_og_t2o_mou  loc_ic_t2o_mou  \\\n",
       "0     7000842753        109             0.0             0.0             0.0   \n",
       "1     7001865778        109             0.0             0.0             0.0   \n",
       "2     7001625959        109             0.0             0.0             0.0   \n",
       "3     7001204172        109             0.0             0.0             0.0   \n",
       "4     7000142493        109             0.0             0.0             0.0   \n",
       "\n",
       "  last_date_of_month_6 last_date_of_month_7 last_date_of_month_8  \\\n",
       "0            6/30/2014            7/31/2014            8/31/2014   \n",
       "1            6/30/2014            7/31/2014            8/31/2014   \n",
       "2            6/30/2014            7/31/2014            8/31/2014   \n",
       "3            6/30/2014            7/31/2014            8/31/2014   \n",
       "4            6/30/2014            7/31/2014            8/31/2014   \n",
       "\n",
       "  last_date_of_month_9   arpu_6  ...  sachet_3g_9  fb_user_6  fb_user_7  \\\n",
       "0            9/30/2014  197.385  ...            0        1.0        1.0   \n",
       "1            9/30/2014   34.047  ...            0        NaN        1.0   \n",
       "2            9/30/2014  167.690  ...            0        NaN        NaN   \n",
       "3            9/30/2014  221.338  ...            0        NaN        NaN   \n",
       "4            9/30/2014  261.636  ...            0        0.0        NaN   \n",
       "\n",
       "   fb_user_8  fb_user_9   aon  aug_vbc_3g  jul_vbc_3g  jun_vbc_3g  sep_vbc_3g  \n",
       "0        1.0        NaN   968        30.4         0.0      101.20        3.58  \n",
       "1        1.0        NaN  1006         0.0         0.0        0.00        0.00  \n",
       "2        NaN        1.0  1103         0.0         0.0        4.17        0.00  \n",
       "3        NaN        NaN  2491         0.0         0.0        0.00        0.00  \n",
       "4        NaN        NaN  1526         0.0         0.0        0.00        0.00  \n",
       "\n",
       "[5 rows x 226 columns]"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(99999, 226)"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 99999 entries, 0 to 99998\n",
      "Data columns (total 226 columns):\n",
      " #   Column                    Non-Null Count  Dtype  \n",
      "---  ------                    --------------  -----  \n",
      " 0   mobile_number             99999 non-null  int64  \n",
      " 1   circle_id                 99999 non-null  int64  \n",
      " 2   loc_og_t2o_mou            98981 non-null  float64\n",
      " 3   std_og_t2o_mou            98981 non-null  float64\n",
      " 4   loc_ic_t2o_mou            98981 non-null  float64\n",
      " 5   last_date_of_month_6      99999 non-null  object \n",
      " 6   last_date_of_month_7      99398 non-null  object \n",
      " 7   last_date_of_month_8      98899 non-null  object \n",
      " 8   last_date_of_month_9      98340 non-null  object \n",
      " 9   arpu_6                    99999 non-null  float64\n",
      " 10  arpu_7                    99999 non-null  float64\n",
      " 11  arpu_8                    99999 non-null  float64\n",
      " 12  arpu_9                    99999 non-null  float64\n",
      " 13  onnet_mou_6               96062 non-null  float64\n",
      " 14  onnet_mou_7               96140 non-null  float64\n",
      " 15  onnet_mou_8               94621 non-null  float64\n",
      " 16  onnet_mou_9               92254 non-null  float64\n",
      " 17  offnet_mou_6              96062 non-null  float64\n",
      " 18  offnet_mou_7              96140 non-null  float64\n",
      " 19  offnet_mou_8              94621 non-null  float64\n",
      " 20  offnet_mou_9              92254 non-null  float64\n",
      " 21  roam_ic_mou_6             96062 non-null  float64\n",
      " 22  roam_ic_mou_7             96140 non-null  float64\n",
      " 23  roam_ic_mou_8             94621 non-null  float64\n",
      " 24  roam_ic_mou_9             92254 non-null  float64\n",
      " 25  roam_og_mou_6             96062 non-null  float64\n",
      " 26  roam_og_mou_7             96140 non-null  float64\n",
      " 27  roam_og_mou_8             94621 non-null  float64\n",
      " 28  roam_og_mou_9             92254 non-null  float64\n",
      " 29  loc_og_t2t_mou_6          96062 non-null  float64\n",
      " 30  loc_og_t2t_mou_7          96140 non-null  float64\n",
      " 31  loc_og_t2t_mou_8          94621 non-null  float64\n",
      " 32  loc_og_t2t_mou_9          92254 non-null  float64\n",
      " 33  loc_og_t2m_mou_6          96062 non-null  float64\n",
      " 34  loc_og_t2m_mou_7          96140 non-null  float64\n",
      " 35  loc_og_t2m_mou_8          94621 non-null  float64\n",
      " 36  loc_og_t2m_mou_9          92254 non-null  float64\n",
      " 37  loc_og_t2f_mou_6          96062 non-null  float64\n",
      " 38  loc_og_t2f_mou_7          96140 non-null  float64\n",
      " 39  loc_og_t2f_mou_8          94621 non-null  float64\n",
      " 40  loc_og_t2f_mou_9          92254 non-null  float64\n",
      " 41  loc_og_t2c_mou_6          96062 non-null  float64\n",
      " 42  loc_og_t2c_mou_7          96140 non-null  float64\n",
      " 43  loc_og_t2c_mou_8          94621 non-null  float64\n",
      " 44  loc_og_t2c_mou_9          92254 non-null  float64\n",
      " 45  loc_og_mou_6              96062 non-null  float64\n",
      " 46  loc_og_mou_7              96140 non-null  float64\n",
      " 47  loc_og_mou_8              94621 non-null  float64\n",
      " 48  loc_og_mou_9              92254 non-null  float64\n",
      " 49  std_og_t2t_mou_6          96062 non-null  float64\n",
      " 50  std_og_t2t_mou_7          96140 non-null  float64\n",
      " 51  std_og_t2t_mou_8          94621 non-null  float64\n",
      " 52  std_og_t2t_mou_9          92254 non-null  float64\n",
      " 53  std_og_t2m_mou_6          96062 non-null  float64\n",
      " 54  std_og_t2m_mou_7          96140 non-null  float64\n",
      " 55  std_og_t2m_mou_8          94621 non-null  float64\n",
      " 56  std_og_t2m_mou_9          92254 non-null  float64\n",
      " 57  std_og_t2f_mou_6          96062 non-null  float64\n",
      " 58  std_og_t2f_mou_7          96140 non-null  float64\n",
      " 59  std_og_t2f_mou_8          94621 non-null  float64\n",
      " 60  std_og_t2f_mou_9          92254 non-null  float64\n",
      " 61  std_og_t2c_mou_6          96062 non-null  float64\n",
      " 62  std_og_t2c_mou_7          96140 non-null  float64\n",
      " 63  std_og_t2c_mou_8          94621 non-null  float64\n",
      " 64  std_og_t2c_mou_9          92254 non-null  float64\n",
      " 65  std_og_mou_6              96062 non-null  float64\n",
      " 66  std_og_mou_7              96140 non-null  float64\n",
      " 67  std_og_mou_8              94621 non-null  float64\n",
      " 68  std_og_mou_9              92254 non-null  float64\n",
      " 69  isd_og_mou_6              96062 non-null  float64\n",
      " 70  isd_og_mou_7              96140 non-null  float64\n",
      " 71  isd_og_mou_8              94621 non-null  float64\n",
      " 72  isd_og_mou_9              92254 non-null  float64\n",
      " 73  spl_og_mou_6              96062 non-null  float64\n",
      " 74  spl_og_mou_7              96140 non-null  float64\n",
      " 75  spl_og_mou_8              94621 non-null  float64\n",
      " 76  spl_og_mou_9              92254 non-null  float64\n",
      " 77  og_others_6               96062 non-null  float64\n",
      " 78  og_others_7               96140 non-null  float64\n",
      " 79  og_others_8               94621 non-null  float64\n",
      " 80  og_others_9               92254 non-null  float64\n",
      " 81  total_og_mou_6            99999 non-null  float64\n",
      " 82  total_og_mou_7            99999 non-null  float64\n",
      " 83  total_og_mou_8            99999 non-null  float64\n",
      " 84  total_og_mou_9            99999 non-null  float64\n",
      " 85  loc_ic_t2t_mou_6          96062 non-null  float64\n",
      " 86  loc_ic_t2t_mou_7          96140 non-null  float64\n",
      " 87  loc_ic_t2t_mou_8          94621 non-null  float64\n",
      " 88  loc_ic_t2t_mou_9          92254 non-null  float64\n",
      " 89  loc_ic_t2m_mou_6          96062 non-null  float64\n",
      " 90  loc_ic_t2m_mou_7          96140 non-null  float64\n",
      " 91  loc_ic_t2m_mou_8          94621 non-null  float64\n",
      " 92  loc_ic_t2m_mou_9          92254 non-null  float64\n",
      " 93  loc_ic_t2f_mou_6          96062 non-null  float64\n",
      " 94  loc_ic_t2f_mou_7          96140 non-null  float64\n",
      " 95  loc_ic_t2f_mou_8          94621 non-null  float64\n",
      " 96  loc_ic_t2f_mou_9          92254 non-null  float64\n",
      " 97  loc_ic_mou_6              96062 non-null  float64\n",
      " 98  loc_ic_mou_7              96140 non-null  float64\n",
      " 99  loc_ic_mou_8              94621 non-null  float64\n",
      " 100 loc_ic_mou_9              92254 non-null  float64\n",
      " 101 std_ic_t2t_mou_6          96062 non-null  float64\n",
      " 102 std_ic_t2t_mou_7          96140 non-null  float64\n",
      " 103 std_ic_t2t_mou_8          94621 non-null  float64\n",
      " 104 std_ic_t2t_mou_9          92254 non-null  float64\n",
      " 105 std_ic_t2m_mou_6          96062 non-null  float64\n",
      " 106 std_ic_t2m_mou_7          96140 non-null  float64\n",
      " 107 std_ic_t2m_mou_8          94621 non-null  float64\n",
      " 108 std_ic_t2m_mou_9          92254 non-null  float64\n",
      " 109 std_ic_t2f_mou_6          96062 non-null  float64\n",
      " 110 std_ic_t2f_mou_7          96140 non-null  float64\n",
      " 111 std_ic_t2f_mou_8          94621 non-null  float64\n",
      " 112 std_ic_t2f_mou_9          92254 non-null  float64\n",
      " 113 std_ic_t2o_mou_6          96062 non-null  float64\n",
      " 114 std_ic_t2o_mou_7          96140 non-null  float64\n",
      " 115 std_ic_t2o_mou_8          94621 non-null  float64\n",
      " 116 std_ic_t2o_mou_9          92254 non-null  float64\n",
      " 117 std_ic_mou_6              96062 non-null  float64\n",
      " 118 std_ic_mou_7              96140 non-null  float64\n",
      " 119 std_ic_mou_8              94621 non-null  float64\n",
      " 120 std_ic_mou_9              92254 non-null  float64\n",
      " 121 total_ic_mou_6            99999 non-null  float64\n",
      " 122 total_ic_mou_7            99999 non-null  float64\n",
      " 123 total_ic_mou_8            99999 non-null  float64\n",
      " 124 total_ic_mou_9            99999 non-null  float64\n",
      " 125 spl_ic_mou_6              96062 non-null  float64\n",
      " 126 spl_ic_mou_7              96140 non-null  float64\n",
      " 127 spl_ic_mou_8              94621 non-null  float64\n",
      " 128 spl_ic_mou_9              92254 non-null  float64\n",
      " 129 isd_ic_mou_6              96062 non-null  float64\n",
      " 130 isd_ic_mou_7              96140 non-null  float64\n",
      " 131 isd_ic_mou_8              94621 non-null  float64\n",
      " 132 isd_ic_mou_9              92254 non-null  float64\n",
      " 133 ic_others_6               96062 non-null  float64\n",
      " 134 ic_others_7               96140 non-null  float64\n",
      " 135 ic_others_8               94621 non-null  float64\n",
      " 136 ic_others_9               92254 non-null  float64\n",
      " 137 total_rech_num_6          99999 non-null  int64  \n",
      " 138 total_rech_num_7          99999 non-null  int64  \n",
      " 139 total_rech_num_8          99999 non-null  int64  \n",
      " 140 total_rech_num_9          99999 non-null  int64  \n",
      " 141 total_rech_amt_6          99999 non-null  int64  \n",
      " 142 total_rech_amt_7          99999 non-null  int64  \n",
      " 143 total_rech_amt_8          99999 non-null  int64  \n",
      " 144 total_rech_amt_9          99999 non-null  int64  \n",
      " 145 max_rech_amt_6            99999 non-null  int64  \n",
      " 146 max_rech_amt_7            99999 non-null  int64  \n",
      " 147 max_rech_amt_8            99999 non-null  int64  \n",
      " 148 max_rech_amt_9            99999 non-null  int64  \n",
      " 149 date_of_last_rech_6       98392 non-null  object \n",
      " 150 date_of_last_rech_7       98232 non-null  object \n",
      " 151 date_of_last_rech_8       96377 non-null  object \n",
      " 152 date_of_last_rech_9       95239 non-null  object \n",
      " 153 last_day_rch_amt_6        99999 non-null  int64  \n",
      " 154 last_day_rch_amt_7        99999 non-null  int64  \n",
      " 155 last_day_rch_amt_8        99999 non-null  int64  \n",
      " 156 last_day_rch_amt_9        99999 non-null  int64  \n",
      " 157 date_of_last_rech_data_6  25153 non-null  object \n",
      " 158 date_of_last_rech_data_7  25571 non-null  object \n",
      " 159 date_of_last_rech_data_8  26339 non-null  object \n",
      " 160 date_of_last_rech_data_9  25922 non-null  object \n",
      " 161 total_rech_data_6         25153 non-null  float64\n",
      " 162 total_rech_data_7         25571 non-null  float64\n",
      " 163 total_rech_data_8         26339 non-null  float64\n",
      " 164 total_rech_data_9         25922 non-null  float64\n",
      " 165 max_rech_data_6           25153 non-null  float64\n",
      " 166 max_rech_data_7           25571 non-null  float64\n",
      " 167 max_rech_data_8           26339 non-null  float64\n",
      " 168 max_rech_data_9           25922 non-null  float64\n",
      " 169 count_rech_2g_6           25153 non-null  float64\n",
      " 170 count_rech_2g_7           25571 non-null  float64\n",
      " 171 count_rech_2g_8           26339 non-null  float64\n",
      " 172 count_rech_2g_9           25922 non-null  float64\n",
      " 173 count_rech_3g_6           25153 non-null  float64\n",
      " 174 count_rech_3g_7           25571 non-null  float64\n",
      " 175 count_rech_3g_8           26339 non-null  float64\n",
      " 176 count_rech_3g_9           25922 non-null  float64\n",
      " 177 av_rech_amt_data_6        25153 non-null  float64\n",
      " 178 av_rech_amt_data_7        25571 non-null  float64\n",
      " 179 av_rech_amt_data_8        26339 non-null  float64\n",
      " 180 av_rech_amt_data_9        25922 non-null  float64\n",
      " 181 vol_2g_mb_6               99999 non-null  float64\n",
      " 182 vol_2g_mb_7               99999 non-null  float64\n",
      " 183 vol_2g_mb_8               99999 non-null  float64\n",
      " 184 vol_2g_mb_9               99999 non-null  float64\n",
      " 185 vol_3g_mb_6               99999 non-null  float64\n",
      " 186 vol_3g_mb_7               99999 non-null  float64\n",
      " 187 vol_3g_mb_8               99999 non-null  float64\n",
      " 188 vol_3g_mb_9               99999 non-null  float64\n",
      " 189 arpu_3g_6                 25153 non-null  float64\n",
      " 190 arpu_3g_7                 25571 non-null  float64\n",
      " 191 arpu_3g_8                 26339 non-null  float64\n",
      " 192 arpu_3g_9                 25922 non-null  float64\n",
      " 193 arpu_2g_6                 25153 non-null  float64\n",
      " 194 arpu_2g_7                 25571 non-null  float64\n",
      " 195 arpu_2g_8                 26339 non-null  float64\n",
      " 196 arpu_2g_9                 25922 non-null  float64\n",
      " 197 night_pck_user_6          25153 non-null  float64\n",
      " 198 night_pck_user_7          25571 non-null  float64\n",
      " 199 night_pck_user_8          26339 non-null  float64\n",
      " 200 night_pck_user_9          25922 non-null  float64\n",
      " 201 monthly_2g_6              99999 non-null  int64  \n",
      " 202 monthly_2g_7              99999 non-null  int64  \n",
      " 203 monthly_2g_8              99999 non-null  int64  \n",
      " 204 monthly_2g_9              99999 non-null  int64  \n",
      " 205 sachet_2g_6               99999 non-null  int64  \n",
      " 206 sachet_2g_7               99999 non-null  int64  \n",
      " 207 sachet_2g_8               99999 non-null  int64  \n",
      " 208 sachet_2g_9               99999 non-null  int64  \n",
      " 209 monthly_3g_6              99999 non-null  int64  \n",
      " 210 monthly_3g_7              99999 non-null  int64  \n",
      " 211 monthly_3g_8              99999 non-null  int64  \n",
      " 212 monthly_3g_9              99999 non-null  int64  \n",
      " 213 sachet_3g_6               99999 non-null  int64  \n",
      " 214 sachet_3g_7               99999 non-null  int64  \n",
      " 215 sachet_3g_8               99999 non-null  int64  \n",
      " 216 sachet_3g_9               99999 non-null  int64  \n",
      " 217 fb_user_6                 25153 non-null  float64\n",
      " 218 fb_user_7                 25571 non-null  float64\n",
      " 219 fb_user_8                 26339 non-null  float64\n",
      " 220 fb_user_9                 25922 non-null  float64\n",
      " 221 aon                       99999 non-null  int64  \n",
      " 222 aug_vbc_3g                99999 non-null  float64\n",
      " 223 jul_vbc_3g                99999 non-null  float64\n",
      " 224 jun_vbc_3g                99999 non-null  float64\n",
      " 225 sep_vbc_3g                99999 non-null  float64\n",
      "dtypes: float64(179), int64(35), object(12)\n",
      "memory usage: 172.4+ MB\n"
     ]
    }
   ],
   "source": [
    "df.info(verbose=True, null_counts=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Data Cleaning"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "mobile_number        0\n",
       "circle_id            0\n",
       "loc_og_t2o_mou    1018\n",
       "std_og_t2o_mou    1018\n",
       "loc_ic_t2o_mou    1018\n",
       "                  ... \n",
       "aon                  0\n",
       "aug_vbc_3g           0\n",
       "jul_vbc_3g           0\n",
       "jun_vbc_3g           0\n",
       "sep_vbc_3g           0\n",
       "Length: 226, dtype: int64"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "# function to sort missing percentages\n",
    "def sortMissing(data, size):\n",
    "    missing_df = round((data.isnull().sum() * 100/ len(data)),2).sort_values(ascending=False)\n",
    "    print(missing_df.head(size))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "count_rech_2g_6             74.85\n",
      "date_of_last_rech_data_6    74.85\n",
      "count_rech_3g_6             74.85\n",
      "av_rech_amt_data_6          74.85\n",
      "max_rech_data_6             74.85\n",
      "total_rech_data_6           74.85\n",
      "arpu_3g_6                   74.85\n",
      "arpu_2g_6                   74.85\n",
      "night_pck_user_6            74.85\n",
      "fb_user_6                   74.85\n",
      "arpu_3g_7                   74.43\n",
      "count_rech_2g_7             74.43\n",
      "fb_user_7                   74.43\n",
      "count_rech_3g_7             74.43\n",
      "arpu_2g_7                   74.43\n",
      "av_rech_amt_data_7          74.43\n",
      "max_rech_data_7             74.43\n",
      "night_pck_user_7            74.43\n",
      "total_rech_data_7           74.43\n",
      "date_of_last_rech_data_7    74.43\n",
      "night_pck_user_9            74.08\n",
      "date_of_last_rech_data_9    74.08\n",
      "fb_user_9                   74.08\n",
      "arpu_2g_9                   74.08\n",
      "max_rech_data_9             74.08\n",
      "arpu_3g_9                   74.08\n",
      "total_rech_data_9           74.08\n",
      "av_rech_amt_data_9          74.08\n",
      "count_rech_3g_9             74.08\n",
      "count_rech_2g_9             74.08\n",
      "fb_user_8                   73.66\n",
      "av_rech_amt_data_8          73.66\n",
      "count_rech_3g_8             73.66\n",
      "count_rech_2g_8             73.66\n",
      "date_of_last_rech_data_8    73.66\n",
      "total_rech_data_8           73.66\n",
      "max_rech_data_8             73.66\n",
      "arpu_3g_8                   73.66\n",
      "arpu_2g_8                   73.66\n",
      "night_pck_user_8            73.66\n",
      "std_ic_t2m_mou_9             7.75\n",
      "spl_ic_mou_9                 7.75\n",
      "loc_ic_mou_9                 7.75\n",
      "isd_ic_mou_9                 7.75\n",
      "std_ic_t2o_mou_9             7.75\n",
      "loc_ic_t2f_mou_9             7.75\n",
      "ic_others_9                  7.75\n",
      "loc_og_t2f_mou_9             7.75\n",
      "loc_ic_t2m_mou_9             7.75\n",
      "loc_og_t2c_mou_9             7.75\n",
      "dtype: float64\n"
     ]
    }
   ],
   "source": [
    "# get the percentage of missing values\n",
    "sortMissing(df, 50)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "# dropping all columns having high percentages of missing values >70%\n",
    "for column in df:\n",
    "    col = df[column]\n",
    "    if col.isnull().sum() * 100 / len(df) > 70.0:\n",
    "        df.drop([column], axis=1, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 99999 entries, 0 to 99998\n",
      "Columns: 186 entries, mobile_number to sep_vbc_3g\n",
      "dtypes: float64(143), int64(35), object(8)\n",
      "memory usage: 141.9+ MB\n"
     ]
    }
   ],
   "source": [
    "df.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "loc_ic_t2m_mou_9    7.75\n",
      "std_og_t2t_mou_9    7.75\n",
      "std_ic_t2f_mou_9    7.75\n",
      "std_ic_t2t_mou_9    7.75\n",
      "roam_og_mou_9       7.75\n",
      "loc_ic_mou_9        7.75\n",
      "loc_og_t2t_mou_9    7.75\n",
      "loc_ic_t2f_mou_9    7.75\n",
      "roam_ic_mou_9       7.75\n",
      "ic_others_9         7.75\n",
      "dtype: float64\n"
     ]
    }
   ],
   "source": [
    "sortMissing(df, 10)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Dropping columns having single value"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(16,)"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.columns[df.nunique() <= 1].shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "for column in df.columns[df.nunique() <= 1]:\n",
    "    df.drop([column], axis=1, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 99999 entries, 0 to 99998\n",
      "Columns: 170 entries, mobile_number to sep_vbc_3g\n",
      "dtypes: float64(132), int64(34), object(4)\n",
      "memory usage: 129.7+ MB\n"
     ]
    }
   ],
   "source": [
    "df.info()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Filter High Value customers"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "368.5"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "avg_rech_amt = (df.total_rech_amt_6 + df.total_rech_amt_7) / 2\n",
    "df['avg_rech_amt'] = avg_rech_amt\n",
    "df['avg_rech_amt'].quantile(0.7)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_high_val = df[df['avg_rech_amt'] >= df['avg_rech_amt'].quantile(0.7)]\n",
    "df_high_val.reset_index(inplace=True, drop=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(30011, 171)"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_high_val.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "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>mobile_number</th>\n",
       "      <th>arpu_6</th>\n",
       "      <th>arpu_7</th>\n",
       "      <th>arpu_8</th>\n",
       "      <th>arpu_9</th>\n",
       "      <th>onnet_mou_6</th>\n",
       "      <th>onnet_mou_7</th>\n",
       "      <th>onnet_mou_8</th>\n",
       "      <th>onnet_mou_9</th>\n",
       "      <th>offnet_mou_6</th>\n",
       "      <th>...</th>\n",
       "      <th>sachet_3g_6</th>\n",
       "      <th>sachet_3g_7</th>\n",
       "      <th>sachet_3g_8</th>\n",
       "      <th>sachet_3g_9</th>\n",
       "      <th>aon</th>\n",
       "      <th>aug_vbc_3g</th>\n",
       "      <th>jul_vbc_3g</th>\n",
       "      <th>jun_vbc_3g</th>\n",
       "      <th>sep_vbc_3g</th>\n",
       "      <th>avg_rech_amt</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>7000701601</td>\n",
       "      <td>1069.180</td>\n",
       "      <td>1349.850</td>\n",
       "      <td>3171.480</td>\n",
       "      <td>500.000</td>\n",
       "      <td>57.84</td>\n",
       "      <td>54.68</td>\n",
       "      <td>52.29</td>\n",
       "      <td>NaN</td>\n",
       "      <td>453.43</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>802</td>\n",
       "      <td>57.74</td>\n",
       "      <td>19.38</td>\n",
       "      <td>18.74</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1185.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>7001524846</td>\n",
       "      <td>378.721</td>\n",
       "      <td>492.223</td>\n",
       "      <td>137.362</td>\n",
       "      <td>166.787</td>\n",
       "      <td>413.69</td>\n",
       "      <td>351.03</td>\n",
       "      <td>35.08</td>\n",
       "      <td>33.46</td>\n",
       "      <td>94.66</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>315</td>\n",
       "      <td>21.03</td>\n",
       "      <td>910.65</td>\n",
       "      <td>122.16</td>\n",
       "      <td>0.0</td>\n",
       "      <td>519.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>7002191713</td>\n",
       "      <td>492.846</td>\n",
       "      <td>205.671</td>\n",
       "      <td>593.260</td>\n",
       "      <td>322.732</td>\n",
       "      <td>501.76</td>\n",
       "      <td>108.39</td>\n",
       "      <td>534.24</td>\n",
       "      <td>244.81</td>\n",
       "      <td>413.31</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>2607</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.0</td>\n",
       "      <td>380.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>7000875565</td>\n",
       "      <td>430.975</td>\n",
       "      <td>299.869</td>\n",
       "      <td>187.894</td>\n",
       "      <td>206.490</td>\n",
       "      <td>50.51</td>\n",
       "      <td>74.01</td>\n",
       "      <td>70.61</td>\n",
       "      <td>31.34</td>\n",
       "      <td>296.29</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>511</td>\n",
       "      <td>0.00</td>\n",
       "      <td>2.45</td>\n",
       "      <td>21.89</td>\n",
       "      <td>0.0</td>\n",
       "      <td>459.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>7000187447</td>\n",
       "      <td>690.008</td>\n",
       "      <td>18.980</td>\n",
       "      <td>25.499</td>\n",
       "      <td>257.583</td>\n",
       "      <td>1185.91</td>\n",
       "      <td>9.28</td>\n",
       "      <td>7.79</td>\n",
       "      <td>558.51</td>\n",
       "      <td>61.64</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>667</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.0</td>\n",
       "      <td>408.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 171 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   mobile_number    arpu_6    arpu_7    arpu_8   arpu_9  onnet_mou_6  \\\n",
       "0     7000701601  1069.180  1349.850  3171.480  500.000        57.84   \n",
       "1     7001524846   378.721   492.223   137.362  166.787       413.69   \n",
       "2     7002191713   492.846   205.671   593.260  322.732       501.76   \n",
       "3     7000875565   430.975   299.869   187.894  206.490        50.51   \n",
       "4     7000187447   690.008    18.980    25.499  257.583      1185.91   \n",
       "\n",
       "   onnet_mou_7  onnet_mou_8  onnet_mou_9  offnet_mou_6  ...  sachet_3g_6  \\\n",
       "0        54.68        52.29          NaN        453.43  ...            0   \n",
       "1       351.03        35.08        33.46         94.66  ...            0   \n",
       "2       108.39       534.24       244.81        413.31  ...            0   \n",
       "3        74.01        70.61        31.34        296.29  ...            0   \n",
       "4         9.28         7.79       558.51         61.64  ...            0   \n",
       "\n",
       "   sachet_3g_7  sachet_3g_8  sachet_3g_9   aon  aug_vbc_3g  jul_vbc_3g  \\\n",
       "0            0            0            0   802       57.74       19.38   \n",
       "1            0            0            0   315       21.03      910.65   \n",
       "2            0            0            0  2607        0.00        0.00   \n",
       "3            0            0            0   511        0.00        2.45   \n",
       "4            0            0            0   667        0.00        0.00   \n",
       "\n",
       "   jun_vbc_3g  sep_vbc_3g  avg_rech_amt  \n",
       "0       18.74         0.0        1185.0  \n",
       "1      122.16         0.0         519.0  \n",
       "2        0.00         0.0         380.0  \n",
       "3       21.89         0.0         459.0  \n",
       "4        0.00         0.0         408.0  \n",
       "\n",
       "[5 rows x 171 columns]"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_high_val.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Derive Churn"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['arpu_9', 'onnet_mou_9', 'offnet_mou_9', 'roam_ic_mou_9',\n",
       "       'roam_og_mou_9', 'loc_og_t2t_mou_9', 'loc_og_t2m_mou_9',\n",
       "       'loc_og_t2f_mou_9', 'loc_og_t2c_mou_9', 'loc_og_mou_9',\n",
       "       'std_og_t2t_mou_9', 'std_og_t2m_mou_9', 'std_og_t2f_mou_9',\n",
       "       'std_og_mou_9', 'isd_og_mou_9', 'spl_og_mou_9', 'og_others_9',\n",
       "       'total_og_mou_9', 'loc_ic_t2t_mou_9', 'loc_ic_t2m_mou_9',\n",
       "       'loc_ic_t2f_mou_9', 'loc_ic_mou_9', 'std_ic_t2t_mou_9',\n",
       "       'std_ic_t2m_mou_9', 'std_ic_t2f_mou_9', 'std_ic_mou_9',\n",
       "       'total_ic_mou_9', 'spl_ic_mou_9', 'isd_ic_mou_9', 'ic_others_9',\n",
       "       'total_rech_num_9', 'total_rech_amt_9', 'max_rech_amt_9',\n",
       "       'date_of_last_rech_9', 'last_day_rch_amt_9', 'vol_2g_mb_9',\n",
       "       'vol_3g_mb_9', 'monthly_2g_9', 'sachet_2g_9', 'monthly_3g_9',\n",
       "       'sachet_3g_9'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_high_val.filter(regex=(\"._9\")).columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.00       3249\n",
       "0.21         20\n",
       "1.01          7\n",
       "0.61          7\n",
       "5.93          6\n",
       "           ... \n",
       "345.84        1\n",
       "97.94         1\n",
       "1391.18       1\n",
       "248.33        1\n",
       "1281.78       1\n",
       "Name: total_og_mou_9, Length: 22615, dtype: int64"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_high_val.total_og_mou_9.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    27302\n",
       "1     2709\n",
       "Name: churn, dtype: int64"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# total_og_mou_9, total_ic_mou_9, monthly_2g_9, monthly_3g_9\n",
    "df_high_val['churn'] = ((df_high_val['total_og_mou_9'] < 1.0) &  (df_high_val['total_ic_mou_9'] < 1.0) & (df_high_val['monthly_2g_9'] < 1.0) & (df_high_val['monthly_3g_9'] < 1.0)).astype(int)\n",
    "df_high_val['churn'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "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>mobile_number</th>\n",
       "      <th>arpu_6</th>\n",
       "      <th>arpu_7</th>\n",
       "      <th>arpu_8</th>\n",
       "      <th>arpu_9</th>\n",
       "      <th>onnet_mou_6</th>\n",
       "      <th>onnet_mou_7</th>\n",
       "      <th>onnet_mou_8</th>\n",
       "      <th>onnet_mou_9</th>\n",
       "      <th>offnet_mou_6</th>\n",
       "      <th>...</th>\n",
       "      <th>sachet_3g_7</th>\n",
       "      <th>sachet_3g_8</th>\n",
       "      <th>sachet_3g_9</th>\n",
       "      <th>aon</th>\n",
       "      <th>aug_vbc_3g</th>\n",
       "      <th>jul_vbc_3g</th>\n",
       "      <th>jun_vbc_3g</th>\n",
       "      <th>sep_vbc_3g</th>\n",
       "      <th>avg_rech_amt</th>\n",
       "      <th>churn</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>7000701601</td>\n",
       "      <td>1069.180</td>\n",
       "      <td>1349.850</td>\n",
       "      <td>3171.480</td>\n",
       "      <td>500.000</td>\n",
       "      <td>57.84</td>\n",
       "      <td>54.68</td>\n",
       "      <td>52.29</td>\n",
       "      <td>NaN</td>\n",
       "      <td>453.43</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>802</td>\n",
       "      <td>57.74</td>\n",
       "      <td>19.38</td>\n",
       "      <td>18.74</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1185.0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>7001524846</td>\n",
       "      <td>378.721</td>\n",
       "      <td>492.223</td>\n",
       "      <td>137.362</td>\n",
       "      <td>166.787</td>\n",
       "      <td>413.69</td>\n",
       "      <td>351.03</td>\n",
       "      <td>35.08</td>\n",
       "      <td>33.46</td>\n",
       "      <td>94.66</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>315</td>\n",
       "      <td>21.03</td>\n",
       "      <td>910.65</td>\n",
       "      <td>122.16</td>\n",
       "      <td>0.0</td>\n",
       "      <td>519.0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>7002191713</td>\n",
       "      <td>492.846</td>\n",
       "      <td>205.671</td>\n",
       "      <td>593.260</td>\n",
       "      <td>322.732</td>\n",
       "      <td>501.76</td>\n",
       "      <td>108.39</td>\n",
       "      <td>534.24</td>\n",
       "      <td>244.81</td>\n",
       "      <td>413.31</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>2607</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.0</td>\n",
       "      <td>380.0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>7000875565</td>\n",
       "      <td>430.975</td>\n",
       "      <td>299.869</td>\n",
       "      <td>187.894</td>\n",
       "      <td>206.490</td>\n",
       "      <td>50.51</td>\n",
       "      <td>74.01</td>\n",
       "      <td>70.61</td>\n",
       "      <td>31.34</td>\n",
       "      <td>296.29</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>511</td>\n",
       "      <td>0.00</td>\n",
       "      <td>2.45</td>\n",
       "      <td>21.89</td>\n",
       "      <td>0.0</td>\n",
       "      <td>459.0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>7000187447</td>\n",
       "      <td>690.008</td>\n",
       "      <td>18.980</td>\n",
       "      <td>25.499</td>\n",
       "      <td>257.583</td>\n",
       "      <td>1185.91</td>\n",
       "      <td>9.28</td>\n",
       "      <td>7.79</td>\n",
       "      <td>558.51</td>\n",
       "      <td>61.64</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>667</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.0</td>\n",
       "      <td>408.0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 172 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   mobile_number    arpu_6    arpu_7    arpu_8   arpu_9  onnet_mou_6  \\\n",
       "0     7000701601  1069.180  1349.850  3171.480  500.000        57.84   \n",
       "1     7001524846   378.721   492.223   137.362  166.787       413.69   \n",
       "2     7002191713   492.846   205.671   593.260  322.732       501.76   \n",
       "3     7000875565   430.975   299.869   187.894  206.490        50.51   \n",
       "4     7000187447   690.008    18.980    25.499  257.583      1185.91   \n",
       "\n",
       "   onnet_mou_7  onnet_mou_8  onnet_mou_9  offnet_mou_6  ...  sachet_3g_7  \\\n",
       "0        54.68        52.29          NaN        453.43  ...            0   \n",
       "1       351.03        35.08        33.46         94.66  ...            0   \n",
       "2       108.39       534.24       244.81        413.31  ...            0   \n",
       "3        74.01        70.61        31.34        296.29  ...            0   \n",
       "4         9.28         7.79       558.51         61.64  ...            0   \n",
       "\n",
       "   sachet_3g_8  sachet_3g_9   aon  aug_vbc_3g  jul_vbc_3g  jun_vbc_3g  \\\n",
       "0            0            0   802       57.74       19.38       18.74   \n",
       "1            0            0   315       21.03      910.65      122.16   \n",
       "2            0            0  2607        0.00        0.00        0.00   \n",
       "3            0            0   511        0.00        2.45       21.89   \n",
       "4            0            0   667        0.00        0.00        0.00   \n",
       "\n",
       "   sep_vbc_3g  avg_rech_amt  churn  \n",
       "0         0.0        1185.0      1  \n",
       "1         0.0         519.0      0  \n",
       "2         0.0         380.0      0  \n",
       "3         0.0         459.0      0  \n",
       "4         0.0         408.0      0  \n",
       "\n",
       "[5 rows x 172 columns]"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_high_val.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "9.02669021358835"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "### Checking the Churn Rate\n",
    "churn = (sum(df_high_val['churn'])/len(df_high_val['churn'].index))*100\n",
    "churn"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "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>total_og_mou_9</th>\n",
       "      <th>total_ic_mou_9</th>\n",
       "      <th>monthly_2g_9</th>\n",
       "      <th>monthly_3g_9</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>142.18</td>\n",
       "      <td>34.91</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>446.09</td>\n",
       "      <td>631.86</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>255.74</td>\n",
       "      <td>334.56</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>631.08</td>\n",
       "      <td>87.13</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   total_og_mou_9  total_ic_mou_9  monthly_2g_9  monthly_3g_9\n",
       "0            0.00            0.00             0             0\n",
       "1          142.18           34.91             0             0\n",
       "2          446.09          631.86             0             0\n",
       "3          255.74          334.56             0             0\n",
       "4          631.08           87.13             0             0"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_high_val[['total_og_mou_9', 'total_ic_mou_9', 'monthly_2g_9', 'monthly_3g_9']].head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Drop all columns for Month 9"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_high_val.drop(df_high_val.filter(regex=(\"._9\")).columns, inplace=True, axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(30011, 131)"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_high_val.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 30011 entries, 0 to 30010\n",
      "Data columns (total 131 columns):\n",
      " #   Column               Non-Null Count  Dtype  \n",
      "---  ------               --------------  -----  \n",
      " 0   mobile_number        30011 non-null  int64  \n",
      " 1   arpu_6               30011 non-null  float64\n",
      " 2   arpu_7               30011 non-null  float64\n",
      " 3   arpu_8               30011 non-null  float64\n",
      " 4   onnet_mou_6          29695 non-null  float64\n",
      " 5   onnet_mou_7          29708 non-null  float64\n",
      " 6   onnet_mou_8          29073 non-null  float64\n",
      " 7   offnet_mou_6         29695 non-null  float64\n",
      " 8   offnet_mou_7         29708 non-null  float64\n",
      " 9   offnet_mou_8         29073 non-null  float64\n",
      " 10  roam_ic_mou_6        29695 non-null  float64\n",
      " 11  roam_ic_mou_7        29708 non-null  float64\n",
      " 12  roam_ic_mou_8        29073 non-null  float64\n",
      " 13  roam_og_mou_6        29695 non-null  float64\n",
      " 14  roam_og_mou_7        29708 non-null  float64\n",
      " 15  roam_og_mou_8        29073 non-null  float64\n",
      " 16  loc_og_t2t_mou_6     29695 non-null  float64\n",
      " 17  loc_og_t2t_mou_7     29708 non-null  float64\n",
      " 18  loc_og_t2t_mou_8     29073 non-null  float64\n",
      " 19  loc_og_t2m_mou_6     29695 non-null  float64\n",
      " 20  loc_og_t2m_mou_7     29708 non-null  float64\n",
      " 21  loc_og_t2m_mou_8     29073 non-null  float64\n",
      " 22  loc_og_t2f_mou_6     29695 non-null  float64\n",
      " 23  loc_og_t2f_mou_7     29708 non-null  float64\n",
      " 24  loc_og_t2f_mou_8     29073 non-null  float64\n",
      " 25  loc_og_t2c_mou_6     29695 non-null  float64\n",
      " 26  loc_og_t2c_mou_7     29708 non-null  float64\n",
      " 27  loc_og_t2c_mou_8     29073 non-null  float64\n",
      " 28  loc_og_mou_6         29695 non-null  float64\n",
      " 29  loc_og_mou_7         29708 non-null  float64\n",
      " 30  loc_og_mou_8         29073 non-null  float64\n",
      " 31  std_og_t2t_mou_6     29695 non-null  float64\n",
      " 32  std_og_t2t_mou_7     29708 non-null  float64\n",
      " 33  std_og_t2t_mou_8     29073 non-null  float64\n",
      " 34  std_og_t2m_mou_6     29695 non-null  float64\n",
      " 35  std_og_t2m_mou_7     29708 non-null  float64\n",
      " 36  std_og_t2m_mou_8     29073 non-null  float64\n",
      " 37  std_og_t2f_mou_6     29695 non-null  float64\n",
      " 38  std_og_t2f_mou_7     29708 non-null  float64\n",
      " 39  std_og_t2f_mou_8     29073 non-null  float64\n",
      " 40  std_og_mou_6         29695 non-null  float64\n",
      " 41  std_og_mou_7         29708 non-null  float64\n",
      " 42  std_og_mou_8         29073 non-null  float64\n",
      " 43  isd_og_mou_6         29695 non-null  float64\n",
      " 44  isd_og_mou_7         29708 non-null  float64\n",
      " 45  isd_og_mou_8         29073 non-null  float64\n",
      " 46  spl_og_mou_6         29695 non-null  float64\n",
      " 47  spl_og_mou_7         29708 non-null  float64\n",
      " 48  spl_og_mou_8         29073 non-null  float64\n",
      " 49  og_others_6          29695 non-null  float64\n",
      " 50  og_others_7          29708 non-null  float64\n",
      " 51  og_others_8          29073 non-null  float64\n",
      " 52  total_og_mou_6       30011 non-null  float64\n",
      " 53  total_og_mou_7       30011 non-null  float64\n",
      " 54  total_og_mou_8       30011 non-null  float64\n",
      " 55  loc_ic_t2t_mou_6     29695 non-null  float64\n",
      " 56  loc_ic_t2t_mou_7     29708 non-null  float64\n",
      " 57  loc_ic_t2t_mou_8     29073 non-null  float64\n",
      " 58  loc_ic_t2m_mou_6     29695 non-null  float64\n",
      " 59  loc_ic_t2m_mou_7     29708 non-null  float64\n",
      " 60  loc_ic_t2m_mou_8     29073 non-null  float64\n",
      " 61  loc_ic_t2f_mou_6     29695 non-null  float64\n",
      " 62  loc_ic_t2f_mou_7     29708 non-null  float64\n",
      " 63  loc_ic_t2f_mou_8     29073 non-null  float64\n",
      " 64  loc_ic_mou_6         29695 non-null  float64\n",
      " 65  loc_ic_mou_7         29708 non-null  float64\n",
      " 66  loc_ic_mou_8         29073 non-null  float64\n",
      " 67  std_ic_t2t_mou_6     29695 non-null  float64\n",
      " 68  std_ic_t2t_mou_7     29708 non-null  float64\n",
      " 69  std_ic_t2t_mou_8     29073 non-null  float64\n",
      " 70  std_ic_t2m_mou_6     29695 non-null  float64\n",
      " 71  std_ic_t2m_mou_7     29708 non-null  float64\n",
      " 72  std_ic_t2m_mou_8     29073 non-null  float64\n",
      " 73  std_ic_t2f_mou_6     29695 non-null  float64\n",
      " 74  std_ic_t2f_mou_7     29708 non-null  float64\n",
      " 75  std_ic_t2f_mou_8     29073 non-null  float64\n",
      " 76  std_ic_mou_6         29695 non-null  float64\n",
      " 77  std_ic_mou_7         29708 non-null  float64\n",
      " 78  std_ic_mou_8         29073 non-null  float64\n",
      " 79  total_ic_mou_6       30011 non-null  float64\n",
      " 80  total_ic_mou_7       30011 non-null  float64\n",
      " 81  total_ic_mou_8       30011 non-null  float64\n",
      " 82  spl_ic_mou_6         29695 non-null  float64\n",
      " 83  spl_ic_mou_7         29708 non-null  float64\n",
      " 84  spl_ic_mou_8         29073 non-null  float64\n",
      " 85  isd_ic_mou_6         29695 non-null  float64\n",
      " 86  isd_ic_mou_7         29708 non-null  float64\n",
      " 87  isd_ic_mou_8         29073 non-null  float64\n",
      " 88  ic_others_6          29695 non-null  float64\n",
      " 89  ic_others_7          29708 non-null  float64\n",
      " 90  ic_others_8          29073 non-null  float64\n",
      " 91  total_rech_num_6     30011 non-null  int64  \n",
      " 92  total_rech_num_7     30011 non-null  int64  \n",
      " 93  total_rech_num_8     30011 non-null  int64  \n",
      " 94  total_rech_amt_6     30011 non-null  int64  \n",
      " 95  total_rech_amt_7     30011 non-null  int64  \n",
      " 96  total_rech_amt_8     30011 non-null  int64  \n",
      " 97  max_rech_amt_6       30011 non-null  int64  \n",
      " 98  max_rech_amt_7       30011 non-null  int64  \n",
      " 99  max_rech_amt_8       30011 non-null  int64  \n",
      " 100 date_of_last_rech_6  29949 non-null  object \n",
      " 101 date_of_last_rech_7  29897 non-null  object \n",
      " 102 date_of_last_rech_8  29417 non-null  object \n",
      " 103 last_day_rch_amt_6   30011 non-null  int64  \n",
      " 104 last_day_rch_amt_7   30011 non-null  int64  \n",
      " 105 last_day_rch_amt_8   30011 non-null  int64  \n",
      " 106 vol_2g_mb_6          30011 non-null  float64\n",
      " 107 vol_2g_mb_7          30011 non-null  float64\n",
      " 108 vol_2g_mb_8          30011 non-null  float64\n",
      " 109 vol_3g_mb_6          30011 non-null  float64\n",
      " 110 vol_3g_mb_7          30011 non-null  float64\n",
      " 111 vol_3g_mb_8          30011 non-null  float64\n",
      " 112 monthly_2g_6         30011 non-null  int64  \n",
      " 113 monthly_2g_7         30011 non-null  int64  \n",
      " 114 monthly_2g_8         30011 non-null  int64  \n",
      " 115 sachet_2g_6          30011 non-null  int64  \n",
      " 116 sachet_2g_7          30011 non-null  int64  \n",
      " 117 sachet_2g_8          30011 non-null  int64  \n",
      " 118 monthly_3g_6         30011 non-null  int64  \n",
      " 119 monthly_3g_7         30011 non-null  int64  \n",
      " 120 monthly_3g_8         30011 non-null  int64  \n",
      " 121 sachet_3g_6          30011 non-null  int64  \n",
      " 122 sachet_3g_7          30011 non-null  int64  \n",
      " 123 sachet_3g_8          30011 non-null  int64  \n",
      " 124 aon                  30011 non-null  int64  \n",
      " 125 aug_vbc_3g           30011 non-null  float64\n",
      " 126 jul_vbc_3g           30011 non-null  float64\n",
      " 127 jun_vbc_3g           30011 non-null  float64\n",
      " 128 sep_vbc_3g           30011 non-null  float64\n",
      " 129 avg_rech_amt         30011 non-null  float64\n",
      " 130 churn                30011 non-null  int32  \n",
      "dtypes: float64(101), int32(1), int64(26), object(3)\n",
      "memory usage: 29.9+ MB\n"
     ]
    }
   ],
   "source": [
    "df_high_val.info(verbose=True, null_counts=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(30011, 131)"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_high_val.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "# removing mobile number\n",
    "df_high_val.drop(['mobile_number', 'date_of_last_rech_6', 'date_of_last_rech_7', 'date_of_last_rech_8'], inplace=True, axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "isd_og_mou_8        3.13\n",
      "loc_ic_t2f_mou_8    3.13\n",
      "spl_ic_mou_8        3.13\n",
      "ic_others_8         3.13\n",
      "std_og_t2m_mou_8    3.13\n",
      "std_og_t2f_mou_8    3.13\n",
      "std_og_t2t_mou_8    3.13\n",
      "std_ic_mou_8        3.13\n",
      "loc_og_mou_8        3.13\n",
      "std_ic_t2f_mou_8    3.13\n",
      "loc_og_t2c_mou_8    3.13\n",
      "std_ic_t2m_mou_8    3.13\n",
      "loc_og_t2f_mou_8    3.13\n",
      "std_ic_t2t_mou_8    3.13\n",
      "loc_og_t2m_mou_8    3.13\n",
      "loc_og_t2t_mou_8    3.13\n",
      "loc_ic_mou_8        3.13\n",
      "offnet_mou_8        3.13\n",
      "loc_ic_t2t_mou_8    3.13\n",
      "spl_og_mou_8        3.13\n",
      "og_others_8         3.13\n",
      "onnet_mou_8         3.13\n",
      "std_og_mou_8        3.13\n",
      "roam_og_mou_8       3.13\n",
      "isd_ic_mou_8        3.13\n",
      "roam_ic_mou_8       3.13\n",
      "loc_ic_t2m_mou_8    3.13\n",
      "std_ic_t2m_mou_6    1.05\n",
      "loc_ic_t2t_mou_6    1.05\n",
      "std_ic_t2f_mou_6    1.05\n",
      "loc_ic_t2m_mou_6    1.05\n",
      "std_ic_t2t_mou_6    1.05\n",
      "std_ic_mou_6        1.05\n",
      "isd_og_mou_6        1.05\n",
      "isd_ic_mou_6        1.05\n",
      "std_og_t2f_mou_6    1.05\n",
      "og_others_6         1.05\n",
      "std_og_mou_6        1.05\n",
      "spl_ic_mou_6        1.05\n",
      "spl_og_mou_6        1.05\n",
      "loc_ic_t2f_mou_6    1.05\n",
      "loc_ic_mou_6        1.05\n",
      "roam_ic_mou_6       1.05\n",
      "std_og_t2t_mou_6    1.05\n",
      "loc_og_t2m_mou_6    1.05\n",
      "ic_others_6         1.05\n",
      "loc_og_t2f_mou_6    1.05\n",
      "roam_og_mou_6       1.05\n",
      "loc_og_t2c_mou_6    1.05\n",
      "loc_og_mou_6        1.05\n",
      "dtype: float64\n"
     ]
    }
   ],
   "source": [
    "# let's check the distribution of churn column\n",
    "sortMissing(df_high_val, 50)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Imputing missing values with KNN imputer\n",
    "imputer = KNNImputer(n_neighbors=2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_high_val_imputed = pd.DataFrame(imputer.fit_transform(df_high_val),columns = df_high_val.columns)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "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>arpu_6</th>\n",
       "      <th>arpu_7</th>\n",
       "      <th>arpu_8</th>\n",
       "      <th>onnet_mou_6</th>\n",
       "      <th>onnet_mou_7</th>\n",
       "      <th>onnet_mou_8</th>\n",
       "      <th>offnet_mou_6</th>\n",
       "      <th>offnet_mou_7</th>\n",
       "      <th>offnet_mou_8</th>\n",
       "      <th>roam_ic_mou_6</th>\n",
       "      <th>...</th>\n",
       "      <th>sachet_3g_6</th>\n",
       "      <th>sachet_3g_7</th>\n",
       "      <th>sachet_3g_8</th>\n",
       "      <th>aon</th>\n",
       "      <th>aug_vbc_3g</th>\n",
       "      <th>jul_vbc_3g</th>\n",
       "      <th>jun_vbc_3g</th>\n",
       "      <th>sep_vbc_3g</th>\n",
       "      <th>avg_rech_amt</th>\n",
       "      <th>churn</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1069.180</td>\n",
       "      <td>1349.850</td>\n",
       "      <td>3171.480</td>\n",
       "      <td>57.84</td>\n",
       "      <td>54.68</td>\n",
       "      <td>52.29</td>\n",
       "      <td>453.43</td>\n",
       "      <td>567.16</td>\n",
       "      <td>325.91</td>\n",
       "      <td>16.23</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>802.0</td>\n",
       "      <td>57.74</td>\n",
       "      <td>19.38</td>\n",
       "      <td>18.74</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1185.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>378.721</td>\n",
       "      <td>492.223</td>\n",
       "      <td>137.362</td>\n",
       "      <td>413.69</td>\n",
       "      <td>351.03</td>\n",
       "      <td>35.08</td>\n",
       "      <td>94.66</td>\n",
       "      <td>80.63</td>\n",
       "      <td>136.48</td>\n",
       "      <td>0.00</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>315.0</td>\n",
       "      <td>21.03</td>\n",
       "      <td>910.65</td>\n",
       "      <td>122.16</td>\n",
       "      <td>0.0</td>\n",
       "      <td>519.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>492.846</td>\n",
       "      <td>205.671</td>\n",
       "      <td>593.260</td>\n",
       "      <td>501.76</td>\n",
       "      <td>108.39</td>\n",
       "      <td>534.24</td>\n",
       "      <td>413.31</td>\n",
       "      <td>119.28</td>\n",
       "      <td>482.46</td>\n",
       "      <td>23.53</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2607.0</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.0</td>\n",
       "      <td>380.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>430.975</td>\n",
       "      <td>299.869</td>\n",
       "      <td>187.894</td>\n",
       "      <td>50.51</td>\n",
       "      <td>74.01</td>\n",
       "      <td>70.61</td>\n",
       "      <td>296.29</td>\n",
       "      <td>229.74</td>\n",
       "      <td>162.76</td>\n",
       "      <td>0.00</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>511.0</td>\n",
       "      <td>0.00</td>\n",
       "      <td>2.45</td>\n",
       "      <td>21.89</td>\n",
       "      <td>0.0</td>\n",
       "      <td>459.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>690.008</td>\n",
       "      <td>18.980</td>\n",
       "      <td>25.499</td>\n",
       "      <td>1185.91</td>\n",
       "      <td>9.28</td>\n",
       "      <td>7.79</td>\n",
       "      <td>61.64</td>\n",
       "      <td>0.00</td>\n",
       "      <td>5.54</td>\n",
       "      <td>0.00</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>667.0</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.0</td>\n",
       "      <td>408.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 127 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "     arpu_6    arpu_7    arpu_8  onnet_mou_6  onnet_mou_7  onnet_mou_8  \\\n",
       "0  1069.180  1349.850  3171.480        57.84        54.68        52.29   \n",
       "1   378.721   492.223   137.362       413.69       351.03        35.08   \n",
       "2   492.846   205.671   593.260       501.76       108.39       534.24   \n",
       "3   430.975   299.869   187.894        50.51        74.01        70.61   \n",
       "4   690.008    18.980    25.499      1185.91         9.28         7.79   \n",
       "\n",
       "   offnet_mou_6  offnet_mou_7  offnet_mou_8  roam_ic_mou_6  ...  sachet_3g_6  \\\n",
       "0        453.43        567.16        325.91          16.23  ...          0.0   \n",
       "1         94.66         80.63        136.48           0.00  ...          0.0   \n",
       "2        413.31        119.28        482.46          23.53  ...          0.0   \n",
       "3        296.29        229.74        162.76           0.00  ...          0.0   \n",
       "4         61.64          0.00          5.54           0.00  ...          0.0   \n",
       "\n",
       "   sachet_3g_7  sachet_3g_8     aon  aug_vbc_3g  jul_vbc_3g  jun_vbc_3g  \\\n",
       "0          0.0          0.0   802.0       57.74       19.38       18.74   \n",
       "1          0.0          0.0   315.0       21.03      910.65      122.16   \n",
       "2          0.0          0.0  2607.0        0.00        0.00        0.00   \n",
       "3          0.0          0.0   511.0        0.00        2.45       21.89   \n",
       "4          0.0          0.0   667.0        0.00        0.00        0.00   \n",
       "\n",
       "   sep_vbc_3g  avg_rech_amt  churn  \n",
       "0         0.0        1185.0    1.0  \n",
       "1         0.0         519.0    0.0  \n",
       "2         0.0         380.0    0.0  \n",
       "3         0.0         459.0    0.0  \n",
       "4         0.0         408.0    0.0  \n",
       "\n",
       "[5 rows x 127 columns]"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_high_val_imputed.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "churn               0.0\n",
      "spl_og_mou_7        0.0\n",
      "std_og_t2m_mou_6    0.0\n",
      "std_og_t2m_mou_7    0.0\n",
      "std_og_t2m_mou_8    0.0\n",
      "std_og_t2f_mou_6    0.0\n",
      "std_og_t2f_mou_7    0.0\n",
      "std_og_t2f_mou_8    0.0\n",
      "std_og_mou_6        0.0\n",
      "std_og_mou_7        0.0\n",
      "dtype: float64\n"
     ]
    }
   ],
   "source": [
    "sortMissing(df_high_val_imputed, 10)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Data Preparation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 30011 entries, 0 to 30010\n",
      "Data columns (total 127 columns):\n",
      " #   Column              Dtype  \n",
      "---  ------              -----  \n",
      " 0   arpu_6              float64\n",
      " 1   arpu_7              float64\n",
      " 2   arpu_8              float64\n",
      " 3   onnet_mou_6         float64\n",
      " 4   onnet_mou_7         float64\n",
      " 5   onnet_mou_8         float64\n",
      " 6   offnet_mou_6        float64\n",
      " 7   offnet_mou_7        float64\n",
      " 8   offnet_mou_8        float64\n",
      " 9   roam_ic_mou_6       float64\n",
      " 10  roam_ic_mou_7       float64\n",
      " 11  roam_ic_mou_8       float64\n",
      " 12  roam_og_mou_6       float64\n",
      " 13  roam_og_mou_7       float64\n",
      " 14  roam_og_mou_8       float64\n",
      " 15  loc_og_t2t_mou_6    float64\n",
      " 16  loc_og_t2t_mou_7    float64\n",
      " 17  loc_og_t2t_mou_8    float64\n",
      " 18  loc_og_t2m_mou_6    float64\n",
      " 19  loc_og_t2m_mou_7    float64\n",
      " 20  loc_og_t2m_mou_8    float64\n",
      " 21  loc_og_t2f_mou_6    float64\n",
      " 22  loc_og_t2f_mou_7    float64\n",
      " 23  loc_og_t2f_mou_8    float64\n",
      " 24  loc_og_t2c_mou_6    float64\n",
      " 25  loc_og_t2c_mou_7    float64\n",
      " 26  loc_og_t2c_mou_8    float64\n",
      " 27  loc_og_mou_6        float64\n",
      " 28  loc_og_mou_7        float64\n",
      " 29  loc_og_mou_8        float64\n",
      " 30  std_og_t2t_mou_6    float64\n",
      " 31  std_og_t2t_mou_7    float64\n",
      " 32  std_og_t2t_mou_8    float64\n",
      " 33  std_og_t2m_mou_6    float64\n",
      " 34  std_og_t2m_mou_7    float64\n",
      " 35  std_og_t2m_mou_8    float64\n",
      " 36  std_og_t2f_mou_6    float64\n",
      " 37  std_og_t2f_mou_7    float64\n",
      " 38  std_og_t2f_mou_8    float64\n",
      " 39  std_og_mou_6        float64\n",
      " 40  std_og_mou_7        float64\n",
      " 41  std_og_mou_8        float64\n",
      " 42  isd_og_mou_6        float64\n",
      " 43  isd_og_mou_7        float64\n",
      " 44  isd_og_mou_8        float64\n",
      " 45  spl_og_mou_6        float64\n",
      " 46  spl_og_mou_7        float64\n",
      " 47  spl_og_mou_8        float64\n",
      " 48  og_others_6         float64\n",
      " 49  og_others_7         float64\n",
      " 50  og_others_8         float64\n",
      " 51  total_og_mou_6      float64\n",
      " 52  total_og_mou_7      float64\n",
      " 53  total_og_mou_8      float64\n",
      " 54  loc_ic_t2t_mou_6    float64\n",
      " 55  loc_ic_t2t_mou_7    float64\n",
      " 56  loc_ic_t2t_mou_8    float64\n",
      " 57  loc_ic_t2m_mou_6    float64\n",
      " 58  loc_ic_t2m_mou_7    float64\n",
      " 59  loc_ic_t2m_mou_8    float64\n",
      " 60  loc_ic_t2f_mou_6    float64\n",
      " 61  loc_ic_t2f_mou_7    float64\n",
      " 62  loc_ic_t2f_mou_8    float64\n",
      " 63  loc_ic_mou_6        float64\n",
      " 64  loc_ic_mou_7        float64\n",
      " 65  loc_ic_mou_8        float64\n",
      " 66  std_ic_t2t_mou_6    float64\n",
      " 67  std_ic_t2t_mou_7    float64\n",
      " 68  std_ic_t2t_mou_8    float64\n",
      " 69  std_ic_t2m_mou_6    float64\n",
      " 70  std_ic_t2m_mou_7    float64\n",
      " 71  std_ic_t2m_mou_8    float64\n",
      " 72  std_ic_t2f_mou_6    float64\n",
      " 73  std_ic_t2f_mou_7    float64\n",
      " 74  std_ic_t2f_mou_8    float64\n",
      " 75  std_ic_mou_6        float64\n",
      " 76  std_ic_mou_7        float64\n",
      " 77  std_ic_mou_8        float64\n",
      " 78  total_ic_mou_6      float64\n",
      " 79  total_ic_mou_7      float64\n",
      " 80  total_ic_mou_8      float64\n",
      " 81  spl_ic_mou_6        float64\n",
      " 82  spl_ic_mou_7        float64\n",
      " 83  spl_ic_mou_8        float64\n",
      " 84  isd_ic_mou_6        float64\n",
      " 85  isd_ic_mou_7        float64\n",
      " 86  isd_ic_mou_8        float64\n",
      " 87  ic_others_6         float64\n",
      " 88  ic_others_7         float64\n",
      " 89  ic_others_8         float64\n",
      " 90  total_rech_num_6    float64\n",
      " 91  total_rech_num_7    float64\n",
      " 92  total_rech_num_8    float64\n",
      " 93  total_rech_amt_6    float64\n",
      " 94  total_rech_amt_7    float64\n",
      " 95  total_rech_amt_8    float64\n",
      " 96  max_rech_amt_6      float64\n",
      " 97  max_rech_amt_7      float64\n",
      " 98  max_rech_amt_8      float64\n",
      " 99  last_day_rch_amt_6  float64\n",
      " 100 last_day_rch_amt_7  float64\n",
      " 101 last_day_rch_amt_8  float64\n",
      " 102 vol_2g_mb_6         float64\n",
      " 103 vol_2g_mb_7         float64\n",
      " 104 vol_2g_mb_8         float64\n",
      " 105 vol_3g_mb_6         float64\n",
      " 106 vol_3g_mb_7         float64\n",
      " 107 vol_3g_mb_8         float64\n",
      " 108 monthly_2g_6        float64\n",
      " 109 monthly_2g_7        float64\n",
      " 110 monthly_2g_8        float64\n",
      " 111 sachet_2g_6         float64\n",
      " 112 sachet_2g_7         float64\n",
      " 113 sachet_2g_8         float64\n",
      " 114 monthly_3g_6        float64\n",
      " 115 monthly_3g_7        float64\n",
      " 116 monthly_3g_8        float64\n",
      " 117 sachet_3g_6         float64\n",
      " 118 sachet_3g_7         float64\n",
      " 119 sachet_3g_8         float64\n",
      " 120 aon                 float64\n",
      " 121 aug_vbc_3g          float64\n",
      " 122 jul_vbc_3g          float64\n",
      " 123 jun_vbc_3g          float64\n",
      " 124 sep_vbc_3g          float64\n",
      " 125 avg_rech_amt        float64\n",
      " 126 churn               float64\n",
      "dtypes: float64(127)\n",
      "memory usage: 29.1 MB\n"
     ]
    }
   ],
   "source": [
    "df_high_val_imputed.info(verbose=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "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>arpu_6</th>\n",
       "      <th>arpu_7</th>\n",
       "      <th>arpu_8</th>\n",
       "      <th>onnet_mou_6</th>\n",
       "      <th>onnet_mou_7</th>\n",
       "      <th>onnet_mou_8</th>\n",
       "      <th>offnet_mou_6</th>\n",
       "      <th>offnet_mou_7</th>\n",
       "      <th>offnet_mou_8</th>\n",
       "      <th>roam_ic_mou_6</th>\n",
       "      <th>...</th>\n",
       "      <th>sachet_3g_6</th>\n",
       "      <th>sachet_3g_7</th>\n",
       "      <th>sachet_3g_8</th>\n",
       "      <th>aon</th>\n",
       "      <th>aug_vbc_3g</th>\n",
       "      <th>jul_vbc_3g</th>\n",
       "      <th>jun_vbc_3g</th>\n",
       "      <th>sep_vbc_3g</th>\n",
       "      <th>avg_rech_amt</th>\n",
       "      <th>churn</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>30011.000000</td>\n",
       "      <td>30011.000000</td>\n",
       "      <td>30011.000000</td>\n",
       "      <td>30011.000000</td>\n",
       "      <td>30011.000000</td>\n",
       "      <td>30011.000000</td>\n",
       "      <td>30011.000000</td>\n",
       "      <td>30011.000000</td>\n",
       "      <td>30011.000000</td>\n",
       "      <td>30011.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>30011.000000</td>\n",
       "      <td>30011.000000</td>\n",
       "      <td>30011.000000</td>\n",
       "      <td>30011.000000</td>\n",
       "      <td>30011.000000</td>\n",
       "      <td>30011.000000</td>\n",
       "      <td>30011.000000</td>\n",
       "      <td>30011.000000</td>\n",
       "      <td>30011.000000</td>\n",
       "      <td>30011.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>587.284404</td>\n",
       "      <td>589.135427</td>\n",
       "      <td>534.857433</td>\n",
       "      <td>296.570022</td>\n",
       "      <td>304.701251</td>\n",
       "      <td>268.467774</td>\n",
       "      <td>418.892374</td>\n",
       "      <td>424.709081</td>\n",
       "      <td>376.694954</td>\n",
       "      <td>17.706985</td>\n",
       "      <td>...</td>\n",
       "      <td>0.137916</td>\n",
       "      <td>0.150212</td>\n",
       "      <td>0.140415</td>\n",
       "      <td>1264.064776</td>\n",
       "      <td>129.439626</td>\n",
       "      <td>135.127102</td>\n",
       "      <td>121.360548</td>\n",
       "      <td>6.562685</td>\n",
       "      <td>696.664356</td>\n",
       "      <td>0.090267</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>442.722413</td>\n",
       "      <td>462.897814</td>\n",
       "      <td>492.259586</td>\n",
       "      <td>460.614696</td>\n",
       "      <td>481.593939</td>\n",
       "      <td>466.180291</td>\n",
       "      <td>469.956346</td>\n",
       "      <td>485.984268</td>\n",
       "      <td>476.437782</td>\n",
       "      <td>80.065676</td>\n",
       "      <td>...</td>\n",
       "      <td>0.875528</td>\n",
       "      <td>0.969286</td>\n",
       "      <td>0.974727</td>\n",
       "      <td>975.263117</td>\n",
       "      <td>390.478591</td>\n",
       "      <td>408.024394</td>\n",
       "      <td>389.726031</td>\n",
       "      <td>48.638658</td>\n",
       "      <td>488.782088</td>\n",
       "      <td>0.286569</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>-2258.709000</td>\n",
       "      <td>-2014.045000</td>\n",
       "      <td>-945.808000</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>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>180.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>368.500000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>364.161000</td>\n",
       "      <td>365.004500</td>\n",
       "      <td>289.609500</td>\n",
       "      <td>41.950000</td>\n",
       "      <td>41.480000</td>\n",
       "      <td>28.670000</td>\n",
       "      <td>138.430000</td>\n",
       "      <td>136.360000</td>\n",
       "      <td>98.285000</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>480.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>450.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>495.682000</td>\n",
       "      <td>493.561000</td>\n",
       "      <td>452.091000</td>\n",
       "      <td>126.060000</td>\n",
       "      <td>125.730000</td>\n",
       "      <td>100.360000</td>\n",
       "      <td>282.560000</td>\n",
       "      <td>282.330000</td>\n",
       "      <td>241.580000</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>914.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>568.500000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>703.922000</td>\n",
       "      <td>700.788000</td>\n",
       "      <td>671.150000</td>\n",
       "      <td>353.380000</td>\n",
       "      <td>359.925000</td>\n",
       "      <td>298.140000</td>\n",
       "      <td>523.295000</td>\n",
       "      <td>532.775000</td>\n",
       "      <td>482.735000</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>1924.000000</td>\n",
       "      <td>1.600000</td>\n",
       "      <td>1.990000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>795.500000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>90%</th>\n",
       "      <td>994.099000</td>\n",
       "      <td>995.557000</td>\n",
       "      <td>985.202000</td>\n",
       "      <td>795.340000</td>\n",
       "      <td>824.380000</td>\n",
       "      <td>723.610000</td>\n",
       "      <td>916.030000</td>\n",
       "      <td>935.690000</td>\n",
       "      <td>853.790000</td>\n",
       "      <td>33.390000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>2846.000000</td>\n",
       "      <td>438.970000</td>\n",
       "      <td>454.830000</td>\n",
       "      <td>394.500000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1118.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>95%</th>\n",
       "      <td>1240.793500</td>\n",
       "      <td>1261.421500</td>\n",
       "      <td>1258.555500</td>\n",
       "      <td>1154.405000</td>\n",
       "      <td>1208.175000</td>\n",
       "      <td>1115.660000</td>\n",
       "      <td>1271.470000</td>\n",
       "      <td>1286.285000</td>\n",
       "      <td>1188.460000</td>\n",
       "      <td>95.565000</td>\n",
       "      <td>...</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>3179.000000</td>\n",
       "      <td>818.750000</td>\n",
       "      <td>845.315000</td>\n",
       "      <td>761.740000</td>\n",
       "      <td>18.635000</td>\n",
       "      <td>1406.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>99%</th>\n",
       "      <td>1983.089500</td>\n",
       "      <td>1999.374500</td>\n",
       "      <td>1987.934400</td>\n",
       "      <td>2166.370000</td>\n",
       "      <td>2220.367000</td>\n",
       "      <td>2188.504000</td>\n",
       "      <td>2326.290000</td>\n",
       "      <td>2410.104000</td>\n",
       "      <td>2211.642000</td>\n",
       "      <td>350.374000</td>\n",
       "      <td>...</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>3651.000000</td>\n",
       "      <td>1822.115000</td>\n",
       "      <td>1941.598000</td>\n",
       "      <td>1866.386000</td>\n",
       "      <td>173.662000</td>\n",
       "      <td>2216.300000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>27731.088000</td>\n",
       "      <td>35145.834000</td>\n",
       "      <td>33543.624000</td>\n",
       "      <td>7376.710000</td>\n",
       "      <td>8157.780000</td>\n",
       "      <td>10752.560000</td>\n",
       "      <td>8362.360000</td>\n",
       "      <td>9667.130000</td>\n",
       "      <td>14007.340000</td>\n",
       "      <td>2613.310000</td>\n",
       "      <td>...</td>\n",
       "      <td>29.000000</td>\n",
       "      <td>35.000000</td>\n",
       "      <td>41.000000</td>\n",
       "      <td>4321.000000</td>\n",
       "      <td>12916.220000</td>\n",
       "      <td>9165.600000</td>\n",
       "      <td>11166.210000</td>\n",
       "      <td>2618.570000</td>\n",
       "      <td>37762.500000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>11 rows × 127 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "             arpu_6        arpu_7        arpu_8   onnet_mou_6   onnet_mou_7  \\\n",
       "count  30011.000000  30011.000000  30011.000000  30011.000000  30011.000000   \n",
       "mean     587.284404    589.135427    534.857433    296.570022    304.701251   \n",
       "std      442.722413    462.897814    492.259586    460.614696    481.593939   \n",
       "min    -2258.709000  -2014.045000   -945.808000      0.000000      0.000000   \n",
       "25%      364.161000    365.004500    289.609500     41.950000     41.480000   \n",
       "50%      495.682000    493.561000    452.091000    126.060000    125.730000   \n",
       "75%      703.922000    700.788000    671.150000    353.380000    359.925000   \n",
       "90%      994.099000    995.557000    985.202000    795.340000    824.380000   \n",
       "95%     1240.793500   1261.421500   1258.555500   1154.405000   1208.175000   \n",
       "99%     1983.089500   1999.374500   1987.934400   2166.370000   2220.367000   \n",
       "max    27731.088000  35145.834000  33543.624000   7376.710000   8157.780000   \n",
       "\n",
       "        onnet_mou_8  offnet_mou_6  offnet_mou_7  offnet_mou_8  roam_ic_mou_6  \\\n",
       "count  30011.000000  30011.000000  30011.000000  30011.000000   30011.000000   \n",
       "mean     268.467774    418.892374    424.709081    376.694954      17.706985   \n",
       "std      466.180291    469.956346    485.984268    476.437782      80.065676   \n",
       "min        0.000000      0.000000      0.000000      0.000000       0.000000   \n",
       "25%       28.670000    138.430000    136.360000     98.285000       0.000000   \n",
       "50%      100.360000    282.560000    282.330000    241.580000       0.000000   \n",
       "75%      298.140000    523.295000    532.775000    482.735000       0.000000   \n",
       "90%      723.610000    916.030000    935.690000    853.790000      33.390000   \n",
       "95%     1115.660000   1271.470000   1286.285000   1188.460000      95.565000   \n",
       "99%     2188.504000   2326.290000   2410.104000   2211.642000     350.374000   \n",
       "max    10752.560000   8362.360000   9667.130000  14007.340000    2613.310000   \n",
       "\n",
       "       ...   sachet_3g_6   sachet_3g_7   sachet_3g_8           aon  \\\n",
       "count  ...  30011.000000  30011.000000  30011.000000  30011.000000   \n",
       "mean   ...      0.137916      0.150212      0.140415   1264.064776   \n",
       "std    ...      0.875528      0.969286      0.974727    975.263117   \n",
       "min    ...      0.000000      0.000000      0.000000    180.000000   \n",
       "25%    ...      0.000000      0.000000      0.000000    480.000000   \n",
       "50%    ...      0.000000      0.000000      0.000000    914.000000   \n",
       "75%    ...      0.000000      0.000000      0.000000   1924.000000   \n",
       "90%    ...      0.000000      0.000000      0.000000   2846.000000   \n",
       "95%    ...      1.000000      1.000000      1.000000   3179.000000   \n",
       "99%    ...      3.000000      3.000000      3.000000   3651.000000   \n",
       "max    ...     29.000000     35.000000     41.000000   4321.000000   \n",
       "\n",
       "         aug_vbc_3g    jul_vbc_3g    jun_vbc_3g    sep_vbc_3g  avg_rech_amt  \\\n",
       "count  30011.000000  30011.000000  30011.000000  30011.000000  30011.000000   \n",
       "mean     129.439626    135.127102    121.360548      6.562685    696.664356   \n",
       "std      390.478591    408.024394    389.726031     48.638658    488.782088   \n",
       "min        0.000000      0.000000      0.000000      0.000000    368.500000   \n",
       "25%        0.000000      0.000000      0.000000      0.000000    450.000000   \n",
       "50%        0.000000      0.000000      0.000000      0.000000    568.500000   \n",
       "75%        1.600000      1.990000      0.000000      0.000000    795.500000   \n",
       "90%      438.970000    454.830000    394.500000      0.000000   1118.000000   \n",
       "95%      818.750000    845.315000    761.740000     18.635000   1406.000000   \n",
       "99%     1822.115000   1941.598000   1866.386000    173.662000   2216.300000   \n",
       "max    12916.220000   9165.600000  11166.210000   2618.570000  37762.500000   \n",
       "\n",
       "              churn  \n",
       "count  30011.000000  \n",
       "mean       0.090267  \n",
       "std        0.286569  \n",
       "min        0.000000  \n",
       "25%        0.000000  \n",
       "50%        0.000000  \n",
       "75%        0.000000  \n",
       "90%        0.000000  \n",
       "95%        1.000000  \n",
       "99%        1.000000  \n",
       "max        1.000000  \n",
       "\n",
       "[11 rows x 127 columns]"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Finding outliers\n",
    "df_high_val_imputed.describe(percentiles=[.25,.5,.75,.90,.95,.99])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['arpu_6',\n",
       " 'arpu_7',\n",
       " 'arpu_8',\n",
       " 'onnet_mou_6',\n",
       " 'onnet_mou_7',\n",
       " 'onnet_mou_8',\n",
       " 'offnet_mou_6',\n",
       " 'offnet_mou_7',\n",
       " 'offnet_mou_8',\n",
       " 'roam_ic_mou_6',\n",
       " 'roam_ic_mou_7',\n",
       " 'roam_ic_mou_8',\n",
       " 'roam_og_mou_6',\n",
       " 'roam_og_mou_7',\n",
       " 'roam_og_mou_8',\n",
       " 'loc_og_t2t_mou_6',\n",
       " 'loc_og_t2t_mou_7',\n",
       " 'loc_og_t2t_mou_8',\n",
       " 'loc_og_t2m_mou_6',\n",
       " 'loc_og_t2m_mou_7',\n",
       " 'loc_og_t2m_mou_8',\n",
       " 'loc_og_t2f_mou_6',\n",
       " 'loc_og_t2f_mou_7',\n",
       " 'loc_og_t2f_mou_8',\n",
       " 'loc_og_t2c_mou_6',\n",
       " 'loc_og_t2c_mou_7',\n",
       " 'loc_og_t2c_mou_8',\n",
       " 'loc_og_mou_6',\n",
       " 'loc_og_mou_7',\n",
       " 'loc_og_mou_8',\n",
       " 'std_og_t2t_mou_6',\n",
       " 'std_og_t2t_mou_7',\n",
       " 'std_og_t2t_mou_8',\n",
       " 'std_og_t2m_mou_6',\n",
       " 'std_og_t2m_mou_7',\n",
       " 'std_og_t2m_mou_8',\n",
       " 'std_og_t2f_mou_6',\n",
       " 'std_og_t2f_mou_7',\n",
       " 'std_og_t2f_mou_8',\n",
       " 'std_og_mou_6',\n",
       " 'std_og_mou_7',\n",
       " 'std_og_mou_8',\n",
       " 'isd_og_mou_6',\n",
       " 'isd_og_mou_7',\n",
       " 'isd_og_mou_8',\n",
       " 'spl_og_mou_6',\n",
       " 'spl_og_mou_7',\n",
       " 'spl_og_mou_8',\n",
       " 'og_others_6',\n",
       " 'og_others_7',\n",
       " 'og_others_8',\n",
       " 'total_og_mou_6',\n",
       " 'total_og_mou_7',\n",
       " 'total_og_mou_8',\n",
       " 'loc_ic_t2t_mou_6',\n",
       " 'loc_ic_t2t_mou_7',\n",
       " 'loc_ic_t2t_mou_8',\n",
       " 'loc_ic_t2m_mou_6',\n",
       " 'loc_ic_t2m_mou_7',\n",
       " 'loc_ic_t2m_mou_8',\n",
       " 'loc_ic_t2f_mou_6',\n",
       " 'loc_ic_t2f_mou_7',\n",
       " 'loc_ic_t2f_mou_8',\n",
       " 'loc_ic_mou_6',\n",
       " 'loc_ic_mou_7',\n",
       " 'loc_ic_mou_8',\n",
       " 'std_ic_t2t_mou_6',\n",
       " 'std_ic_t2t_mou_7',\n",
       " 'std_ic_t2t_mou_8',\n",
       " 'std_ic_t2m_mou_6',\n",
       " 'std_ic_t2m_mou_7',\n",
       " 'std_ic_t2m_mou_8',\n",
       " 'std_ic_t2f_mou_6',\n",
       " 'std_ic_t2f_mou_7',\n",
       " 'std_ic_t2f_mou_8',\n",
       " 'std_ic_mou_6',\n",
       " 'std_ic_mou_7',\n",
       " 'std_ic_mou_8',\n",
       " 'total_ic_mou_6',\n",
       " 'total_ic_mou_7',\n",
       " 'total_ic_mou_8',\n",
       " 'spl_ic_mou_6',\n",
       " 'spl_ic_mou_7',\n",
       " 'spl_ic_mou_8',\n",
       " 'isd_ic_mou_6',\n",
       " 'isd_ic_mou_7',\n",
       " 'isd_ic_mou_8',\n",
       " 'ic_others_6',\n",
       " 'ic_others_7',\n",
       " 'ic_others_8',\n",
       " 'total_rech_num_6',\n",
       " 'total_rech_num_7',\n",
       " 'total_rech_num_8',\n",
       " 'total_rech_amt_6',\n",
       " 'total_rech_amt_7',\n",
       " 'total_rech_amt_8',\n",
       " 'max_rech_amt_6',\n",
       " 'max_rech_amt_7',\n",
       " 'max_rech_amt_8',\n",
       " 'last_day_rch_amt_6',\n",
       " 'last_day_rch_amt_7',\n",
       " 'last_day_rch_amt_8',\n",
       " 'vol_2g_mb_6',\n",
       " 'vol_2g_mb_7',\n",
       " 'vol_2g_mb_8',\n",
       " 'vol_3g_mb_6',\n",
       " 'vol_3g_mb_7',\n",
       " 'vol_3g_mb_8',\n",
       " 'monthly_2g_6',\n",
       " 'monthly_2g_7',\n",
       " 'monthly_2g_8',\n",
       " 'sachet_2g_6',\n",
       " 'sachet_2g_7',\n",
       " 'sachet_2g_8',\n",
       " 'monthly_3g_6',\n",
       " 'monthly_3g_7',\n",
       " 'monthly_3g_8',\n",
       " 'sachet_3g_6',\n",
       " 'sachet_3g_7',\n",
       " 'sachet_3g_8',\n",
       " 'aon',\n",
       " 'aug_vbc_3g',\n",
       " 'jul_vbc_3g',\n",
       " 'jun_vbc_3g',\n",
       " 'sep_vbc_3g',\n",
       " 'avg_rech_amt',\n",
       " 'churn']"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# List of columns with outliers\n",
    "out_list = []\n",
    "for col in df_high_val_imputed:\n",
    "    IQR = df_high_val_imputed[col].quantile(0.75) - df_high_val_imputed[col].quantile(0.5)\n",
    "    if df_high_val_imputed[col].max()  >  df_high_val_imputed[col].quantile(0.75) + 1.5*IQR:\n",
    "        out_list.append(col)\n",
    "out_list"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Modelling"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "# Putting feature variable to X\n",
    "X = df_high_val_imputed.drop(['churn'],axis=1)\n",
    "\n",
    "# Putting response variable to y\n",
    "y = df_high_val_imputed['churn']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Feature Standardization**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [],
   "source": [
    "normalized_X = (X-X.mean())/X.std()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Splitting the data into train and test\n",
    "X_train, X_test, y_train, y_test = train_test_split(normalized_X, y, stratify=y, train_size=0.7,test_size=0.3,random_state=100)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Prediction using PCA"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.decomposition import PCA\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn import metrics"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [],
   "source": [
    "pca = PCA(0.9)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_train_pca = pca.fit_transform(X_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(21007, 51)"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_train_pca.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 0.17239538,  0.19334366,  0.19309593, ...,  0.04279805,\n",
       "         0.02414144,  0.19990799],\n",
       "       [ 0.10337755,  0.13055398,  0.1125503 , ..., -0.0550364 ,\n",
       "        -0.01305077,  0.12841284],\n",
       "       [ 0.13591175,  0.15522281,  0.14198261, ...,  0.16552445,\n",
       "         0.05371914,  0.1614639 ],\n",
       "       ...,\n",
       "       [ 0.01376578,  0.00868891,  0.03052416, ...,  0.22173834,\n",
       "        -0.05357231,  0.00290679],\n",
       "       [ 0.02349623, -0.0628219 ,  0.08108614, ..., -0.02680103,\n",
       "        -0.09036594, -0.01462256],\n",
       "       [ 0.01928314,  0.02725527,  0.00289483, ...,  0.00405342,\n",
       "        -0.0304203 ,  0.02394918]])"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pca.components_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.09935914, 0.08599588, 0.06082667, 0.04622431, 0.0388145 ,\n",
       "       0.03548485, 0.03419657, 0.02698564, 0.02682936, 0.02528813,\n",
       "       0.02438988, 0.02346342, 0.0221981 , 0.02182124, 0.02059907,\n",
       "       0.01854065, 0.01799106, 0.01622217, 0.01450871, 0.01355473,\n",
       "       0.01324722, 0.0128263 , 0.01171314, 0.0110843 , 0.01062756,\n",
       "       0.00991381, 0.00954316, 0.0091771 , 0.00855625, 0.00821418,\n",
       "       0.00813792, 0.0076926 , 0.00746489, 0.00725879, 0.00717252,\n",
       "       0.00679253, 0.00661354, 0.00646007, 0.00619653, 0.00607072,\n",
       "       0.00589991, 0.00567913, 0.00564531, 0.00547843, 0.00517971,\n",
       "       0.00486636, 0.00481998, 0.00465068, 0.00429689, 0.0040401 ,\n",
       "       0.00398166])"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pca.explained_variance_ratio_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [],
   "source": [
    "var_cumu = np.cumsum(pca.explained_variance_ratio_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Scree plot\n",
    "fig = plt.figure(figsize=[12,8])\n",
    "plt.vlines(x=15, ymax=1, ymin=0, colors=\"r\", linestyles=\"--\")\n",
    "plt.hlines(y=0.95, xmax=30, xmin=0, colors=\"g\", linestyles=\"--\")\n",
    "plt.plot(var_cumu)\n",
    "plt.ylabel(\"Cumulative variance explained\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [],
   "source": [
    "corrmat = np.corrcoef(df_train_pca.transpose())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [],
   "source": [
    "learner_pca = LogisticRegression()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [],
   "source": [
    "model_pca = learner_pca.fit(df_train_pca, y_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Making Predictions**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_test_pca = pca.transform(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(9004, 51)"
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_test_pca.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "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>0</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.000819</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.066795</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.025579</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.038265</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.134676</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          0\n",
       "0  0.000819\n",
       "1  0.066795\n",
       "2  0.025579\n",
       "3  0.038265\n",
       "4  0.134676"
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_probs_test = model_pca.predict_proba(df_test_pca)[:,1]\n",
    "y_pred_1 = pd.DataFrame(pred_probs_test)\n",
    "y_pred_1.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "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>churn</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>23530</th>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4825</th>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14228</th>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21259</th>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23625</th>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       churn\n",
       "23530    0.0\n",
       "4825     0.0\n",
       "14228    0.0\n",
       "21259    0.0\n",
       "23625    0.0"
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Converting y_test to dataframe\n",
    "y_test_df = pd.DataFrame(y_test)\n",
    "y_test_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [
    {
     "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>CustID</th>\n",
       "      <th>churn</th>\n",
       "      <th>Churn_Prob</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>23530</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000819</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>4825</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.066795</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>14228</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.025579</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>21259</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.038265</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>23625</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.134676</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   CustID  churn  Churn_Prob\n",
       "0   23530    0.0    0.000819\n",
       "1    4825    0.0    0.066795\n",
       "2   14228    0.0    0.025579\n",
       "3   21259    0.0    0.038265\n",
       "4   23625    0.0    0.134676"
      ]
     },
     "execution_count": 57,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Putting CustID to index\n",
    "y_test_df['CustID'] = y_test_df.index\n",
    "# Removing index for both dataframes to append them side by side \n",
    "y_pred_1.reset_index(drop=True, inplace=True)\n",
    "y_test_df.reset_index(drop=True, inplace=True)\n",
    "# Appending y_test_df and y_pred_1\n",
    "y_pred_final = pd.concat([y_test_df,y_pred_1],axis=1)\n",
    "# Renaming the column \n",
    "y_pred_final= y_pred_final.rename(columns={ 0 : 'Churn_Prob'})\n",
    "# Rearranging the columns\n",
    "y_pred_final = y_pred_final.reindex(['CustID','churn','Churn_Prob'], axis=1)\n",
    "# Let's see the head of y_pred_final\n",
    "y_pred_final.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [
    {
     "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>CustID</th>\n",
       "      <th>churn</th>\n",
       "      <th>Churn_Prob</th>\n",
       "      <th>predicted</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>23530</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000819</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>4825</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.066795</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>14228</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.025579</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>21259</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.038265</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>23625</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.134676</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   CustID  churn  Churn_Prob  predicted\n",
       "0   23530    0.0    0.000819          0\n",
       "1    4825    0.0    0.066795          0\n",
       "2   14228    0.0    0.025579          0\n",
       "3   21259    0.0    0.038265          0\n",
       "4   23625    0.0    0.134676          0"
      ]
     },
     "execution_count": 58,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Creating new column 'predicted' with 1 if Churn_Prob>0.5 else 0\n",
    "y_pred_final['predicted'] = y_pred_final.Churn_Prob.map( lambda x: 1 if x > 0.5 else 0)\n",
    "# Let's see the head\n",
    "y_pred_final.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Model Evaluation**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[8083,  108],\n",
       "       [ 610,  203]], dtype=int64)"
      ]
     },
     "execution_count": 59,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Confusion matrix \n",
    "confusion = metrics.confusion_matrix( y_pred_final.churn, y_pred_final.predicted )\n",
    "confusion"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.920257663260773"
      ]
     },
     "execution_count": 60,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "metrics.accuracy_score(y_pred_final.churn, y_pred_final.predicted)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.6527331189710611"
      ]
     },
     "execution_count": 61,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "metrics.precision_score(y_pred_final.churn, y_pred_final.predicted)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [],
   "source": [
    "def draw_roc( actual, probs ):\n",
    "    fpr, tpr, thresholds = metrics.roc_curve( actual, probs,\n",
    "                                              drop_intermediate = False )\n",
    "    auc_score = metrics.roc_auc_score( actual, probs )\n",
    "    plt.figure(figsize=(6, 6))\n",
    "    plt.plot( fpr, tpr, label='ROC curve (area = %0.2f)' % auc_score )\n",
    "    plt.plot([0, 1], [0, 1], 'k--')\n",
    "    plt.xlim([0.0, 1.0])\n",
    "    plt.ylim([0.0, 1.05])\n",
    "    plt.xlabel('False Positive Rate or [1 - True Negative Rate]')\n",
    "    plt.ylabel('True Positive Rate')\n",
    "    plt.title('Receiver operating characteristic example')\n",
    "    plt.legend(loc=\"lower right\")\n",
    "    plt.show()\n",
    "\n",
    "    return fpr, tpr, thresholds"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "(array([0.       , 0.0131852, 1.       ]),\n",
       " array([0.       , 0.2496925, 1.       ]),\n",
       " array([2, 1, 0], dtype=int64))"
      ]
     },
     "execution_count": 63,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "draw_roc(y_pred_final.churn, y_pred_final.predicted)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'0.88'"
      ]
     },
     "execution_count": 64,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\"{:2.2}\".format(metrics.roc_auc_score(y_test, pred_probs_test))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Feature Selection"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Splitting the data into train and test\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, train_size=0.7,test_size=0.3,random_state=100)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "import statsmodels.api as sm"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table class=\"simpletable\">\n",
       "<caption>Generalized Linear Model Regression Results</caption>\n",
       "<tr>\n",
       "  <th>Dep. Variable:</th>         <td>churn</td>      <th>  No. Observations:  </th>  <td> 21007</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Model:</th>                  <td>GLM</td>       <th>  Df Residuals:      </th>  <td> 20881</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Model Family:</th>        <td>Binomial</td>     <th>  Df Model:          </th>  <td>   125</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Link Function:</th>         <td>logit</td>      <th>  Scale:             </th> <td>  1.0000</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Method:</th>                <td>IRLS</td>       <th>  Log-Likelihood:    </th> <td> -3972.0</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Date:</th>            <td>Wed, 17 Feb 2021</td> <th>  Deviance:          </th> <td>  7944.1</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Time:</th>                <td>18:34:22</td>     <th>  Pearson chi2:      </th> <td>1.70e+05</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>No. Iterations:</th>          <td>9</td>        <th>                     </th>     <td> </td>   \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Covariance Type:</th>     <td>nonrobust</td>    <th>                     </th>     <td> </td>   \n",
       "</tr>\n",
       "</table>\n",
       "<table class=\"simpletable\">\n",
       "<tr>\n",
       "           <td></td>             <th>coef</th>     <th>std err</th>      <th>z</th>      <th>P>|z|</th>  <th>[0.025</th>    <th>0.975]</th>  \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>const</th>              <td>   -0.4706</td> <td>    0.110</td> <td>   -4.272</td> <td> 0.000</td> <td>   -0.686</td> <td>   -0.255</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>arpu_6</th>             <td>    0.0003</td> <td>    0.000</td> <td>    1.171</td> <td> 0.242</td> <td>   -0.000</td> <td>    0.001</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>arpu_7</th>             <td>    0.0005</td> <td>    0.000</td> <td>    1.709</td> <td> 0.087</td> <td>-7.46e-05</td> <td>    0.001</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>arpu_8</th>             <td>    0.0006</td> <td>    0.000</td> <td>    1.710</td> <td> 0.087</td> <td>-9.02e-05</td> <td>    0.001</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>onnet_mou_6</th>        <td>   -0.0032</td> <td>    0.006</td> <td>   -0.539</td> <td> 0.590</td> <td>   -0.015</td> <td>    0.009</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>onnet_mou_7</th>        <td>   -0.0041</td> <td>    0.004</td> <td>   -0.907</td> <td> 0.364</td> <td>   -0.013</td> <td>    0.005</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>onnet_mou_8</th>        <td>    0.0009</td> <td>    0.004</td> <td>    0.250</td> <td> 0.802</td> <td>   -0.006</td> <td>    0.008</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>offnet_mou_6</th>       <td>   -0.0024</td> <td>    0.006</td> <td>   -0.403</td> <td> 0.687</td> <td>   -0.014</td> <td>    0.009</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>offnet_mou_7</th>       <td>   -0.0046</td> <td>    0.004</td> <td>   -1.042</td> <td> 0.297</td> <td>   -0.013</td> <td>    0.004</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>offnet_mou_8</th>       <td>   -0.0003</td> <td>    0.004</td> <td>   -0.091</td> <td> 0.927</td> <td>   -0.007</td> <td>    0.007</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>roam_ic_mou_6</th>      <td>    0.0004</td> <td>    0.000</td> <td>    0.941</td> <td> 0.347</td> <td>   -0.000</td> <td>    0.001</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>roam_ic_mou_7</th>      <td>    0.0001</td> <td>    0.000</td> <td>    0.231</td> <td> 0.817</td> <td>   -0.001</td> <td>    0.001</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>roam_ic_mou_8</th>      <td>    0.0004</td> <td>    0.000</td> <td>    0.819</td> <td> 0.413</td> <td>   -0.000</td> <td>    0.001</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>roam_og_mou_6</th>      <td>    0.0029</td> <td>    0.006</td> <td>    0.489</td> <td> 0.625</td> <td>   -0.009</td> <td>    0.015</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>roam_og_mou_7</th>      <td>    0.0053</td> <td>    0.004</td> <td>    1.191</td> <td> 0.234</td> <td>   -0.003</td> <td>    0.014</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>roam_og_mou_8</th>      <td> -9.64e-05</td> <td>    0.004</td> <td>   -0.027</td> <td> 0.978</td> <td>   -0.007</td> <td>    0.007</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>loc_og_t2t_mou_6</th>   <td>   -0.0141</td> <td>    5.391</td> <td>   -0.003</td> <td> 0.998</td> <td>  -10.581</td> <td>   10.553</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>loc_og_t2t_mou_7</th>   <td>    7.7659</td> <td>    5.376</td> <td>    1.444</td> <td> 0.149</td> <td>   -2.772</td> <td>   18.303</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>loc_og_t2t_mou_8</th>   <td>   27.0011</td> <td>    6.093</td> <td>    4.432</td> <td> 0.000</td> <td>   15.059</td> <td>   38.943</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>loc_og_t2m_mou_6</th>   <td>   -0.0159</td> <td>    5.391</td> <td>   -0.003</td> <td> 0.998</td> <td>  -10.583</td> <td>   10.551</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>loc_og_t2m_mou_7</th>   <td>    7.7663</td> <td>    5.376</td> <td>    1.445</td> <td> 0.149</td> <td>   -2.771</td> <td>   18.304</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>loc_og_t2m_mou_8</th>   <td>   27.0035</td> <td>    6.093</td> <td>    4.432</td> <td> 0.000</td> <td>   15.062</td> <td>   38.945</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>loc_og_t2f_mou_6</th>   <td>   -0.0151</td> <td>    5.392</td> <td>   -0.003</td> <td> 0.998</td> <td>  -10.583</td> <td>   10.553</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>loc_og_t2f_mou_7</th>   <td>    7.7669</td> <td>    5.377</td> <td>    1.445</td> <td> 0.149</td> <td>   -2.771</td> <td>   18.305</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>loc_og_t2f_mou_8</th>   <td>   27.0067</td> <td>    6.093</td> <td>    4.432</td> <td> 0.000</td> <td>   15.064</td> <td>   38.949</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>loc_og_t2c_mou_6</th>   <td>   -0.0007</td> <td>    0.004</td> <td>   -0.152</td> <td> 0.879</td> <td>   -0.009</td> <td>    0.008</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>loc_og_t2c_mou_7</th>   <td>    0.0090</td> <td>    0.005</td> <td>    1.867</td> <td> 0.062</td> <td>   -0.000</td> <td>    0.018</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>loc_og_t2c_mou_8</th>   <td>    0.0018</td> <td>    0.006</td> <td>    0.284</td> <td> 0.776</td> <td>   -0.011</td> <td>    0.014</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>loc_og_mou_6</th>       <td>    0.0208</td> <td>    5.391</td> <td>    0.004</td> <td> 0.997</td> <td>  -10.546</td> <td>   10.588</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>loc_og_mou_7</th>       <td>   -7.7559</td> <td>    5.376</td> <td>   -1.443</td> <td> 0.149</td> <td>  -18.294</td> <td>    2.782</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>loc_og_mou_8</th>       <td>  -27.0010</td> <td>    6.093</td> <td>   -4.432</td> <td> 0.000</td> <td>  -38.943</td> <td>  -15.059</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>std_og_t2t_mou_6</th>   <td>    1.9890</td> <td>    6.239</td> <td>    0.319</td> <td> 0.750</td> <td>  -10.238</td> <td>   14.216</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>std_og_t2t_mou_7</th>   <td>   -3.8984</td> <td>    6.352</td> <td>   -0.614</td> <td> 0.539</td> <td>  -16.347</td> <td>    8.551</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>std_og_t2t_mou_8</th>   <td>   26.9068</td> <td>    7.451</td> <td>    3.611</td> <td> 0.000</td> <td>   12.304</td> <td>   41.510</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>std_og_t2m_mou_6</th>   <td>    1.9882</td> <td>    6.239</td> <td>    0.319</td> <td> 0.750</td> <td>  -10.239</td> <td>   14.215</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>std_og_t2m_mou_7</th>   <td>   -3.8975</td> <td>    6.352</td> <td>   -0.614</td> <td> 0.539</td> <td>  -16.347</td> <td>    8.552</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>std_og_t2m_mou_8</th>   <td>   26.9075</td> <td>    7.451</td> <td>    3.611</td> <td> 0.000</td> <td>   12.304</td> <td>   41.511</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>std_og_t2f_mou_6</th>   <td>    1.9920</td> <td>    6.239</td> <td>    0.319</td> <td> 0.750</td> <td>  -10.236</td> <td>   14.220</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>std_og_t2f_mou_7</th>   <td>   -3.9050</td> <td>    6.352</td> <td>   -0.615</td> <td> 0.539</td> <td>  -16.355</td> <td>    8.545</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>std_og_t2f_mou_8</th>   <td>   26.9073</td> <td>    7.451</td> <td>    3.611</td> <td> 0.000</td> <td>   12.303</td> <td>   41.511</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>std_og_mou_6</th>       <td>   -1.9827</td> <td>    6.239</td> <td>   -0.318</td> <td> 0.751</td> <td>  -14.210</td> <td>   10.245</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>std_og_mou_7</th>       <td>    3.9083</td> <td>    6.352</td> <td>    0.615</td> <td> 0.538</td> <td>   -8.541</td> <td>   16.358</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>std_og_mou_8</th>       <td>  -26.9049</td> <td>    7.451</td> <td>   -3.611</td> <td> 0.000</td> <td>  -41.508</td> <td>  -12.302</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>isd_og_mou_6</th>       <td>    0.0040</td> <td>    0.007</td> <td>    0.615</td> <td> 0.539</td> <td>   -0.009</td> <td>    0.017</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>isd_og_mou_7</th>       <td>    0.0125</td> <td>    0.006</td> <td>    2.225</td> <td> 0.026</td> <td>    0.001</td> <td>    0.024</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>isd_og_mou_8</th>       <td>    0.0021</td> <td>    0.005</td> <td>    0.442</td> <td> 0.658</td> <td>   -0.007</td> <td>    0.012</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>spl_og_mou_6</th>       <td>    0.0054</td> <td>    0.006</td> <td>    0.903</td> <td> 0.366</td> <td>   -0.006</td> <td>    0.017</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>spl_og_mou_7</th>       <td>    0.0084</td> <td>    0.005</td> <td>    1.807</td> <td> 0.071</td> <td>   -0.001</td> <td>    0.017</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>spl_og_mou_8</th>       <td>    0.0011</td> <td>    0.004</td> <td>    0.250</td> <td> 0.803</td> <td>   -0.007</td> <td>    0.010</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>og_others_6</th>        <td>   -0.0307</td> <td>    0.015</td> <td>   -2.113</td> <td> 0.035</td> <td>   -0.059</td> <td>   -0.002</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>og_others_7</th>        <td>    0.0074</td> <td>    0.015</td> <td>    0.486</td> <td> 0.627</td> <td>   -0.022</td> <td>    0.037</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>og_others_8</th>        <td>    0.0128</td> <td>    0.018</td> <td>    0.714</td> <td> 0.475</td> <td>   -0.022</td> <td>    0.048</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>total_og_mou_6</th>     <td>   -0.0029</td> <td>    0.002</td> <td>   -1.486</td> <td> 0.137</td> <td>   -0.007</td> <td>    0.001</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>total_og_mou_7</th>     <td>   -0.0058</td> <td>    0.002</td> <td>   -2.520</td> <td> 0.012</td> <td>   -0.010</td> <td>   -0.001</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>total_og_mou_8</th>     <td>   -0.0033</td> <td>    0.002</td> <td>   -2.035</td> <td> 0.042</td> <td>   -0.007</td> <td>   -0.000</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>loc_ic_t2t_mou_6</th>   <td>   -8.2573</td> <td>    5.217</td> <td>   -1.583</td> <td> 0.113</td> <td>  -18.482</td> <td>    1.968</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>loc_ic_t2t_mou_7</th>   <td>   21.6690</td> <td>    5.260</td> <td>    4.120</td> <td> 0.000</td> <td>   11.360</td> <td>   31.978</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>loc_ic_t2t_mou_8</th>   <td>   39.2584</td> <td>    5.708</td> <td>    6.878</td> <td> 0.000</td> <td>   28.071</td> <td>   50.445</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>loc_ic_t2m_mou_6</th>   <td>   -8.2570</td> <td>    5.217</td> <td>   -1.583</td> <td> 0.113</td> <td>  -18.482</td> <td>    1.968</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>loc_ic_t2m_mou_7</th>   <td>   21.6703</td> <td>    5.260</td> <td>    4.120</td> <td> 0.000</td> <td>   11.361</td> <td>   31.980</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>loc_ic_t2m_mou_8</th>   <td>   39.2566</td> <td>    5.708</td> <td>    6.878</td> <td> 0.000</td> <td>   28.070</td> <td>   50.444</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>loc_ic_t2f_mou_6</th>   <td>   -8.2579</td> <td>    5.217</td> <td>   -1.583</td> <td> 0.113</td> <td>  -18.483</td> <td>    1.967</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>loc_ic_t2f_mou_7</th>   <td>   21.6695</td> <td>    5.260</td> <td>    4.120</td> <td> 0.000</td> <td>   11.360</td> <td>   31.979</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>loc_ic_t2f_mou_8</th>   <td>   39.2570</td> <td>    5.708</td> <td>    6.878</td> <td> 0.000</td> <td>   28.069</td> <td>   50.445</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>loc_ic_mou_6</th>       <td>    8.2586</td> <td>    5.217</td> <td>    1.583</td> <td> 0.113</td> <td>   -1.966</td> <td>   18.483</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>loc_ic_mou_7</th>       <td>  -21.6694</td> <td>    5.260</td> <td>   -4.120</td> <td> 0.000</td> <td>  -31.978</td> <td>  -11.360</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>loc_ic_mou_8</th>       <td>  -39.2392</td> <td>    5.708</td> <td>   -6.875</td> <td> 0.000</td> <td>  -50.426</td> <td>  -28.052</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>std_ic_t2t_mou_6</th>   <td>  -14.2072</td> <td>    6.160</td> <td>   -2.306</td> <td> 0.021</td> <td>  -26.280</td> <td>   -2.134</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>std_ic_t2t_mou_7</th>   <td>   -4.6899</td> <td>    6.424</td> <td>   -0.730</td> <td> 0.465</td> <td>  -17.281</td> <td>    7.901</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>std_ic_t2t_mou_8</th>   <td>   11.3391</td> <td>    7.696</td> <td>    1.473</td> <td> 0.141</td> <td>   -3.745</td> <td>   26.424</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>std_ic_t2m_mou_6</th>   <td>  -14.2084</td> <td>    6.160</td> <td>   -2.307</td> <td> 0.021</td> <td>  -26.282</td> <td>   -2.135</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>std_ic_t2m_mou_7</th>   <td>   -4.6915</td> <td>    6.424</td> <td>   -0.730</td> <td> 0.465</td> <td>  -17.283</td> <td>    7.900</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>std_ic_t2m_mou_8</th>   <td>   11.3441</td> <td>    7.696</td> <td>    1.474</td> <td> 0.140</td> <td>   -3.740</td> <td>   26.428</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>std_ic_t2f_mou_6</th>   <td>  -14.2085</td> <td>    6.160</td> <td>   -2.306</td> <td> 0.021</td> <td>  -26.282</td> <td>   -2.135</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>std_ic_t2f_mou_7</th>   <td>   -4.6884</td> <td>    6.424</td> <td>   -0.730</td> <td> 0.466</td> <td>  -17.280</td> <td>    7.903</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>std_ic_t2f_mou_8</th>   <td>   11.3408</td> <td>    7.697</td> <td>    1.473</td> <td> 0.141</td> <td>   -3.744</td> <td>   26.426</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>std_ic_mou_6</th>       <td>   14.2099</td> <td>    6.160</td> <td>    2.307</td> <td> 0.021</td> <td>    2.137</td> <td>   26.283</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>std_ic_mou_7</th>       <td>    4.6899</td> <td>    6.424</td> <td>    0.730</td> <td> 0.465</td> <td>   -7.901</td> <td>   17.281</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>std_ic_mou_8</th>       <td>  -11.3220</td> <td>    7.696</td> <td>   -1.471</td> <td> 0.141</td> <td>  -26.406</td> <td>    3.762</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>total_ic_mou_6</th>     <td>   -0.0015</td> <td>    0.003</td> <td>   -0.577</td> <td> 0.564</td> <td>   -0.007</td> <td>    0.004</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>total_ic_mou_7</th>     <td>    0.0013</td> <td>    0.003</td> <td>    0.466</td> <td> 0.641</td> <td>   -0.004</td> <td>    0.007</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>total_ic_mou_8</th>     <td>   -0.0229</td> <td>    0.002</td> <td>   -9.761</td> <td> 0.000</td> <td>   -0.027</td> <td>   -0.018</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>spl_ic_mou_6</th>       <td>    0.0787</td> <td>    0.193</td> <td>    0.409</td> <td> 0.683</td> <td>   -0.299</td> <td>    0.456</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>spl_ic_mou_7</th>       <td>   -0.2649</td> <td>    0.342</td> <td>   -0.774</td> <td> 0.439</td> <td>   -0.936</td> <td>    0.406</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>spl_ic_mou_8</th>       <td>   -3.2593</td> <td>    0.502</td> <td>   -6.488</td> <td> 0.000</td> <td>   -4.244</td> <td>   -2.275</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>isd_ic_mou_6</th>       <td>    0.0021</td> <td>    0.003</td> <td>    0.787</td> <td> 0.431</td> <td>   -0.003</td> <td>    0.007</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>isd_ic_mou_7</th>       <td>   -0.0004</td> <td>    0.003</td> <td>   -0.142</td> <td> 0.887</td> <td>   -0.006</td> <td>    0.005</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>isd_ic_mou_8</th>       <td>    0.0215</td> <td>    0.003</td> <td>    8.290</td> <td> 0.000</td> <td>    0.016</td> <td>    0.027</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>ic_others_6</th>        <td>    0.0007</td> <td>    0.006</td> <td>    0.113</td> <td> 0.910</td> <td>   -0.011</td> <td>    0.013</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>ic_others_7</th>        <td>   -0.0069</td> <td>    0.008</td> <td>   -0.896</td> <td> 0.370</td> <td>   -0.022</td> <td>    0.008</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>ic_others_8</th>        <td>    0.0219</td> <td>    0.008</td> <td>    2.741</td> <td> 0.006</td> <td>    0.006</td> <td>    0.038</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>total_rech_num_6</th>   <td>   -0.0021</td> <td>    0.005</td> <td>   -0.454</td> <td> 0.650</td> <td>   -0.011</td> <td>    0.007</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>total_rech_num_7</th>   <td>    0.0146</td> <td>    0.006</td> <td>    2.281</td> <td> 0.023</td> <td>    0.002</td> <td>    0.027</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>total_rech_num_8</th>   <td>   -0.0449</td> <td>    0.008</td> <td>   -5.592</td> <td> 0.000</td> <td>   -0.061</td> <td>   -0.029</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>total_rech_amt_6</th>   <td>    0.0002</td> <td>    0.000</td> <td>    0.773</td> <td> 0.439</td> <td>   -0.000</td> <td>    0.001</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>total_rech_amt_7</th>   <td>   -0.0003</td> <td>    0.000</td> <td>   -1.208</td> <td> 0.227</td> <td>   -0.001</td> <td>    0.000</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>total_rech_amt_8</th>   <td>   -0.0006</td> <td>    0.000</td> <td>   -1.966</td> <td> 0.049</td> <td>   -0.001</td> <td>-1.85e-06</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>max_rech_amt_6</th>     <td>   -0.0002</td> <td>    0.000</td> <td>   -0.614</td> <td> 0.539</td> <td>   -0.001</td> <td>    0.000</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>max_rech_amt_7</th>     <td> 7.392e-05</td> <td>    0.000</td> <td>    0.252</td> <td> 0.801</td> <td>   -0.000</td> <td>    0.001</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>max_rech_amt_8</th>     <td>    0.0007</td> <td>    0.000</td> <td>    1.867</td> <td> 0.062</td> <td>-3.49e-05</td> <td>    0.001</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>last_day_rch_amt_6</th> <td>   -0.0002</td> <td>    0.000</td> <td>   -0.524</td> <td> 0.601</td> <td>   -0.001</td> <td>    0.000</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>last_day_rch_amt_7</th> <td>   -0.0004</td> <td>    0.000</td> <td>   -1.279</td> <td> 0.201</td> <td>   -0.001</td> <td>    0.000</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>last_day_rch_amt_8</th> <td>   -0.0042</td> <td>    0.000</td> <td>   -8.526</td> <td> 0.000</td> <td>   -0.005</td> <td>   -0.003</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>vol_2g_mb_6</th>        <td> 7.671e-05</td> <td>    0.000</td> <td>    0.468</td> <td> 0.640</td> <td>   -0.000</td> <td>    0.000</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>vol_2g_mb_7</th>        <td>    0.0003</td> <td>    0.000</td> <td>    1.806</td> <td> 0.071</td> <td>-2.82e-05</td> <td>    0.001</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>vol_2g_mb_8</th>        <td>   -0.0006</td> <td>    0.000</td> <td>   -2.230</td> <td> 0.026</td> <td>   -0.001</td> <td>-6.89e-05</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>vol_3g_mb_6</th>        <td>   -0.0001</td> <td>  8.5e-05</td> <td>   -1.558</td> <td> 0.119</td> <td>   -0.000</td> <td> 3.41e-05</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>vol_3g_mb_7</th>        <td> 7.421e-05</td> <td> 9.98e-05</td> <td>    0.744</td> <td> 0.457</td> <td>   -0.000</td> <td>    0.000</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>vol_3g_mb_8</th>        <td>   -0.0002</td> <td>    0.000</td> <td>   -1.497</td> <td> 0.134</td> <td>   -0.000</td> <td> 5.97e-05</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>monthly_2g_6</th>       <td>   -0.2986</td> <td>    0.148</td> <td>   -2.013</td> <td> 0.044</td> <td>   -0.589</td> <td>   -0.008</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>monthly_2g_7</th>       <td>   -0.1963</td> <td>    0.150</td> <td>   -1.313</td> <td> 0.189</td> <td>   -0.489</td> <td>    0.097</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>monthly_2g_8</th>       <td>   -0.9160</td> <td>    0.225</td> <td>   -4.069</td> <td> 0.000</td> <td>   -1.357</td> <td>   -0.475</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>sachet_2g_6</th>        <td>    0.0755</td> <td>    0.028</td> <td>    2.688</td> <td> 0.007</td> <td>    0.020</td> <td>    0.131</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>sachet_2g_7</th>        <td>   -0.0126</td> <td>    0.029</td> <td>   -0.442</td> <td> 0.659</td> <td>   -0.069</td> <td>    0.043</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>sachet_2g_8</th>        <td>   -0.1185</td> <td>    0.036</td> <td>   -3.266</td> <td> 0.001</td> <td>   -0.190</td> <td>   -0.047</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>monthly_3g_6</th>       <td>    0.0007</td> <td>    0.107</td> <td>    0.006</td> <td> 0.995</td> <td>   -0.210</td> <td>    0.211</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>monthly_3g_7</th>       <td>   -0.0353</td> <td>    0.116</td> <td>   -0.305</td> <td> 0.760</td> <td>   -0.262</td> <td>    0.192</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>monthly_3g_8</th>       <td>   -0.7178</td> <td>    0.179</td> <td>   -4.014</td> <td> 0.000</td> <td>   -1.068</td> <td>   -0.367</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>sachet_3g_6</th>        <td>    0.0204</td> <td>    0.054</td> <td>    0.375</td> <td> 0.708</td> <td>   -0.086</td> <td>    0.127</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>sachet_3g_7</th>        <td>    0.1558</td> <td>    0.061</td> <td>    2.562</td> <td> 0.010</td> <td>    0.037</td> <td>    0.275</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>sachet_3g_8</th>        <td>   -0.1877</td> <td>    0.090</td> <td>   -2.085</td> <td> 0.037</td> <td>   -0.364</td> <td>   -0.011</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>aon</th>                <td>   -0.0003</td> <td> 4.23e-05</td> <td>   -6.163</td> <td> 0.000</td> <td>   -0.000</td> <td>   -0.000</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>aug_vbc_3g</th>         <td> 4.788e-05</td> <td>    0.000</td> <td>    0.222</td> <td> 0.824</td> <td>   -0.000</td> <td>    0.000</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>jul_vbc_3g</th>         <td>   -0.0002</td> <td>    0.000</td> <td>   -0.993</td> <td> 0.321</td> <td>   -0.001</td> <td>    0.000</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>jun_vbc_3g</th>         <td>    0.0002</td> <td>    0.000</td> <td>    1.237</td> <td> 0.216</td> <td>   -0.000</td> <td>    0.001</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>sep_vbc_3g</th>         <td>   -0.0127</td> <td>    0.004</td> <td>   -3.137</td> <td> 0.002</td> <td>   -0.021</td> <td>   -0.005</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>avg_rech_amt</th>       <td>-5.299e-05</td> <td>    0.000</td> <td>   -0.428</td> <td> 0.669</td> <td>   -0.000</td> <td>    0.000</td>\n",
       "</tr>\n",
       "</table>"
      ],
      "text/plain": [
       "<class 'statsmodels.iolib.summary.Summary'>\n",
       "\"\"\"\n",
       "                 Generalized Linear Model Regression Results                  \n",
       "==============================================================================\n",
       "Dep. Variable:                  churn   No. Observations:                21007\n",
       "Model:                            GLM   Df Residuals:                    20881\n",
       "Model Family:                Binomial   Df Model:                          125\n",
       "Link Function:                  logit   Scale:                          1.0000\n",
       "Method:                          IRLS   Log-Likelihood:                -3972.0\n",
       "Date:                Wed, 17 Feb 2021   Deviance:                       7944.1\n",
       "Time:                        18:34:22   Pearson chi2:                 1.70e+05\n",
       "No. Iterations:                     9                                         \n",
       "Covariance Type:            nonrobust                                         \n",
       "======================================================================================\n",
       "                         coef    std err          z      P>|z|      [0.025      0.975]\n",
       "--------------------------------------------------------------------------------------\n",
       "const                 -0.4706      0.110     -4.272      0.000      -0.686      -0.255\n",
       "arpu_6                 0.0003      0.000      1.171      0.242      -0.000       0.001\n",
       "arpu_7                 0.0005      0.000      1.709      0.087   -7.46e-05       0.001\n",
       "arpu_8                 0.0006      0.000      1.710      0.087   -9.02e-05       0.001\n",
       "onnet_mou_6           -0.0032      0.006     -0.539      0.590      -0.015       0.009\n",
       "onnet_mou_7           -0.0041      0.004     -0.907      0.364      -0.013       0.005\n",
       "onnet_mou_8            0.0009      0.004      0.250      0.802      -0.006       0.008\n",
       "offnet_mou_6          -0.0024      0.006     -0.403      0.687      -0.014       0.009\n",
       "offnet_mou_7          -0.0046      0.004     -1.042      0.297      -0.013       0.004\n",
       "offnet_mou_8          -0.0003      0.004     -0.091      0.927      -0.007       0.007\n",
       "roam_ic_mou_6          0.0004      0.000      0.941      0.347      -0.000       0.001\n",
       "roam_ic_mou_7          0.0001      0.000      0.231      0.817      -0.001       0.001\n",
       "roam_ic_mou_8          0.0004      0.000      0.819      0.413      -0.000       0.001\n",
       "roam_og_mou_6          0.0029      0.006      0.489      0.625      -0.009       0.015\n",
       "roam_og_mou_7          0.0053      0.004      1.191      0.234      -0.003       0.014\n",
       "roam_og_mou_8       -9.64e-05      0.004     -0.027      0.978      -0.007       0.007\n",
       "loc_og_t2t_mou_6      -0.0141      5.391     -0.003      0.998     -10.581      10.553\n",
       "loc_og_t2t_mou_7       7.7659      5.376      1.444      0.149      -2.772      18.303\n",
       "loc_og_t2t_mou_8      27.0011      6.093      4.432      0.000      15.059      38.943\n",
       "loc_og_t2m_mou_6      -0.0159      5.391     -0.003      0.998     -10.583      10.551\n",
       "loc_og_t2m_mou_7       7.7663      5.376      1.445      0.149      -2.771      18.304\n",
       "loc_og_t2m_mou_8      27.0035      6.093      4.432      0.000      15.062      38.945\n",
       "loc_og_t2f_mou_6      -0.0151      5.392     -0.003      0.998     -10.583      10.553\n",
       "loc_og_t2f_mou_7       7.7669      5.377      1.445      0.149      -2.771      18.305\n",
       "loc_og_t2f_mou_8      27.0067      6.093      4.432      0.000      15.064      38.949\n",
       "loc_og_t2c_mou_6      -0.0007      0.004     -0.152      0.879      -0.009       0.008\n",
       "loc_og_t2c_mou_7       0.0090      0.005      1.867      0.062      -0.000       0.018\n",
       "loc_og_t2c_mou_8       0.0018      0.006      0.284      0.776      -0.011       0.014\n",
       "loc_og_mou_6           0.0208      5.391      0.004      0.997     -10.546      10.588\n",
       "loc_og_mou_7          -7.7559      5.376     -1.443      0.149     -18.294       2.782\n",
       "loc_og_mou_8         -27.0010      6.093     -4.432      0.000     -38.943     -15.059\n",
       "std_og_t2t_mou_6       1.9890      6.239      0.319      0.750     -10.238      14.216\n",
       "std_og_t2t_mou_7      -3.8984      6.352     -0.614      0.539     -16.347       8.551\n",
       "std_og_t2t_mou_8      26.9068      7.451      3.611      0.000      12.304      41.510\n",
       "std_og_t2m_mou_6       1.9882      6.239      0.319      0.750     -10.239      14.215\n",
       "std_og_t2m_mou_7      -3.8975      6.352     -0.614      0.539     -16.347       8.552\n",
       "std_og_t2m_mou_8      26.9075      7.451      3.611      0.000      12.304      41.511\n",
       "std_og_t2f_mou_6       1.9920      6.239      0.319      0.750     -10.236      14.220\n",
       "std_og_t2f_mou_7      -3.9050      6.352     -0.615      0.539     -16.355       8.545\n",
       "std_og_t2f_mou_8      26.9073      7.451      3.611      0.000      12.303      41.511\n",
       "std_og_mou_6          -1.9827      6.239     -0.318      0.751     -14.210      10.245\n",
       "std_og_mou_7           3.9083      6.352      0.615      0.538      -8.541      16.358\n",
       "std_og_mou_8         -26.9049      7.451     -3.611      0.000     -41.508     -12.302\n",
       "isd_og_mou_6           0.0040      0.007      0.615      0.539      -0.009       0.017\n",
       "isd_og_mou_7           0.0125      0.006      2.225      0.026       0.001       0.024\n",
       "isd_og_mou_8           0.0021      0.005      0.442      0.658      -0.007       0.012\n",
       "spl_og_mou_6           0.0054      0.006      0.903      0.366      -0.006       0.017\n",
       "spl_og_mou_7           0.0084      0.005      1.807      0.071      -0.001       0.017\n",
       "spl_og_mou_8           0.0011      0.004      0.250      0.803      -0.007       0.010\n",
       "og_others_6           -0.0307      0.015     -2.113      0.035      -0.059      -0.002\n",
       "og_others_7            0.0074      0.015      0.486      0.627      -0.022       0.037\n",
       "og_others_8            0.0128      0.018      0.714      0.475      -0.022       0.048\n",
       "total_og_mou_6        -0.0029      0.002     -1.486      0.137      -0.007       0.001\n",
       "total_og_mou_7        -0.0058      0.002     -2.520      0.012      -0.010      -0.001\n",
       "total_og_mou_8        -0.0033      0.002     -2.035      0.042      -0.007      -0.000\n",
       "loc_ic_t2t_mou_6      -8.2573      5.217     -1.583      0.113     -18.482       1.968\n",
       "loc_ic_t2t_mou_7      21.6690      5.260      4.120      0.000      11.360      31.978\n",
       "loc_ic_t2t_mou_8      39.2584      5.708      6.878      0.000      28.071      50.445\n",
       "loc_ic_t2m_mou_6      -8.2570      5.217     -1.583      0.113     -18.482       1.968\n",
       "loc_ic_t2m_mou_7      21.6703      5.260      4.120      0.000      11.361      31.980\n",
       "loc_ic_t2m_mou_8      39.2566      5.708      6.878      0.000      28.070      50.444\n",
       "loc_ic_t2f_mou_6      -8.2579      5.217     -1.583      0.113     -18.483       1.967\n",
       "loc_ic_t2f_mou_7      21.6695      5.260      4.120      0.000      11.360      31.979\n",
       "loc_ic_t2f_mou_8      39.2570      5.708      6.878      0.000      28.069      50.445\n",
       "loc_ic_mou_6           8.2586      5.217      1.583      0.113      -1.966      18.483\n",
       "loc_ic_mou_7         -21.6694      5.260     -4.120      0.000     -31.978     -11.360\n",
       "loc_ic_mou_8         -39.2392      5.708     -6.875      0.000     -50.426     -28.052\n",
       "std_ic_t2t_mou_6     -14.2072      6.160     -2.306      0.021     -26.280      -2.134\n",
       "std_ic_t2t_mou_7      -4.6899      6.424     -0.730      0.465     -17.281       7.901\n",
       "std_ic_t2t_mou_8      11.3391      7.696      1.473      0.141      -3.745      26.424\n",
       "std_ic_t2m_mou_6     -14.2084      6.160     -2.307      0.021     -26.282      -2.135\n",
       "std_ic_t2m_mou_7      -4.6915      6.424     -0.730      0.465     -17.283       7.900\n",
       "std_ic_t2m_mou_8      11.3441      7.696      1.474      0.140      -3.740      26.428\n",
       "std_ic_t2f_mou_6     -14.2085      6.160     -2.306      0.021     -26.282      -2.135\n",
       "std_ic_t2f_mou_7      -4.6884      6.424     -0.730      0.466     -17.280       7.903\n",
       "std_ic_t2f_mou_8      11.3408      7.697      1.473      0.141      -3.744      26.426\n",
       "std_ic_mou_6          14.2099      6.160      2.307      0.021       2.137      26.283\n",
       "std_ic_mou_7           4.6899      6.424      0.730      0.465      -7.901      17.281\n",
       "std_ic_mou_8         -11.3220      7.696     -1.471      0.141     -26.406       3.762\n",
       "total_ic_mou_6        -0.0015      0.003     -0.577      0.564      -0.007       0.004\n",
       "total_ic_mou_7         0.0013      0.003      0.466      0.641      -0.004       0.007\n",
       "total_ic_mou_8        -0.0229      0.002     -9.761      0.000      -0.027      -0.018\n",
       "spl_ic_mou_6           0.0787      0.193      0.409      0.683      -0.299       0.456\n",
       "spl_ic_mou_7          -0.2649      0.342     -0.774      0.439      -0.936       0.406\n",
       "spl_ic_mou_8          -3.2593      0.502     -6.488      0.000      -4.244      -2.275\n",
       "isd_ic_mou_6           0.0021      0.003      0.787      0.431      -0.003       0.007\n",
       "isd_ic_mou_7          -0.0004      0.003     -0.142      0.887      -0.006       0.005\n",
       "isd_ic_mou_8           0.0215      0.003      8.290      0.000       0.016       0.027\n",
       "ic_others_6            0.0007      0.006      0.113      0.910      -0.011       0.013\n",
       "ic_others_7           -0.0069      0.008     -0.896      0.370      -0.022       0.008\n",
       "ic_others_8            0.0219      0.008      2.741      0.006       0.006       0.038\n",
       "total_rech_num_6      -0.0021      0.005     -0.454      0.650      -0.011       0.007\n",
       "total_rech_num_7       0.0146      0.006      2.281      0.023       0.002       0.027\n",
       "total_rech_num_8      -0.0449      0.008     -5.592      0.000      -0.061      -0.029\n",
       "total_rech_amt_6       0.0002      0.000      0.773      0.439      -0.000       0.001\n",
       "total_rech_amt_7      -0.0003      0.000     -1.208      0.227      -0.001       0.000\n",
       "total_rech_amt_8      -0.0006      0.000     -1.966      0.049      -0.001   -1.85e-06\n",
       "max_rech_amt_6        -0.0002      0.000     -0.614      0.539      -0.001       0.000\n",
       "max_rech_amt_7      7.392e-05      0.000      0.252      0.801      -0.000       0.001\n",
       "max_rech_amt_8         0.0007      0.000      1.867      0.062   -3.49e-05       0.001\n",
       "last_day_rch_amt_6    -0.0002      0.000     -0.524      0.601      -0.001       0.000\n",
       "last_day_rch_amt_7    -0.0004      0.000     -1.279      0.201      -0.001       0.000\n",
       "last_day_rch_amt_8    -0.0042      0.000     -8.526      0.000      -0.005      -0.003\n",
       "vol_2g_mb_6         7.671e-05      0.000      0.468      0.640      -0.000       0.000\n",
       "vol_2g_mb_7            0.0003      0.000      1.806      0.071   -2.82e-05       0.001\n",
       "vol_2g_mb_8           -0.0006      0.000     -2.230      0.026      -0.001   -6.89e-05\n",
       "vol_3g_mb_6           -0.0001    8.5e-05     -1.558      0.119      -0.000    3.41e-05\n",
       "vol_3g_mb_7         7.421e-05   9.98e-05      0.744      0.457      -0.000       0.000\n",
       "vol_3g_mb_8           -0.0002      0.000     -1.497      0.134      -0.000    5.97e-05\n",
       "monthly_2g_6          -0.2986      0.148     -2.013      0.044      -0.589      -0.008\n",
       "monthly_2g_7          -0.1963      0.150     -1.313      0.189      -0.489       0.097\n",
       "monthly_2g_8          -0.9160      0.225     -4.069      0.000      -1.357      -0.475\n",
       "sachet_2g_6            0.0755      0.028      2.688      0.007       0.020       0.131\n",
       "sachet_2g_7           -0.0126      0.029     -0.442      0.659      -0.069       0.043\n",
       "sachet_2g_8           -0.1185      0.036     -3.266      0.001      -0.190      -0.047\n",
       "monthly_3g_6           0.0007      0.107      0.006      0.995      -0.210       0.211\n",
       "monthly_3g_7          -0.0353      0.116     -0.305      0.760      -0.262       0.192\n",
       "monthly_3g_8          -0.7178      0.179     -4.014      0.000      -1.068      -0.367\n",
       "sachet_3g_6            0.0204      0.054      0.375      0.708      -0.086       0.127\n",
       "sachet_3g_7            0.1558      0.061      2.562      0.010       0.037       0.275\n",
       "sachet_3g_8           -0.1877      0.090     -2.085      0.037      -0.364      -0.011\n",
       "aon                   -0.0003   4.23e-05     -6.163      0.000      -0.000      -0.000\n",
       "aug_vbc_3g          4.788e-05      0.000      0.222      0.824      -0.000       0.000\n",
       "jul_vbc_3g            -0.0002      0.000     -0.993      0.321      -0.001       0.000\n",
       "jun_vbc_3g             0.0002      0.000      1.237      0.216      -0.000       0.001\n",
       "sep_vbc_3g            -0.0127      0.004     -3.137      0.002      -0.021      -0.005\n",
       "avg_rech_amt       -5.299e-05      0.000     -0.428      0.669      -0.000       0.000\n",
       "======================================================================================\n",
       "\"\"\""
      ]
     },
     "execution_count": 67,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Logistic regression model\n",
    "logm1 = sm.GLM(y_train,(sm.add_constant(X_train)), family = sm.families.Binomial())\n",
    "logm1.fit().summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Correlation Matrix**"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<href>https://stackoverflow.com/questions/17778394/list-highest-correlation-pairs-from-a-large-correlation-matrix-in-pandas</href>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Top Absolute Correlations\n",
      "arpu_8            total_rech_amt_8    0.953725\n",
      "isd_og_mou_7      isd_og_mou_8        0.941612\n",
      "arpu_6            total_rech_amt_6    0.941256\n",
      "arpu_7            total_rech_amt_7    0.937114\n",
      "isd_og_mou_6      isd_og_mou_7        0.917551\n",
      "                  isd_og_mou_8        0.917194\n",
      "loc_ic_mou_6      total_ic_mou_6      0.894930\n",
      "total_rech_amt_7  avg_rech_amt        0.892365\n",
      "arpu_7            avg_rech_amt        0.890903\n",
      "loc_ic_mou_8      total_ic_mou_8      0.887645\n",
      "total_rech_amt_6  avg_rech_amt        0.882449\n",
      "loc_ic_mou_7      total_ic_mou_7      0.881862\n",
      "arpu_6            avg_rech_amt        0.868980\n",
      "onnet_mou_8       std_og_t2t_mou_8    0.857278\n",
      "onnet_mou_7       std_og_t2t_mou_7    0.853077\n",
      "loc_ic_t2m_mou_8  loc_ic_mou_8        0.852665\n",
      "loc_ic_t2m_mou_6  loc_ic_mou_6        0.851521\n",
      "offnet_mou_7      std_og_t2m_mou_7    0.851181\n",
      "onnet_mou_6       std_og_t2t_mou_6    0.851023\n",
      "offnet_mou_8      std_og_t2m_mou_8    0.848822\n",
      "std_ic_t2m_mou_8  std_ic_mou_8        0.845906\n",
      "std_og_mou_8      total_og_mou_8      0.842622\n",
      "std_og_mou_7      total_og_mou_7      0.837532\n",
      "loc_ic_mou_7      loc_ic_mou_8        0.836470\n",
      "loc_ic_t2m_mou_7  loc_ic_mou_7        0.834070\n",
      "loc_og_mou_7      loc_og_mou_8        0.832403\n",
      "loc_og_t2t_mou_7  loc_og_t2t_mou_8    0.829139\n",
      "offnet_mou_6      std_og_t2m_mou_6    0.825206\n",
      "std_og_mou_6      total_og_mou_6      0.819237\n",
      "std_ic_t2m_mou_6  std_ic_mou_6        0.814796\n",
      "loc_ic_t2m_mou_7  loc_ic_t2m_mou_8    0.814581\n",
      "loc_og_t2m_mou_7  loc_og_t2m_mou_8    0.814198\n",
      "loc_ic_mou_6      loc_ic_mou_7        0.812772\n",
      "std_ic_t2m_mou_7  std_ic_mou_7        0.812732\n",
      "total_ic_mou_7    total_ic_mou_8      0.812222\n",
      "loc_og_t2f_mou_6  loc_og_t2f_mou_7    0.805848\n",
      "loc_ic_t2t_mou_7  loc_ic_t2t_mou_8    0.796115\n",
      "std_og_t2t_mou_7  std_og_t2t_mou_8    0.795879\n",
      "onnet_mou_7       onnet_mou_8         0.794406\n",
      "loc_og_mou_6      loc_og_mou_7        0.792674\n",
      "dtype: float64\n"
     ]
    }
   ],
   "source": [
    "def get_redundant_pairs(df):\n",
    "    '''Get diagonal and lower triangular pairs of correlation matrix'''\n",
    "    pairs_to_drop = set()\n",
    "    cols = df.columns\n",
    "    for i in range(0, df.shape[1]):\n",
    "        for j in range(0, i+1):\n",
    "            pairs_to_drop.add((cols[i], cols[j]))\n",
    "    return pairs_to_drop\n",
    "\n",
    "def get_top_abs_correlations(df, n=5):\n",
    "    au_corr = df.corr().abs().unstack()\n",
    "    labels_to_drop = get_redundant_pairs(df)\n",
    "    au_corr = au_corr.drop(labels=labels_to_drop).sort_values(ascending=False)\n",
    "    return au_corr[0:n]\n",
    "\n",
    "print(\"Top Absolute Correlations\")\n",
    "print(get_top_abs_correlations(df_high_val_imputed, 40))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {},
   "outputs": [],
   "source": [
    "top_corr = get_top_abs_correlations(df_high_val_imputed, 40)\n",
    "top_cor_list = top_corr[top_corr>0.80].index.droplevel().tolist()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Dropping highly correlated variables**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_test2 = X_test.drop(top_cor_list,1)\n",
    "X_train2 = X_train.drop(top_cor_list,1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Checking the Correlation Matrix**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Top Absolute Correlations\n",
      "onnet_mou_7       onnet_mou_8         0.796592\n",
      "loc_og_t2m_mou_6  loc_og_mou_6        0.792065\n",
      "loc_ic_t2f_mou_7  loc_ic_t2f_mou_8    0.783386\n",
      "loc_ic_t2t_mou_7  loc_ic_t2t_mou_8    0.783045\n",
      "loc_ic_t2m_mou_6  loc_ic_t2m_mou_7    0.782690\n",
      "std_og_mou_7      std_og_mou_8        0.782015\n",
      "loc_og_mou_6      loc_og_mou_7        0.778974\n",
      "std_og_t2f_mou_7  std_og_t2f_mou_8    0.774053\n",
      "loc_og_t2t_mou_6  loc_og_t2t_mou_7    0.771000\n",
      "loc_og_t2t_mou_7  loc_og_mou_7        0.769586\n",
      "loc_ic_t2t_mou_6  loc_ic_t2t_mou_7    0.766510\n",
      "loc_og_t2m_mou_6  loc_og_t2m_mou_7    0.758898\n",
      "std_ic_t2t_mou_6  std_ic_t2t_mou_7    0.756551\n",
      "isd_ic_mou_6      isd_ic_mou_7        0.753450\n",
      "loc_og_t2t_mou_6  loc_og_mou_6        0.752914\n",
      "total_rech_num_7  total_rech_num_8    0.752340\n",
      "loc_og_t2m_mou_7  loc_og_mou_7        0.747564\n",
      "vol_3g_mb_7       vol_3g_mb_8         0.747065\n",
      "loc_ic_t2f_mou_6  loc_ic_t2f_mou_7    0.744602\n",
      "ic_others_7       ic_others_8         0.741534\n",
      "dtype: float64\n"
     ]
    }
   ],
   "source": [
    "print(\"Top Absolute Correlations\")\n",
    "print(get_top_abs_correlations(X_train2, 20))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Re-Running the Model**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table class=\"simpletable\">\n",
       "<caption>Generalized Linear Model Regression Results</caption>\n",
       "<tr>\n",
       "  <th>Dep. Variable:</th>         <td>churn</td>      <th>  No. Observations:  </th>  <td> 21007</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Model:</th>                  <td>GLM</td>       <th>  Df Residuals:      </th>  <td> 20909</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Model Family:</th>        <td>Binomial</td>     <th>  Df Model:          </th>  <td>    97</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Link Function:</th>         <td>logit</td>      <th>  Scale:             </th> <td>  1.0000</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Method:</th>                <td>IRLS</td>       <th>  Log-Likelihood:    </th> <td> -4322.5</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Date:</th>            <td>Wed, 17 Feb 2021</td> <th>  Deviance:          </th> <td>  8645.1</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Time:</th>                <td>18:34:29</td>     <th>  Pearson chi2:      </th> <td>5.05e+06</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>No. Iterations:</th>          <td>9</td>        <th>                     </th>     <td> </td>   \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Covariance Type:</th>     <td>nonrobust</td>    <th>                     </th>     <td> </td>   \n",
       "</tr>\n",
       "</table>\n",
       "<table class=\"simpletable\">\n",
       "<tr>\n",
       "           <td></td>             <th>coef</th>     <th>std err</th>      <th>z</th>      <th>P>|z|</th>  <th>[0.025</th>    <th>0.975]</th>  \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>const</th>              <td>   -0.6826</td> <td>    0.096</td> <td>   -7.088</td> <td> 0.000</td> <td>   -0.871</td> <td>   -0.494</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>arpu_6</th>             <td>    0.0005</td> <td>    0.000</td> <td>    3.080</td> <td> 0.002</td> <td>    0.000</td> <td>    0.001</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>arpu_7</th>             <td>    0.0005</td> <td>    0.000</td> <td>    2.891</td> <td> 0.004</td> <td>    0.000</td> <td>    0.001</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>arpu_8</th>             <td>   -0.0007</td> <td>    0.000</td> <td>   -2.965</td> <td> 0.003</td> <td>   -0.001</td> <td>   -0.000</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>onnet_mou_6</th>        <td>   -0.0064</td> <td>    0.005</td> <td>   -1.192</td> <td> 0.233</td> <td>   -0.017</td> <td>    0.004</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>onnet_mou_7</th>        <td>    0.0017</td> <td>    0.002</td> <td>    0.806</td> <td> 0.420</td> <td>   -0.002</td> <td>    0.006</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>onnet_mou_8</th>        <td>   -0.0027</td> <td>    0.000</td> <td>   -6.739</td> <td> 0.000</td> <td>   -0.003</td> <td>   -0.002</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>offnet_mou_6</th>       <td>   -0.0063</td> <td>    0.005</td> <td>   -1.185</td> <td> 0.236</td> <td>   -0.017</td> <td>    0.004</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>offnet_mou_7</th>       <td>    0.0020</td> <td>    0.002</td> <td>    0.948</td> <td> 0.343</td> <td>   -0.002</td> <td>    0.006</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>offnet_mou_8</th>       <td>   -0.0033</td> <td>    0.000</td> <td>   -8.767</td> <td> 0.000</td> <td>   -0.004</td> <td>   -0.003</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>roam_ic_mou_6</th>      <td>    0.0005</td> <td>    0.000</td> <td>    1.164</td> <td> 0.244</td> <td>   -0.000</td> <td>    0.001</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>roam_ic_mou_7</th>      <td>   -0.0001</td> <td>    0.000</td> <td>   -0.234</td> <td> 0.815</td> <td>   -0.001</td> <td>    0.001</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>roam_ic_mou_8</th>      <td>    0.0006</td> <td>    0.000</td> <td>    1.360</td> <td> 0.174</td> <td>   -0.000</td> <td>    0.001</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>roam_og_mou_6</th>      <td>    0.0066</td> <td>    0.005</td> <td>    1.223</td> <td> 0.221</td> <td>   -0.004</td> <td>    0.017</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>roam_og_mou_7</th>      <td>   -0.0011</td> <td>    0.002</td> <td>   -0.534</td> <td> 0.593</td> <td>   -0.005</td> <td>    0.003</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>roam_og_mou_8</th>      <td>    0.0039</td> <td>    0.000</td> <td>    8.591</td> <td> 0.000</td> <td>    0.003</td> <td>    0.005</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>loc_og_t2t_mou_6</th>   <td>   10.9494</td> <td>    4.898</td> <td>    2.236</td> <td> 0.025</td> <td>    1.350</td> <td>   20.549</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>loc_og_t2t_mou_7</th>   <td>    0.0043</td> <td>    0.004</td> <td>    1.025</td> <td> 0.305</td> <td>   -0.004</td> <td>    0.013</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>loc_og_t2m_mou_6</th>   <td>   10.9486</td> <td>    4.898</td> <td>    2.235</td> <td> 0.025</td> <td>    1.349</td> <td>   20.548</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>loc_og_t2m_mou_7</th>   <td>    0.0044</td> <td>    0.004</td> <td>    1.028</td> <td> 0.304</td> <td>   -0.004</td> <td>    0.013</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>loc_og_t2f_mou_6</th>   <td>   10.9517</td> <td>    4.898</td> <td>    2.236</td> <td> 0.025</td> <td>    1.352</td> <td>   20.552</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>loc_og_t2f_mou_8</th>   <td>    0.0060</td> <td>    0.004</td> <td>    1.558</td> <td> 0.119</td> <td>   -0.002</td> <td>    0.014</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>loc_og_t2c_mou_6</th>   <td>    0.0006</td> <td>    0.004</td> <td>    0.129</td> <td> 0.897</td> <td>   -0.008</td> <td>    0.009</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>loc_og_t2c_mou_7</th>   <td>    0.0049</td> <td>    0.005</td> <td>    1.010</td> <td> 0.312</td> <td>   -0.005</td> <td>    0.015</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>loc_og_t2c_mou_8</th>   <td>    0.0057</td> <td>    0.007</td> <td>    0.864</td> <td> 0.387</td> <td>   -0.007</td> <td>    0.019</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>loc_og_mou_6</th>       <td>  -10.9424</td> <td>    4.898</td> <td>   -2.234</td> <td> 0.025</td> <td>  -20.542</td> <td>   -1.343</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>loc_og_mou_7</th>       <td>   -0.0061</td> <td>    0.005</td> <td>   -1.283</td> <td> 0.199</td> <td>   -0.015</td> <td>    0.003</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>std_og_t2f_mou_6</th>   <td>    0.0088</td> <td>    0.004</td> <td>    2.145</td> <td> 0.032</td> <td>    0.001</td> <td>    0.017</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>std_og_t2f_mou_7</th>   <td>   -0.0093</td> <td>    0.005</td> <td>   -1.717</td> <td> 0.086</td> <td>   -0.020</td> <td>    0.001</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>std_og_t2f_mou_8</th>   <td>   -0.0058</td> <td>    0.007</td> <td>   -0.785</td> <td> 0.433</td> <td>   -0.020</td> <td>    0.009</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>std_og_mou_6</th>       <td>    0.0065</td> <td>    0.005</td> <td>    1.222</td> <td> 0.222</td> <td>   -0.004</td> <td>    0.017</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>std_og_mou_7</th>       <td>   -0.0018</td> <td>    0.002</td> <td>   -0.850</td> <td> 0.396</td> <td>   -0.006</td> <td>    0.002</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>std_og_mou_8</th>       <td>    0.0023</td> <td>    0.000</td> <td>    6.206</td> <td> 0.000</td> <td>    0.002</td> <td>    0.003</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>isd_og_mou_6</th>       <td>    0.0078</td> <td>    0.005</td> <td>    1.456</td> <td> 0.145</td> <td>   -0.003</td> <td>    0.018</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>spl_og_mou_6</th>       <td>    0.0053</td> <td>    0.005</td> <td>    1.004</td> <td> 0.315</td> <td>   -0.005</td> <td>    0.016</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>spl_og_mou_7</th>       <td>   -0.0013</td> <td>    0.003</td> <td>   -0.442</td> <td> 0.659</td> <td>   -0.007</td> <td>    0.004</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>spl_og_mou_8</th>       <td>   -0.0013</td> <td>    0.003</td> <td>   -0.379</td> <td> 0.705</td> <td>   -0.008</td> <td>    0.005</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>og_others_6</th>        <td>   -0.0291</td> <td>    0.013</td> <td>   -2.327</td> <td> 0.020</td> <td>   -0.054</td> <td>   -0.005</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>og_others_7</th>        <td>    0.0005</td> <td>    0.013</td> <td>    0.037</td> <td> 0.971</td> <td>   -0.025</td> <td>    0.026</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>og_others_8</th>        <td>    0.0149</td> <td>    0.018</td> <td>    0.831</td> <td> 0.406</td> <td>   -0.020</td> <td>    0.050</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>loc_ic_t2t_mou_6</th>   <td>    0.0002</td> <td>    0.000</td> <td>    0.636</td> <td> 0.525</td> <td>   -0.000</td> <td>    0.001</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>loc_ic_t2t_mou_7</th>   <td>    0.0010</td> <td>    0.000</td> <td>    2.645</td> <td> 0.008</td> <td>    0.000</td> <td>    0.002</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>loc_ic_t2t_mou_8</th>   <td>   -0.0056</td> <td>    0.001</td> <td>   -5.801</td> <td> 0.000</td> <td>   -0.008</td> <td>   -0.004</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>loc_ic_t2m_mou_6</th>   <td>    0.0006</td> <td>    0.000</td> <td>    2.853</td> <td> 0.004</td> <td>    0.000</td> <td>    0.001</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>loc_ic_t2m_mou_7</th>   <td>   -0.0006</td> <td>    0.000</td> <td>   -2.027</td> <td> 0.043</td> <td>   -0.001</td> <td>-1.86e-05</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>loc_ic_t2f_mou_6</th>   <td>    0.0007</td> <td>    0.001</td> <td>    0.516</td> <td> 0.606</td> <td>   -0.002</td> <td>    0.003</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>loc_ic_t2f_mou_7</th>   <td>    0.0012</td> <td>    0.001</td> <td>    0.953</td> <td> 0.341</td> <td>   -0.001</td> <td>    0.004</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>loc_ic_t2f_mou_8</th>   <td>   -0.0113</td> <td>    0.003</td> <td>   -4.047</td> <td> 0.000</td> <td>   -0.017</td> <td>   -0.006</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>std_ic_t2t_mou_6</th>   <td>    0.0016</td> <td>    0.001</td> <td>    3.050</td> <td> 0.002</td> <td>    0.001</td> <td>    0.003</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>std_ic_t2t_mou_7</th>   <td>    0.0012</td> <td>    0.001</td> <td>    2.093</td> <td> 0.036</td> <td> 7.94e-05</td> <td>    0.002</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>std_ic_t2t_mou_8</th>   <td>   -0.0067</td> <td>    0.002</td> <td>   -4.193</td> <td> 0.000</td> <td>   -0.010</td> <td>   -0.004</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>std_ic_t2m_mou_6</th>   <td>    0.0003</td> <td>    0.000</td> <td>    0.721</td> <td> 0.471</td> <td>   -0.000</td> <td>    0.001</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>std_ic_t2m_mou_7</th>   <td>   -0.0002</td> <td>    0.000</td> <td>   -0.450</td> <td> 0.653</td> <td>   -0.001</td> <td>    0.001</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>std_ic_t2m_mou_8</th>   <td>   -0.0014</td> <td>    0.001</td> <td>   -2.145</td> <td> 0.032</td> <td>   -0.003</td> <td>   -0.000</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>std_ic_t2f_mou_6</th>   <td>    0.0009</td> <td>    0.003</td> <td>    0.346</td> <td> 0.729</td> <td>   -0.004</td> <td>    0.006</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>std_ic_t2f_mou_7</th>   <td>    0.0013</td> <td>    0.003</td> <td>    0.384</td> <td> 0.701</td> <td>   -0.005</td> <td>    0.008</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>std_ic_t2f_mou_8</th>   <td>   -0.0084</td> <td>    0.006</td> <td>   -1.357</td> <td> 0.175</td> <td>   -0.020</td> <td>    0.004</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>spl_ic_mou_6</th>       <td>   -0.0622</td> <td>    0.188</td> <td>   -0.331</td> <td> 0.741</td> <td>   -0.431</td> <td>    0.306</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>spl_ic_mou_7</th>       <td>   -0.2253</td> <td>    0.325</td> <td>   -0.692</td> <td> 0.489</td> <td>   -0.863</td> <td>    0.412</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>spl_ic_mou_8</th>       <td>   -3.1389</td> <td>    0.468</td> <td>   -6.711</td> <td> 0.000</td> <td>   -4.056</td> <td>   -2.222</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>isd_ic_mou_6</th>       <td>    0.0009</td> <td>    0.001</td> <td>    1.234</td> <td> 0.217</td> <td>   -0.000</td> <td>    0.002</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>isd_ic_mou_7</th>       <td>    0.0009</td> <td>    0.001</td> <td>    1.160</td> <td> 0.246</td> <td>   -0.001</td> <td>    0.002</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>isd_ic_mou_8</th>       <td>   -0.0019</td> <td>    0.001</td> <td>   -1.778</td> <td> 0.075</td> <td>   -0.004</td> <td>    0.000</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>ic_others_6</th>        <td>    0.0042</td> <td>    0.004</td> <td>    1.077</td> <td> 0.281</td> <td>   -0.003</td> <td>    0.012</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>ic_others_7</th>        <td>   -0.0087</td> <td>    0.007</td> <td>   -1.304</td> <td> 0.192</td> <td>   -0.022</td> <td>    0.004</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>ic_others_8</th>        <td>   -0.0045</td> <td>    0.008</td> <td>   -0.566</td> <td> 0.572</td> <td>   -0.020</td> <td>    0.011</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>total_rech_num_6</th>   <td>    0.0007</td> <td>    0.004</td> <td>    0.171</td> <td> 0.864</td> <td>   -0.007</td> <td>    0.009</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>total_rech_num_7</th>   <td>    0.0203</td> <td>    0.006</td> <td>    3.487</td> <td> 0.000</td> <td>    0.009</td> <td>    0.032</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>total_rech_num_8</th>   <td>   -0.0718</td> <td>    0.008</td> <td>   -8.962</td> <td> 0.000</td> <td>   -0.087</td> <td>   -0.056</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>max_rech_amt_6</th>     <td>    0.0003</td> <td>    0.000</td> <td>    1.392</td> <td> 0.164</td> <td>   -0.000</td> <td>    0.001</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>max_rech_amt_7</th>     <td>    0.0005</td> <td>    0.000</td> <td>    2.532</td> <td> 0.011</td> <td>    0.000</td> <td>    0.001</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>max_rech_amt_8</th>     <td>   -0.0004</td> <td>    0.000</td> <td>   -1.151</td> <td> 0.250</td> <td>   -0.001</td> <td>    0.000</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>last_day_rch_amt_6</th> <td> 7.074e-05</td> <td>    0.000</td> <td>    0.245</td> <td> 0.807</td> <td>   -0.000</td> <td>    0.001</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>last_day_rch_amt_7</th> <td>   -0.0004</td> <td>    0.000</td> <td>   -1.475</td> <td> 0.140</td> <td>   -0.001</td> <td>    0.000</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>last_day_rch_amt_8</th> <td>   -0.0049</td> <td>    0.001</td> <td>   -9.818</td> <td> 0.000</td> <td>   -0.006</td> <td>   -0.004</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>vol_2g_mb_6</th>        <td>    0.0001</td> <td>    0.000</td> <td>    0.780</td> <td> 0.435</td> <td>   -0.000</td> <td>    0.000</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>vol_2g_mb_7</th>        <td>    0.0003</td> <td>    0.000</td> <td>    1.465</td> <td> 0.143</td> <td>-8.77e-05</td> <td>    0.001</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>vol_2g_mb_8</th>        <td>   -0.0004</td> <td>    0.000</td> <td>   -1.531</td> <td> 0.126</td> <td>   -0.001</td> <td>    0.000</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>vol_3g_mb_6</th>        <td> 4.824e-06</td> <td> 6.94e-05</td> <td>    0.070</td> <td> 0.945</td> <td>   -0.000</td> <td>    0.000</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>vol_3g_mb_7</th>        <td> 5.012e-05</td> <td> 8.29e-05</td> <td>    0.604</td> <td> 0.546</td> <td>   -0.000</td> <td>    0.000</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>vol_3g_mb_8</th>        <td> 8.511e-06</td> <td> 9.74e-05</td> <td>    0.087</td> <td> 0.930</td> <td>   -0.000</td> <td>    0.000</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>monthly_2g_6</th>       <td>   -0.3566</td> <td>    0.138</td> <td>   -2.589</td> <td> 0.010</td> <td>   -0.627</td> <td>   -0.087</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>monthly_2g_7</th>       <td>   -0.2771</td> <td>    0.141</td> <td>   -1.970</td> <td> 0.049</td> <td>   -0.553</td> <td>   -0.001</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>monthly_2g_8</th>       <td>   -1.1104</td> <td>    0.222</td> <td>   -4.992</td> <td> 0.000</td> <td>   -1.546</td> <td>   -0.674</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>sachet_2g_6</th>        <td>    0.0693</td> <td>    0.027</td> <td>    2.557</td> <td> 0.011</td> <td>    0.016</td> <td>    0.122</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>sachet_2g_7</th>        <td>   -0.0097</td> <td>    0.027</td> <td>   -0.354</td> <td> 0.723</td> <td>   -0.063</td> <td>    0.044</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>sachet_2g_8</th>        <td>   -0.1508</td> <td>    0.037</td> <td>   -4.126</td> <td> 0.000</td> <td>   -0.222</td> <td>   -0.079</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>monthly_3g_6</th>       <td>    0.0344</td> <td>    0.095</td> <td>    0.362</td> <td> 0.717</td> <td>   -0.152</td> <td>    0.221</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>monthly_3g_7</th>       <td>   -0.0852</td> <td>    0.103</td> <td>   -0.826</td> <td> 0.409</td> <td>   -0.287</td> <td>    0.117</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>monthly_3g_8</th>       <td>   -0.6010</td> <td>    0.160</td> <td>   -3.761</td> <td> 0.000</td> <td>   -0.914</td> <td>   -0.288</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>sachet_3g_6</th>        <td>    0.0062</td> <td>    0.052</td> <td>    0.120</td> <td> 0.904</td> <td>   -0.095</td> <td>    0.108</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>sachet_3g_7</th>        <td>    0.1379</td> <td>    0.055</td> <td>    2.497</td> <td> 0.013</td> <td>    0.030</td> <td>    0.246</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>sachet_3g_8</th>        <td>   -0.2139</td> <td>    0.088</td> <td>   -2.444</td> <td> 0.015</td> <td>   -0.385</td> <td>   -0.042</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>aon</th>                <td>   -0.0003</td> <td>  3.9e-05</td> <td>   -7.831</td> <td> 0.000</td> <td>   -0.000</td> <td>   -0.000</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>aug_vbc_3g</th>         <td>   -0.0002</td> <td>    0.000</td> <td>   -1.059</td> <td> 0.290</td> <td>   -0.001</td> <td>    0.000</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>jul_vbc_3g</th>         <td>   -0.0001</td> <td>    0.000</td> <td>   -0.886</td> <td> 0.376</td> <td>   -0.000</td> <td>    0.000</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>jun_vbc_3g</th>         <td>  7.88e-05</td> <td>    0.000</td> <td>    0.548</td> <td> 0.584</td> <td>   -0.000</td> <td>    0.000</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>sep_vbc_3g</th>         <td>   -0.0138</td> <td>    0.004</td> <td>   -3.283</td> <td> 0.001</td> <td>   -0.022</td> <td>   -0.006</td>\n",
       "</tr>\n",
       "</table>"
      ],
      "text/plain": [
       "<class 'statsmodels.iolib.summary.Summary'>\n",
       "\"\"\"\n",
       "                 Generalized Linear Model Regression Results                  \n",
       "==============================================================================\n",
       "Dep. Variable:                  churn   No. Observations:                21007\n",
       "Model:                            GLM   Df Residuals:                    20909\n",
       "Model Family:                Binomial   Df Model:                           97\n",
       "Link Function:                  logit   Scale:                          1.0000\n",
       "Method:                          IRLS   Log-Likelihood:                -4322.5\n",
       "Date:                Wed, 17 Feb 2021   Deviance:                       8645.1\n",
       "Time:                        18:34:29   Pearson chi2:                 5.05e+06\n",
       "No. Iterations:                     9                                         \n",
       "Covariance Type:            nonrobust                                         \n",
       "======================================================================================\n",
       "                         coef    std err          z      P>|z|      [0.025      0.975]\n",
       "--------------------------------------------------------------------------------------\n",
       "const                 -0.6826      0.096     -7.088      0.000      -0.871      -0.494\n",
       "arpu_6                 0.0005      0.000      3.080      0.002       0.000       0.001\n",
       "arpu_7                 0.0005      0.000      2.891      0.004       0.000       0.001\n",
       "arpu_8                -0.0007      0.000     -2.965      0.003      -0.001      -0.000\n",
       "onnet_mou_6           -0.0064      0.005     -1.192      0.233      -0.017       0.004\n",
       "onnet_mou_7            0.0017      0.002      0.806      0.420      -0.002       0.006\n",
       "onnet_mou_8           -0.0027      0.000     -6.739      0.000      -0.003      -0.002\n",
       "offnet_mou_6          -0.0063      0.005     -1.185      0.236      -0.017       0.004\n",
       "offnet_mou_7           0.0020      0.002      0.948      0.343      -0.002       0.006\n",
       "offnet_mou_8          -0.0033      0.000     -8.767      0.000      -0.004      -0.003\n",
       "roam_ic_mou_6          0.0005      0.000      1.164      0.244      -0.000       0.001\n",
       "roam_ic_mou_7         -0.0001      0.000     -0.234      0.815      -0.001       0.001\n",
       "roam_ic_mou_8          0.0006      0.000      1.360      0.174      -0.000       0.001\n",
       "roam_og_mou_6          0.0066      0.005      1.223      0.221      -0.004       0.017\n",
       "roam_og_mou_7         -0.0011      0.002     -0.534      0.593      -0.005       0.003\n",
       "roam_og_mou_8          0.0039      0.000      8.591      0.000       0.003       0.005\n",
       "loc_og_t2t_mou_6      10.9494      4.898      2.236      0.025       1.350      20.549\n",
       "loc_og_t2t_mou_7       0.0043      0.004      1.025      0.305      -0.004       0.013\n",
       "loc_og_t2m_mou_6      10.9486      4.898      2.235      0.025       1.349      20.548\n",
       "loc_og_t2m_mou_7       0.0044      0.004      1.028      0.304      -0.004       0.013\n",
       "loc_og_t2f_mou_6      10.9517      4.898      2.236      0.025       1.352      20.552\n",
       "loc_og_t2f_mou_8       0.0060      0.004      1.558      0.119      -0.002       0.014\n",
       "loc_og_t2c_mou_6       0.0006      0.004      0.129      0.897      -0.008       0.009\n",
       "loc_og_t2c_mou_7       0.0049      0.005      1.010      0.312      -0.005       0.015\n",
       "loc_og_t2c_mou_8       0.0057      0.007      0.864      0.387      -0.007       0.019\n",
       "loc_og_mou_6         -10.9424      4.898     -2.234      0.025     -20.542      -1.343\n",
       "loc_og_mou_7          -0.0061      0.005     -1.283      0.199      -0.015       0.003\n",
       "std_og_t2f_mou_6       0.0088      0.004      2.145      0.032       0.001       0.017\n",
       "std_og_t2f_mou_7      -0.0093      0.005     -1.717      0.086      -0.020       0.001\n",
       "std_og_t2f_mou_8      -0.0058      0.007     -0.785      0.433      -0.020       0.009\n",
       "std_og_mou_6           0.0065      0.005      1.222      0.222      -0.004       0.017\n",
       "std_og_mou_7          -0.0018      0.002     -0.850      0.396      -0.006       0.002\n",
       "std_og_mou_8           0.0023      0.000      6.206      0.000       0.002       0.003\n",
       "isd_og_mou_6           0.0078      0.005      1.456      0.145      -0.003       0.018\n",
       "spl_og_mou_6           0.0053      0.005      1.004      0.315      -0.005       0.016\n",
       "spl_og_mou_7          -0.0013      0.003     -0.442      0.659      -0.007       0.004\n",
       "spl_og_mou_8          -0.0013      0.003     -0.379      0.705      -0.008       0.005\n",
       "og_others_6           -0.0291      0.013     -2.327      0.020      -0.054      -0.005\n",
       "og_others_7            0.0005      0.013      0.037      0.971      -0.025       0.026\n",
       "og_others_8            0.0149      0.018      0.831      0.406      -0.020       0.050\n",
       "loc_ic_t2t_mou_6       0.0002      0.000      0.636      0.525      -0.000       0.001\n",
       "loc_ic_t2t_mou_7       0.0010      0.000      2.645      0.008       0.000       0.002\n",
       "loc_ic_t2t_mou_8      -0.0056      0.001     -5.801      0.000      -0.008      -0.004\n",
       "loc_ic_t2m_mou_6       0.0006      0.000      2.853      0.004       0.000       0.001\n",
       "loc_ic_t2m_mou_7      -0.0006      0.000     -2.027      0.043      -0.001   -1.86e-05\n",
       "loc_ic_t2f_mou_6       0.0007      0.001      0.516      0.606      -0.002       0.003\n",
       "loc_ic_t2f_mou_7       0.0012      0.001      0.953      0.341      -0.001       0.004\n",
       "loc_ic_t2f_mou_8      -0.0113      0.003     -4.047      0.000      -0.017      -0.006\n",
       "std_ic_t2t_mou_6       0.0016      0.001      3.050      0.002       0.001       0.003\n",
       "std_ic_t2t_mou_7       0.0012      0.001      2.093      0.036    7.94e-05       0.002\n",
       "std_ic_t2t_mou_8      -0.0067      0.002     -4.193      0.000      -0.010      -0.004\n",
       "std_ic_t2m_mou_6       0.0003      0.000      0.721      0.471      -0.000       0.001\n",
       "std_ic_t2m_mou_7      -0.0002      0.000     -0.450      0.653      -0.001       0.001\n",
       "std_ic_t2m_mou_8      -0.0014      0.001     -2.145      0.032      -0.003      -0.000\n",
       "std_ic_t2f_mou_6       0.0009      0.003      0.346      0.729      -0.004       0.006\n",
       "std_ic_t2f_mou_7       0.0013      0.003      0.384      0.701      -0.005       0.008\n",
       "std_ic_t2f_mou_8      -0.0084      0.006     -1.357      0.175      -0.020       0.004\n",
       "spl_ic_mou_6          -0.0622      0.188     -0.331      0.741      -0.431       0.306\n",
       "spl_ic_mou_7          -0.2253      0.325     -0.692      0.489      -0.863       0.412\n",
       "spl_ic_mou_8          -3.1389      0.468     -6.711      0.000      -4.056      -2.222\n",
       "isd_ic_mou_6           0.0009      0.001      1.234      0.217      -0.000       0.002\n",
       "isd_ic_mou_7           0.0009      0.001      1.160      0.246      -0.001       0.002\n",
       "isd_ic_mou_8          -0.0019      0.001     -1.778      0.075      -0.004       0.000\n",
       "ic_others_6            0.0042      0.004      1.077      0.281      -0.003       0.012\n",
       "ic_others_7           -0.0087      0.007     -1.304      0.192      -0.022       0.004\n",
       "ic_others_8           -0.0045      0.008     -0.566      0.572      -0.020       0.011\n",
       "total_rech_num_6       0.0007      0.004      0.171      0.864      -0.007       0.009\n",
       "total_rech_num_7       0.0203      0.006      3.487      0.000       0.009       0.032\n",
       "total_rech_num_8      -0.0718      0.008     -8.962      0.000      -0.087      -0.056\n",
       "max_rech_amt_6         0.0003      0.000      1.392      0.164      -0.000       0.001\n",
       "max_rech_amt_7         0.0005      0.000      2.532      0.011       0.000       0.001\n",
       "max_rech_amt_8        -0.0004      0.000     -1.151      0.250      -0.001       0.000\n",
       "last_day_rch_amt_6  7.074e-05      0.000      0.245      0.807      -0.000       0.001\n",
       "last_day_rch_amt_7    -0.0004      0.000     -1.475      0.140      -0.001       0.000\n",
       "last_day_rch_amt_8    -0.0049      0.001     -9.818      0.000      -0.006      -0.004\n",
       "vol_2g_mb_6            0.0001      0.000      0.780      0.435      -0.000       0.000\n",
       "vol_2g_mb_7            0.0003      0.000      1.465      0.143   -8.77e-05       0.001\n",
       "vol_2g_mb_8           -0.0004      0.000     -1.531      0.126      -0.001       0.000\n",
       "vol_3g_mb_6         4.824e-06   6.94e-05      0.070      0.945      -0.000       0.000\n",
       "vol_3g_mb_7         5.012e-05   8.29e-05      0.604      0.546      -0.000       0.000\n",
       "vol_3g_mb_8         8.511e-06   9.74e-05      0.087      0.930      -0.000       0.000\n",
       "monthly_2g_6          -0.3566      0.138     -2.589      0.010      -0.627      -0.087\n",
       "monthly_2g_7          -0.2771      0.141     -1.970      0.049      -0.553      -0.001\n",
       "monthly_2g_8          -1.1104      0.222     -4.992      0.000      -1.546      -0.674\n",
       "sachet_2g_6            0.0693      0.027      2.557      0.011       0.016       0.122\n",
       "sachet_2g_7           -0.0097      0.027     -0.354      0.723      -0.063       0.044\n",
       "sachet_2g_8           -0.1508      0.037     -4.126      0.000      -0.222      -0.079\n",
       "monthly_3g_6           0.0344      0.095      0.362      0.717      -0.152       0.221\n",
       "monthly_3g_7          -0.0852      0.103     -0.826      0.409      -0.287       0.117\n",
       "monthly_3g_8          -0.6010      0.160     -3.761      0.000      -0.914      -0.288\n",
       "sachet_3g_6            0.0062      0.052      0.120      0.904      -0.095       0.108\n",
       "sachet_3g_7            0.1379      0.055      2.497      0.013       0.030       0.246\n",
       "sachet_3g_8           -0.2139      0.088     -2.444      0.015      -0.385      -0.042\n",
       "aon                   -0.0003    3.9e-05     -7.831      0.000      -0.000      -0.000\n",
       "aug_vbc_3g            -0.0002      0.000     -1.059      0.290      -0.001       0.000\n",
       "jul_vbc_3g            -0.0001      0.000     -0.886      0.376      -0.000       0.000\n",
       "jun_vbc_3g           7.88e-05      0.000      0.548      0.584      -0.000       0.000\n",
       "sep_vbc_3g            -0.0138      0.004     -3.283      0.001      -0.022      -0.006\n",
       "======================================================================================\n",
       "\"\"\""
      ]
     },
     "execution_count": 72,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "logm2 = sm.GLM(y_train,(sm.add_constant(X_train2)), family = sm.families.Binomial())\n",
    "logm2.fit().summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Using RFE"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[False False False False False  True  True False  True False False False\n",
      " False False False False False False  True False False False False False\n",
      " False False False  True False False  True False False False False False\n",
      " False False False False False  True False False False False False False\n",
      " False False False  True False  True False False  True False False  True\n",
      " False False False False  True  True False False False False False False\n",
      " False False False False False  True False False  True False False False\n",
      " False False  True False False False False False  True False False False\n",
      " False False False False False  True False False False False False False\n",
      " False False  True False False False False False False False False False\n",
      " False False False False  True False]\n",
      "[ 18  13  28  74  34   1   1  93   1  54  84 102  75  30  90  58  60  87\n",
      "   1  59  20 105  89  88  97  99  77   1  24  43   1  42  17  35  33   5\n",
      "  83  92  82  14  22   1  53  29  86 100  44  78  91  96  61   1  23   1\n",
      "  40  70   1  39  16   1  48  71  12   8   1   1  37  25   4  36 107   3\n",
      "  38  52  95   6  62   1   7  31   1 106  73  10   9  32   1  55  41  19\n",
      "  94  27   1  57  26  15 104  51  72  76  67   1  85  47  45  66  65  49\n",
      "  68  46   1  63  98  21 103  64   2  79  50  69  11  81 101  80   1  56]\n"
     ]
    }
   ],
   "source": [
    "from sklearn.linear_model import LogisticRegression\n",
    "logreg = LogisticRegression(max_iter=1000)\n",
    "from sklearn.feature_selection import RFE\n",
    "rfe = RFE(logreg, n_features_to_select=20)    # running RFE with 20 variables as output\n",
    "rfe = rfe.fit(normalized_X,y)\n",
    "print(rfe.support_)           # Printing the boolean results\n",
    "print(rfe.ranking_)           # Printing the ranking"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['onnet_mou_8', 'offnet_mou_6', 'offnet_mou_8', 'loc_og_t2m_mou_6',\n",
       "       'loc_og_mou_6', 'std_og_t2t_mou_6', 'std_og_mou_8', 'total_og_mou_6',\n",
       "       'total_og_mou_8', 'loc_ic_t2t_mou_8', 'loc_ic_t2m_mou_8',\n",
       "       'loc_ic_mou_7', 'loc_ic_mou_8', 'std_ic_mou_8', 'total_ic_mou_8',\n",
       "       'isd_ic_mou_8', 'total_rech_num_8', 'last_day_rch_amt_8',\n",
       "       'monthly_2g_8', 'sep_vbc_3g'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 74,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Variables selected by RFE \n",
    "col = X.columns[rfe.ranking_==1]\n",
    "col"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Top 20 best features**: 'onnet_mou_8', 'offnet_mou_6', 'offnet_mou_8', 'loc_og_t2m_mou_6', 'loc_og_mou_6', 'std_og_t2t_mou_6', 'std_og_mou_8', 'total_og_mou_6', 'total_og_mou_8', 'loc_ic_t2t_mou_8', 'loc_ic_t2m_mou_8', 'loc_ic_mou_7', 'loc_ic_mou_8', 'std_ic_mou_8', 'total_ic_mou_8', 'isd_ic_mou_8', 'total_rech_num_8', 'last_day_rch_amt_8', 'monthly_2g_8', 'sep_vbc_3g'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LogisticRegression(C=1000000000.0, max_iter=1000)"
      ]
     },
     "execution_count": 75,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Let's run the model using the selected variables\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn import metrics\n",
    "logsk = LogisticRegression(C=1e9, max_iter=1000)\n",
    "logsk.fit(X_train[col], y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table class=\"simpletable\">\n",
       "<caption>Generalized Linear Model Regression Results</caption>\n",
       "<tr>\n",
       "  <th>Dep. Variable:</th>         <td>churn</td>      <th>  No. Observations:  </th>  <td> 21007</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Model:</th>                  <td>GLM</td>       <th>  Df Residuals:      </th>  <td> 20986</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Model Family:</th>        <td>Binomial</td>     <th>  Df Model:          </th>  <td>    20</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Link Function:</th>         <td>logit</td>      <th>  Scale:             </th> <td>  1.0000</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Method:</th>                <td>IRLS</td>       <th>  Log-Likelihood:    </th> <td> -4299.5</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Date:</th>            <td>Wed, 17 Feb 2021</td> <th>  Deviance:          </th> <td>  8598.9</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Time:</th>                <td>18:37:14</td>     <th>  Pearson chi2:      </th> <td>1.25e+06</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>No. Iterations:</th>          <td>9</td>        <th>                     </th>     <td> </td>   \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Covariance Type:</th>     <td>nonrobust</td>    <th>                     </th>     <td> </td>   \n",
       "</tr>\n",
       "</table>\n",
       "<table class=\"simpletable\">\n",
       "<tr>\n",
       "           <td></td>             <th>coef</th>     <th>std err</th>      <th>z</th>      <th>P>|z|</th>  <th>[0.025</th>    <th>0.975]</th>  \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>const</th>              <td>   -0.8234</td> <td>    0.062</td> <td>  -13.268</td> <td> 0.000</td> <td>   -0.945</td> <td>   -0.702</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>onnet_mou_8</th>        <td>    0.0011</td> <td>    0.000</td> <td>    4.931</td> <td> 0.000</td> <td>    0.001</td> <td>    0.002</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>offnet_mou_6</th>       <td>    0.0020</td> <td>    0.000</td> <td>    6.698</td> <td> 0.000</td> <td>    0.001</td> <td>    0.003</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>offnet_mou_8</th>       <td>    0.0008</td> <td>    0.000</td> <td>    3.555</td> <td> 0.000</td> <td>    0.000</td> <td>    0.001</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>loc_og_t2m_mou_6</th>   <td>   -0.0025</td> <td>    0.000</td> <td>   -6.232</td> <td> 0.000</td> <td>   -0.003</td> <td>   -0.002</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>loc_og_mou_6</th>       <td>    0.0018</td> <td>    0.000</td> <td>    5.087</td> <td> 0.000</td> <td>    0.001</td> <td>    0.003</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>std_og_t2t_mou_6</th>   <td>    0.0018</td> <td>    0.000</td> <td>    6.001</td> <td> 0.000</td> <td>    0.001</td> <td>    0.002</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>std_og_mou_8</th>       <td>    0.0011</td> <td>    0.000</td> <td>    3.407</td> <td> 0.001</td> <td>    0.000</td> <td>    0.002</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>total_og_mou_6</th>     <td>   -0.0013</td> <td>    0.000</td> <td>   -4.615</td> <td> 0.000</td> <td>   -0.002</td> <td>   -0.001</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>total_og_mou_8</th>     <td>   -0.0029</td> <td>    0.000</td> <td>   -8.075</td> <td> 0.000</td> <td>   -0.004</td> <td>   -0.002</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>loc_ic_t2t_mou_8</th>   <td>    0.0065</td> <td>    0.003</td> <td>    2.105</td> <td> 0.035</td> <td>    0.000</td> <td>    0.013</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>loc_ic_t2m_mou_8</th>   <td>    0.0067</td> <td>    0.003</td> <td>    2.191</td> <td> 0.028</td> <td>    0.001</td> <td>    0.013</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>loc_ic_mou_7</th>       <td>    0.0021</td> <td>    0.000</td> <td>    9.222</td> <td> 0.000</td> <td>    0.002</td> <td>    0.003</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>loc_ic_mou_8</th>       <td>    0.0080</td> <td>    0.003</td> <td>    2.476</td> <td> 0.013</td> <td>    0.002</td> <td>    0.014</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>std_ic_mou_8</th>       <td>    0.0205</td> <td>    0.002</td> <td>   11.928</td> <td> 0.000</td> <td>    0.017</td> <td>    0.024</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>total_ic_mou_8</th>     <td>   -0.0221</td> <td>    0.002</td> <td>  -13.580</td> <td> 0.000</td> <td>   -0.025</td> <td>   -0.019</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>isd_ic_mou_8</th>       <td>    0.0219</td> <td>    0.002</td> <td>   12.840</td> <td> 0.000</td> <td>    0.019</td> <td>    0.025</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>total_rech_num_8</th>   <td>   -0.0477</td> <td>    0.005</td> <td>   -8.820</td> <td> 0.000</td> <td>   -0.058</td> <td>   -0.037</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>last_day_rch_amt_8</th> <td>   -0.0054</td> <td>    0.000</td> <td>  -12.161</td> <td> 0.000</td> <td>   -0.006</td> <td>   -0.005</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>monthly_2g_8</th>       <td>   -1.3425</td> <td>    0.190</td> <td>   -7.076</td> <td> 0.000</td> <td>   -1.714</td> <td>   -0.971</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>sep_vbc_3g</th>         <td>   -0.0144</td> <td>    0.004</td> <td>   -3.589</td> <td> 0.000</td> <td>   -0.022</td> <td>   -0.007</td>\n",
       "</tr>\n",
       "</table>"
      ],
      "text/plain": [
       "<class 'statsmodels.iolib.summary.Summary'>\n",
       "\"\"\"\n",
       "                 Generalized Linear Model Regression Results                  \n",
       "==============================================================================\n",
       "Dep. Variable:                  churn   No. Observations:                21007\n",
       "Model:                            GLM   Df Residuals:                    20986\n",
       "Model Family:                Binomial   Df Model:                           20\n",
       "Link Function:                  logit   Scale:                          1.0000\n",
       "Method:                          IRLS   Log-Likelihood:                -4299.5\n",
       "Date:                Wed, 17 Feb 2021   Deviance:                       8598.9\n",
       "Time:                        18:37:14   Pearson chi2:                 1.25e+06\n",
       "No. Iterations:                     9                                         \n",
       "Covariance Type:            nonrobust                                         \n",
       "======================================================================================\n",
       "                         coef    std err          z      P>|z|      [0.025      0.975]\n",
       "--------------------------------------------------------------------------------------\n",
       "const                 -0.8234      0.062    -13.268      0.000      -0.945      -0.702\n",
       "onnet_mou_8            0.0011      0.000      4.931      0.000       0.001       0.002\n",
       "offnet_mou_6           0.0020      0.000      6.698      0.000       0.001       0.003\n",
       "offnet_mou_8           0.0008      0.000      3.555      0.000       0.000       0.001\n",
       "loc_og_t2m_mou_6      -0.0025      0.000     -6.232      0.000      -0.003      -0.002\n",
       "loc_og_mou_6           0.0018      0.000      5.087      0.000       0.001       0.003\n",
       "std_og_t2t_mou_6       0.0018      0.000      6.001      0.000       0.001       0.002\n",
       "std_og_mou_8           0.0011      0.000      3.407      0.001       0.000       0.002\n",
       "total_og_mou_6        -0.0013      0.000     -4.615      0.000      -0.002      -0.001\n",
       "total_og_mou_8        -0.0029      0.000     -8.075      0.000      -0.004      -0.002\n",
       "loc_ic_t2t_mou_8       0.0065      0.003      2.105      0.035       0.000       0.013\n",
       "loc_ic_t2m_mou_8       0.0067      0.003      2.191      0.028       0.001       0.013\n",
       "loc_ic_mou_7           0.0021      0.000      9.222      0.000       0.002       0.003\n",
       "loc_ic_mou_8           0.0080      0.003      2.476      0.013       0.002       0.014\n",
       "std_ic_mou_8           0.0205      0.002     11.928      0.000       0.017       0.024\n",
       "total_ic_mou_8        -0.0221      0.002    -13.580      0.000      -0.025      -0.019\n",
       "isd_ic_mou_8           0.0219      0.002     12.840      0.000       0.019       0.025\n",
       "total_rech_num_8      -0.0477      0.005     -8.820      0.000      -0.058      -0.037\n",
       "last_day_rch_amt_8    -0.0054      0.000    -12.161      0.000      -0.006      -0.005\n",
       "monthly_2g_8          -1.3425      0.190     -7.076      0.000      -1.714      -0.971\n",
       "sep_vbc_3g            -0.0144      0.004     -3.589      0.000      -0.022      -0.007\n",
       "======================================================================================\n",
       "\"\"\""
      ]
     },
     "execution_count": 76,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "logm4 = sm.GLM(y_train,(sm.add_constant(X_train[col])), family = sm.families.Binomial())\n",
    "modres = logm4.fit()\n",
    "logm4.fit().summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "8869     0.335748\n",
       "14992    0.143787\n",
       "27111    0.023633\n",
       "26075    0.011680\n",
       "27484    0.019369\n",
       "           ...   \n",
       "6607     0.002287\n",
       "22779    0.038859\n",
       "29195    0.004612\n",
       "12122    0.051627\n",
       "5658     0.218232\n",
       "Length: 9004, dtype: float64"
      ]
     },
     "execution_count": 77,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "res = modres.predict(sm.add_constant(X_test[col]))\n",
    "res"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Making Predictions**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [
    {
     "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>1</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.337124</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.134123</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.020708</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.012784</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.022221</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          1\n",
       "0  0.337124\n",
       "1  0.134123\n",
       "2  0.020708\n",
       "3  0.012784\n",
       "4  0.022221"
      ]
     },
     "execution_count": 78,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Predicted probabilities\n",
    "y_pred = logsk.predict_proba(X_test[col])\n",
    "# Converting y_pred to a dataframe which is an array\n",
    "y_pred_df = pd.DataFrame(y_pred)\n",
    "# Converting to column dataframe\n",
    "y_pred_1 = y_pred_df.iloc[:,[1]]\n",
    "# Let's see the head\n",
    "y_pred_1.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "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>churn</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>8869</th>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14992</th>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27111</th>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26075</th>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27484</th>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       churn\n",
       "8869     0.0\n",
       "14992    0.0\n",
       "27111    0.0\n",
       "26075    0.0\n",
       "27484    0.0"
      ]
     },
     "execution_count": 79,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Converting y_test to dataframe\n",
    "y_test_df = pd.DataFrame(y_test)\n",
    "y_test_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [
    {
     "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>CustID</th>\n",
       "      <th>churn</th>\n",
       "      <th>Churn_Prob</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>8869</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.337124</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>14992</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.134123</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>27111</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.020708</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>26075</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.012784</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>27484</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.022221</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   CustID  churn  Churn_Prob\n",
       "0    8869    0.0    0.337124\n",
       "1   14992    0.0    0.134123\n",
       "2   27111    0.0    0.020708\n",
       "3   26075    0.0    0.012784\n",
       "4   27484    0.0    0.022221"
      ]
     },
     "execution_count": 80,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Putting CustID to index\n",
    "y_test_df['CustID'] = y_test_df.index\n",
    "# Removing index for both dataframes to append them side by side \n",
    "y_pred_1.reset_index(drop=True, inplace=True)\n",
    "y_test_df.reset_index(drop=True, inplace=True)\n",
    "# Appending y_test_df and y_pred_1\n",
    "y_pred_final = pd.concat([y_test_df,y_pred_1],axis=1)\n",
    "# Renaming the column \n",
    "y_pred_final= y_pred_final.rename(columns={ 1 : 'Churn_Prob'})\n",
    "# Rearranging the columns\n",
    "y_pred_final = y_pred_final.reindex(['CustID','churn','Churn_Prob'], axis=1)\n",
    "# Let's see the head of y_pred_final\n",
    "y_pred_final.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [
    {
     "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>CustID</th>\n",
       "      <th>churn</th>\n",
       "      <th>Churn_Prob</th>\n",
       "      <th>predicted</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>8869</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.337124</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>14992</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.134123</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>27111</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.020708</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>26075</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.012784</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>27484</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.022221</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   CustID  churn  Churn_Prob  predicted\n",
       "0    8869    0.0    0.337124          0\n",
       "1   14992    0.0    0.134123          0\n",
       "2   27111    0.0    0.020708          0\n",
       "3   26075    0.0    0.012784          0\n",
       "4   27484    0.0    0.022221          0"
      ]
     },
     "execution_count": 81,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Creating new column 'predicted' with 1 if Churn_Prob>0.5 else 0\n",
    "y_pred_final['predicted'] = y_pred_final.Churn_Prob.map( lambda x: 1 if x > 0.5 else 0)\n",
    "# Let's see the head\n",
    "y_pred_final.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Model Evaluation**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[8121,   66],\n",
       "       [ 627,  190]], dtype=int64)"
      ]
     },
     "execution_count": 82,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Confusion matrix \n",
    "confusion = metrics.confusion_matrix( y_pred_final.churn, y_pred_final.predicted )\n",
    "confusion"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9230342070191027"
      ]
     },
     "execution_count": 83,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "metrics.accuracy_score(y_pred_final.churn, y_pred_final.predicted)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7421875"
      ]
     },
     "execution_count": 84,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "metrics.precision_score(y_pred_final.churn, y_pred_final.predicted)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [],
   "source": [
    "def draw_roc( actual, probs ):\n",
    "    fpr, tpr, thresholds = metrics.roc_curve( actual, probs,\n",
    "                                              drop_intermediate = False )\n",
    "    auc_score = metrics.roc_auc_score( actual, probs )\n",
    "    plt.figure(figsize=(6, 6))\n",
    "    plt.plot( fpr, tpr, label='ROC curve (area = %0.2f)' % auc_score )\n",
    "    plt.plot([0, 1], [0, 1], 'k--')\n",
    "    plt.xlim([0.0, 1.0])\n",
    "    plt.ylim([0.0, 1.05])\n",
    "    plt.xlabel('False Positive Rate or [1 - True Negative Rate]')\n",
    "    plt.ylabel('True Positive Rate')\n",
    "    plt.title('Receiver operating characteristic example')\n",
    "    plt.legend(loc=\"lower right\")\n",
    "    plt.show()\n",
    "\n",
    "    return fpr, tpr, thresholds"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYoAAAGDCAYAAAA1cVfYAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/d3fzzAAAACXBIWXMAAAsTAAALEwEAmpwYAABddElEQVR4nO3dd3gUVRfA4d9JAiT03ntvUiMdAekI0pRqQUVsVBFFBUXsBVEEFRUE+ZDQBemoNEFBMKH33nsJISFl7/fHDLiEkCwhm80m532ePNkyc+/Z2d05e++duSPGGJRSSqk78fF0AEoppVI2TRRKKaXipYlCKaVUvDRRKKWUipcmCqWUUvHSRKGUUipemihSCRHZLiKNPR2Hp4nItyIyPJnrnCQi7yVnne4iIj1FZFki1021n0ERMSJS2tNxeIroeRRJT0QOAfmAGOAqsAToa4y56sm4UhsR6QX0NsY08HAck4BjxphhHo5jBFDaGPNYMtQ1iRTwmpOLiBigjDFmn6dj8QRtUbhPO2NMZqAaUB143bPh3D0R8UuLdXuSbnOVIhlj9C+J/4BDQDOn+58AC53u1wHWAZeAzUBjp+dyAj8CJ4CLwC9Oz7UFQuz11gFVYtcJFATCgZxOz1UHzgHp7PtPAzvt8pcCxZyWNcBLwF7g4B1e38PAdjuOlUCFWHG8Duywy/8R8L+L1/AasAW4DvgBQ4H9QKhdZkd72QpABP+12i7Zj08C3rNvNwaOAYOBM8BJ4Cmn+nIBvwJXgH+A94A/43lfGzi9b0eBXk51jgMW2nGuB0o5rfelvfwVYBPQ0Om5EcAs4H/2872BWsBfdj0ngbFAeqd1KgHLgQvAaeANoBUQCUTZ22OzvWw2YIJdznH7Nfraz/UC1gKj7bLesx/7035e7OfOAJft96Uy0MeuJ9Ku69fYn3vA147rxnu3CShyh+0a5/cBqIf1uS1i369qL1Pevh/nZyOO13YJOGCX18t+L84ATzotPwn41t6uocAqbv9elLZvZwA+A47Y2/9bIMDT+x237tM8HUBq/Iv1hSkMbAW+tO8XAs4DbbBadM3t+3ns5xcC04EcQDqgkf14DfvDXdv+Ej5p15Mhjjr/AJ51iudT4Fv7dgdgH9aO1g8YBqxzWtbYX5accX34gbJAmB13OuBVu7z0TnFsA4rYZazlvx23K68hxF43wH7sUazk5wN0tesuYD/Xi1g7dm5PFNHASDvWNsA1IIf9fJD9lxGoiLUDiTNRAEWxdiDd7bJyAdWc6ryAtYP3A6YCQU7rPmYv74eVtE5hJ0+sRBFlvy8+QABQE2vn6QcUx0rqA+3ls2Dt9AcD/vb92k5l/S9W3L8A44FMQF5gA/Cc0/aLBvrZdQVwa6JoibWDz46VNCo4bfub2/kOn/shWJ/7cva6VYFccWzXhL4P72N9ngOwElVfp3UT+mxEA09hfdbew9qxj8Pa0bew38/MTq8nFHjAfv5LnD4L3JoovgDmY32+s2D92PjQ0/sdt+7TPB1AavyzvzBX7Q+eAX4HstvPvQZMibX8UqydZgHAgb0ji7XMN8C7sR7bzX+JxPlL2hv4w74tWDvAB+z7i4FnnMrwwdp5FrPvG+DBeF7bcGBGrPWP89+vwEPA807PtwH238VreDqBbRsCtLdv9yLhRBEO+Dk9fwZrJ+yLtYMu5/TcHVsUWK2kuXd4bhLwQ6zXvCue13ARqGrfHgGsTuA1D7xRN1aiCr7DciNwShRY42TXcUr49vornLbfkVhl3NymwIPAHnt7+dxpO8f63N/4DO6+8T4l8Nru+H2wb6fDSlZbscb65C4+G3udnrsP67Odz+mx89ya7J2Te2as1uqN1owBSmN9n8K4tcVYlzu0vlPLn45RuE8HY0wWrJ1VeSC3/Xgx4FERuXTjD6tLowDWL+kLxpiLcZRXDBgca70iWL+oYpsF1BWRgli/kAywxqmcL53KuID14S/ktP7ReF5XQeDwjTvGGIe9/J3WP+wUoyuv4Za6ReQJEQlxWr4y/21LV5w3xkQ73b+GtRPIg/Ur2rm++F53Eaxujjs5FUcdAIjIYBHZKSKX7deQjVtfQ+zXXFZEFojIKRG5AnzgtHxCcTgrhrWjPem0/cZjtSzirNuZMeYPrG6vccBpEflORLK6WLerccb3fcAYE4W1E68MjDL2nhlc+mycdrodbpcX+7HMTvdvbgtjHXhygdu/X3mwWqCbnOpdYj+eammicDNjzCqsD/pn9kNHsX5BZXf6y2SM+ch+LqeIZI+jqKPA+7HWy2iMmRZHnZeAZUAXoAcwzekLdhSr68G5nABjzDrnIuJ5SSewvtwAiIhg7RSOOy1TxOl2UXsdV1+D846gGPA90Ber2yI7VreWuBBnQs5idU0UvkPcsR0FSt1tJSLSEOtXcxeslmJ2rP5+cVos9uv4BtiFdZRNVqy+/hvLxxdH7HKOYrUocjtt76zGmErxrHNrgcaMMcbUxBoXKYvVpZTgegnEGXu5O30fEJFCwNtYY12jRCSD/XhCn43EuPn+i0hmrK6lE7GWOYeVYCo5xZvNWAeupFqaKJLHF0BzEamGNWjZTkRaioiviPiLSGMRKWyMOYnVNfS1iOQQkXQi8oBdxvfA8yJSWyyZROQhEclyhzp/Bp4AOtu3b/gWeF1EKgGISDYRefQuXssM4CERaSoi6bD6yq9jDUbe8JKIFBaRnFg7uemJfA2ZsHZIZ+1Yn8L61XjDaaCwiKS/i/gBMMbEAHOAESKSUUTKY22vO5kKNBORLiLiJyK57PczIVmwEtJZwE9E3gIS+lWeBWtg+6od1wtOzy0A8ovIQBHJICJZRKS2/dxpoLiI+Niv8STWD4ZRIpJVRHxEpJSINHIhbkTkfvu9SofV3XLj4IEbdZWMZ/UfgHdFpIz9XlcRkVxxLHfH74P9I2QS1mD8M1hjM+/a6yX02UiMNiLSwP48vQusN8bc0uKyW9DfA6NFJK9ddyERaXmPdadomiiSgTHmLPATMNz+4LXH2oGexfpFNYT/3ovHsfrOd2H1pw+0y9gIPIvVFXARawC5VzzVzgfKAKeNMZudYpkLfAwE2d0a24DWd/FadmMNzn6F9euqHdahwJFOi/2MtYM6YP+9l5jXYIzZAYzCOgLoNFY/81qnRf7AOvrqlIicc/U1OOmL1Q10CpgCTMNKenHFcgRr7GEwVpdECNYAbUKWYiX/PVjdcBHE38UF8ApWSzAUa6d0I9FijAnFGvBtZ8e9F2hiPz3T/n9eRP61bz8BpOe/o9BmYXfruCCrXf9FO/bz/NcyngBUtLtffolj3c+xflQsw0p6E7AGpG+RwPehP9Y4y3C7RfwU8JSINHThs5EYP2O1Xi5gHVDQ8w7LvYb12f3b/g79hjVon2rpCXcqSYl1smFvY8xvno7lbonIx0B+Y8yTno5FJS9JYycQ3i1tUag0S0TK210iIiK1sLo35no6LqVSGj0TU6VlWbC6mwpidfONAuZ5NCKlUiDtelJKKRUv7XpSSikVL00USiml4uV1YxS5c+c2xYsX93QYSinlVTZt2nTOGJOoM8i9LlEUL16cjRs3ejoMpZTyKiJyOOGl4qZdT0oppeKliUIppVS8NFEopZSKlyYKpZRS8dJEoZRSKl6aKJRSSsVLE4VSSql4aaJQSikVL00USiml4qWJQimlVLzclihEZKKInBGRbXd4XkRkjIjsE5EtIlLDXbEopZRKPHe2KCYBreJ5vjXWNZ3LAH2Ab9wYi1JKqURyW6IwxqzGukj5nbQHfjKWv4HsIuLqRd+VUkq5wBjD57NW3VMZnhyjKAQcdbp/zH7sNiLSR0Q2isjGs2fPJktwSinl7S6GRVKnW38Gd3nwnsrxZKKQOB6L87qsxpjvjDGBxpjAPHkSNZ26UkqlKWv3naVS68fZMGMs9zdpc09leTJRHAOKON0vDJzwUCxKKZUqRMU4+GTJLnp+/xdRF0/S9cne/L183j2V6ckLF80H+opIEFAbuGyMOenBeJRSyqsdOhdG3ynr2XzgJD0b38frby0la0Z/ROLqwHGd2xKFiEwDGgO5ReQY8DaQDsAY8y2wCGgD7AOuAU+5KxallErNjDHM/vc4w2f+w7GZ75HbN5yRo0JInz59kpTvtkRhjOmewPMGeMld9SulVFpwOTyKYb9s45e/d3Pt1/e4dngnb0+YkGRJArzwmtlKKaUsGw9dYEBQCMeOHydmwbtcPXGY2bNn06FDhyStR6fwUEopLxMd42D08j10Gf8Xvj5CsZ1TCT13ksWLFyd5kgBtUSillFc5euEaA6eHsOnwRTrVKMQ7D1fiSs+ynDp1ipo1a7qlTm1RKKWUl5i/+QRtvlzDnlOhvFA+mstLvyLATyhUqJDbkgRoi0IppVK8q9ejeWveNub8e5yaxXLQIfcZnnuyB4ULF+b8+fPky5fPrfVrolBKqRQs5OglBgQFc/TCNQY0LUO+c5vo1fNJKlWqxNKlS8mbN6/bY9CuJ6WUSoFiHIZxK/bxyDfriI4xTH+uLtmO/sljPXtSp04dVq5cmSxJArRFoZRSKc6JS+EMmh7C+oMXaFulAO93vI9sAemIrlCBLl268OOPPxIQEJBs8WiiUEqpFGTx1pMMnbOVqBgHnz1alU7VC7JixQqaNm1K3bp1qVu3brLHpF1PSimVAlyLjOb1OVt4Yeq/FM+VkUX9G9Khan6effZZmjVrxpo1azwWm7YolFLKw7Ydv0z/oGAOngvjxcalGNS8LDFRkXTp0oW5c+fy9ttv06BBA4/Fp4lCKaU8xOEwTPjzIJ8s3UWuTBmY2rs29UrlJjQ0lA4dOvDHH3/wxRdfMGDAAI/GqYlCKaU84MyVCAbP3MyavedoWSkfH3WqQo5M1kR+y5cvZ/Xq1UyePJknnnjCw5FqolBKqWT3247TvDp7C9cio/mg4310r1UEESEmJgZfX186derE7t27KVmypKdDBXQwWymlkk1EVAxvzdtG7582kj+rPwv6NaRH7aKICHv37uW+++67OWidUpIEaItCKaWSxa5TV+g/LZg9p6/ybMMSvNKyHBn8fAEICQmhZcuWOBwOMmbM6OFIb6eJQiml3MgYw+R1h/hg8S6y+qfjp6dr8UDZPDefX7NmDW3btiVbtmwsW7aM8uXLezDauGmiUEopNzl39TpDZm5mxe6zPFg+L588UoXcmTPcfH7Lli20aNGCYsWKsWzZMooWLerBaO9ME4VSSrnBqj1nGTxjM1ciohjZvhKP1ymGiNyyTKVKlXj11Vfp27cvefLkuUNJnqeJQimlktD16Bg+XbKbH/48SLl8Wfhf71qUz5/1lmV+/PFHmjZtStGiRXnnnXc8FKnr9KgnpZRKIvvOhNJx3Dp++PMgT9Ytxry+9W9JEsYY3nvvPZ5++mlGjRrlwUjvjrYolFLqHhlj+HnDEd5dsIOM6f2Y8GQgTSvcejEhh8PB4MGD+eKLL3jiiSc0USilVFpxMSySoXO2sHT7aRqWyc2oR6uSN6v/LctER0fTu3dvJk+ezIABA/j888/x8fGeDh1NFEoplUjr9p3j5RmbOR92nWEPVeDp+iXw8ZHblrt27RqbN29m5MiRDBs27LZB7ZROE4VSSt2lyGgHny/fw/jV+ymROxM/PFmfyoWy3bZcaGgofn5+ZM2alb/++gt/f/84Skv5NFEopdRdOHgujAFBwWw5dpnutYoyvG0FMqa/fVd67tw5WrduTeHChZkzZ47XJgnQRKGUUi4xxjBr0zHenr+ddL4+fPtYDVpVLhDnskePHqVFixYcOnSIt99+2+u6mmLTRKGUUgm4HB7FG3O3snDLSeqUzMnortUokC3ua1bv3r2b5s2bc/nyZZYuXcoDDzyQzNEmPU0USikVj38OXWBgUAinr0TwaqtyPPdAKXzjGLAG6xDYTp06cf36dVauXEn16tWTOVr30EShlFJxiI5xMOaPfYz9Yy9FcmZk1gv1qFYke7zr+Pj4MGXKFDJnzkzZsmWTJ9BkoIlCKaViOXrhGgOCgvn3yCU61yjMO+0rkTnDnXeXCxYsIDg4mOHDh1OjRo1kjDR5eM8ZH0oplQzmhRynzZdr2Hv6KmO6V2dUl6rxJompU6fSoUMH5s+fT0RERDJGmny0RaGUUkBoRBRvz9vOnODjBBbLweiu1SiSM/6LCH311Vf079+fBx98kF9++cWrD4GNjyYKpVSaF3zkIgOCQjh28RoDm5Whb5PS+PnG3+Hy3nvvMXz4cDp06MC0adNSbZIATRRKqTQsxmH4ZuU+Rv+2l/xZ/ZnxXF0Ci+d0ad3ixYvz9NNPM378ePz8UveuVIwxno7hrgQGBpqNGzd6OgyllJc7cSmcgdND2HDwAu2qFuS9DpXJFpAu3nWioqL4999/qV27djJFmXREZJMxJjAx66buNKiUUnFYtPUkQ2dvIcZhGPVoVTrVKJTg2dPh4eF07dqVpUuXsmfPHooVK5ZM0XqeJgqlVJpxLTKad+bvYPrGo1QtnI0vu1WneO5MCa53+fJlHn74YdasWcO4cePSVJIATRRKqTRi67HLDAgK5uD5MF5qUoqBzcqSLoEBa4AzZ87QqlUrtm7dytSpU+nevXsyRJuyaKJQSqVqDofhhz8P8OnS3eTKlIGfe9ehbqlcLq8/ceJEdu3axbx582jTpo0bI025dDBbKZVqnb4SweAZm/lz3zlaVcrPR53vI3vG9C6t63A48PHxweFwsGfPHsqXL+/maN3rXgaz9cxspVSqtHzHaVp9sZpNhy/yUaf7+OaxGi4niY0bN1K9enUOHDiAj4+P1yeJe6VdT0qpVCUiKob3F+5kyt+HqVQwK192q07pvJldXn/FihU8/PDD5MqVC4fD4cZIvYcmCqVUqrHz5BX6Twtm75mrPNuwBK+0LEcGP1+X1583bx5du3alVKlSLFu2jEKFCrkxWu+hiUIp5fWMMUxad4gPF+8iW0A6pjxTi4Zl8txVGUuWLKFz587UrFmTRYsWkSuX6wPeqZ0mCqWUVzt39TpDZm5mxe6zNC2fl08eqUKuzBnuupz69evTr18/3n33XTJndr2rKi3QwWyllNdaufsMrb5Yzbr95xnZvhI/PBl4V0nCGMP48eO5evUqWbJkYfTo0Zok4qAtCqWU17keHcPHi3czce1ByuXLwtTedSiXP8tdleFwOOjXrx9ff/01ERERDBgwwE3Rej9NFEopr7L3dCj9g0LYefIKveoVZ2jr8vinc33AGqzJ/Xr16sXPP//MkCFD6N+/v5uiTR00USilvIIxhqnrj/Dugh1kzuDHxF6BPFg+312Xc+3aNbp06cLChQv58MMPGTp0qBuiTV00USilUrwLYZG8NnsLy3ecpmGZ3IzqUpW8WRJ3oaAzZ84QHBzMt99+y3PPPZfEkaZOmiiUUina2n3neHlGCBfDohj2UAWerl8CH5/4pwSPy6VLl8iWLRvFixdn9+7dOmh9F/SoJ6VUihQZ7eCjxbt4bMJ6MmfwY+5L9ejdsGSiksShQ4e4//77GTZsGIAmibvk1kQhIq1EZLeI7BOR2zoCRSSbiPwqIptFZLuIPOXOeJRS3uHA2at0/mYd367aT/daRVnQryGVCmZLVFk7duygQYMGnDt3jrZt2yZxpGmD27qeRMQXGAc0B44B/4jIfGPMDqfFXgJ2GGPaiUgeYLeITDXGRLorLqVUymWMYebGY4z4dTvp/Xz49rGatKqcP9HlbdiwgdatW5M+fXpWrVpFlSpVkjDatMOdYxS1gH3GmAMAIhIEtAecE4UBsoh1DcLMwAUg2o0xKaVSqMvXonjjl60s3HKSuiVzMbprNfJnS9yANcCVK1do3bo12bJlY/ny5ZQqVSoJo01b3JkoCgFHne4fA2JfkXwsMB84AWQBuhpjbpuuUUT6AH0AihYt6pZglVKes+HgBQYGBXMm9DqvtSpPnwdK4puIsQhnWbNmZcqUKVSrVo2CBQsmUaRpkzsTRVzvcuyrJLUEQoAHgVLAchFZY4y5cstKxnwHfAfWhYuSPlSllCdExTgY8/texq3YR9GcGZn9Qj2qFsl+T2X++OOP+Pv707179zR7Rbqk5s7B7GNAEaf7hbFaDs6eAuYYyz7gIJC2rxCiVBpx5Pw1uoz/i6/+2EfnGoVZ2L/hPSeJUaNG8fTTTzN16lS87eqdKZk7WxT/AGVEpARwHOgG9Ii1zBGgKbBGRPIB5YADboxJKZUC/BJ8nGG/bEMEvupenXZV761ryBjDsGHD+OCDD3j00UeZMmUK1tCnSgpuSxTGmGgR6QssBXyBicaY7SLyvP38t8C7wCQR2YrVVfWaMeacu2JSSnlWaEQUb83bztzg49xfPAeju1ajcI6M91SmMYYXXniB8ePH8+yzz/LNN9/g63t3cz+p+Ln1zGxjzCJgUazHvnW6fQJo4c4YlFIpw79HLjIgKJgTlyJ4uXlZXmxcCj/fe+/9FhFy5szJa6+9xocffqgtCTfQKTyUUm4V4zB8vWIfX/y+lwLZ/JnxXB1qFst5z+Veu3aNw4cPU6FCBd5//31NEG6kiUIp5TbHL4UzKCiEDYcu8HDVgrzXsTJZ/dPdc7kXL16kbdu2HDhwgL179+qUHG6miUIp5RYLt5zk9TlbiHEYPu9SlY7VCyXJr/5Tp07RsmVLdu7cyc8//6xJIhloolBKJamw69G88+t2Zmw8RrUi2fmyWzWK5cqUJGUfPHiQ5s2bc+rUKRYuXEjz5s2TpFwVP00USqkks/XYZfoHBXPofBh9m5RmQLMypEuCAesb3n33XS5cuMBvv/1GnTp1kqxcFT/xtpNSAgMDzcaNGz0dhlLKicNh+G7NAUYt203uzBkY3bUadUrmSrLyjTGICNeuXePo0aOUK1cuycpOK0RkkzEmMDHr6vUolFL35NTlCB6fuJ6PFu+iecV8LB7QMEmTxPLly2ncuDFXrlwhY8aMmiQ8QLuelFKJtmz7KV6bvYWIKAcfd76PLoFFkvQw1dmzZ9O9e3cqVKhAeHg4WbNmTbKyles0USil7lp4ZAzvLdzB1PVHqFwoK192q06pPEl79NGECRPo06cPderUYcGCBeTIkSNJy1eu00ShlLorO05coX9QMPvOXOW5B0oyuEU50vslbS/2hAkT6N27N61atWLWrFlkypQ0R02pxNFEoZRyiTGGH9ce4qPFu8ieMR3/e6Y2DcrkdktdzZo1o1+/fnz22WekT5/eLXUo1+lgtlIqQWdDr9Prx38YuWAHD5TNzeIBDZM8ScTExPDjjz/icDgoVqwYY8aM0SSRQmiLQikVrxW7zzBk5mZCI6J5t0NlHqtdNMnnVYqMjOSxxx5j5syZ5MmTh7Zt2yZp+ereaKJQSsUpIiqGj5fs4se1hyifPws/P1uHsvmyJHk9YWFhdOrUiWXLlvHZZ59pkkiBNFEopW6z53Qo/acFs+tUKL3qFWdo6/L4p0v6azxcuHCBtm3bsn79eiZMmMDTTz+d5HWoe6eJQil1kzGG/60/wnsLdpA5gx8/9rqfJuXzuq2+3bt3s2PHDmbOnEmnTp3cVo+6N5oolFIAXAiL5NVZW/ht52kalc3Dp49WIW8Wf7fUFRoaSpYsWahbty6HDh0ie/bsbqlHJQ096kkpxZ97z9Hqi9Ws3nOW4W0r8mOv+92WJLZu3Uq5cuWYMmUKgCYJL6AtCqXSsMhoB6OW7Wb86gOUzpuZSU/VomJB902TsW7dOh566CEyZcpEzZo13VaPSlqaKJRKow6cvcqAoBC2Hr9Mz9pFGfZQRQLSJ/2A9Q1Lly6lU6dOFCxYkOXLl1O8eHG31aWSliYKpdIYYwwzNh5lxPwdZEjnw/jHa9KyUn631nngwAHatWtHpUqVWLJkCfny5XNrfSppaaJQKg25fC2K1+duYdHWU9QrlYvPu1Qjfzb3jEU4K1myJN9//z3t27fXMQkvpIlCqTRi/YHzDJoewpnQ6wxtXZ4+DUvi45O0Z1jHNmrUKBo0aEDt2rV58skn3VqXch896kmpVC4qxhqw7vb936T382HOi/V4vlEptyYJYwyvvvoqr7zyCj/99JPb6lHJQ1sUSqViR85fo39QMCFHL9ElsDBvt6tEpgzu/drHxMTw3HPPMWHCBF566SXGjBnj1vqU+2miUCqVmht8jOG/bEcExvaoTtsqBd1eZ2RkJD169GD27NkMHz6cd955J8knEFTJTxOFUqnMlYgo3vplG7+EnKBW8ZyM7laNQtkDkqVuHx8fHA4Ho0ePZuDAgclSp3I/TRRKpSKbDl9kQFAwJy9HMLh5WV5sUhpfNw9YA5w/f57IyEgKFCjA7NmztRWRymiiUCoViHEYxq3Yx5e/76VANn9mPFeXmsWS5xrTx48fp0WLFmTKlIm///4bHx89Ria10UShlJc7fimcgUHB/HPoIh2qFWRkh8pk9U+XLHXv27eP5s2bc+7cOebPn69JIpXSRKGUF1uw5QSvz9mKMTC6a1U6Vi+cbHVv3ryZli1bEh0dzYoVKwgMDEy2ulXy0kShlBcKux7NiPnbmbnpGNWKZGdMt+oUzZUx2eo3xvDSSy+RLl06VqxYQYUKFZKtbpX8NFEo5WU2H73EgKBgDl+4Rr8HS9O/aRnS+SZvl4+IMH36dKKjoylWrFiy1q2Sn3YoKuUlHA7DNyv30/mbdURGOwh6tg6DW5RL1iQRFBREjx49iImJoVChQpok0ghNFEp5gVOXI3hswno+XrKLFpXysXjAA9QumStZY/j222/p0aMHx48fJzw8PFnrVp6lXU9KpXBLt5/itdlbuB7l4JPOVXg0sHCynqdgjOHDDz/kzTffpG3btsyYMYOAgOQ5gU+lDJoolEqhwiNjeHfhDn5ef4T7CmXjy27VKJknc7LHMWLECEaOHMljjz3GxIkTSZcueQ69VSmHy4lCRDIZY8LcGYxSyrL9xGX6Twtm/9kwnmtUksHNy5HezzM9xa1atSI8PJyPPvpIz5NIoxJ810WknojsAHba96uKyNduj0ypNMjhMPyw5gAdx60jNCKa/z1Tm9dbV0j2JBEREcHs2bMBqFu3Lp988okmiTTMlRbFaKAlMB/AGLNZRB5wa1RKpUFnQiN4ZeYWVu85S7MK+fjkkSrkzJQ+2eMIDQ2lQ4cOrFixgi1btlC5cuVkj0GlLC51PRljjsYaPItxTzhKpU0rdp3hlZmbuXo9mvc6VKZn7aIemVjv3LlztGnThn///ZfJkydrklCAa4niqIjUA4yIpAf6Y3dDKaXuTURUDB8t3sWkdYconz8LQX3qUCZfFo/EcuzYMVq0aMGBAweYO3cu7dq180gcKuVxJVE8D3wJFAKOAcuAF90ZlFJpwe5ToQwICmbXqVCerl+CV1uVwz+dr8fiWbVqFcePH2fp0qU0atTIY3GolMeVRFHOGNPT+QERqQ+sdU9ISqVuxhj+9/dh3lu4kyz+fvz41P00KZfXY/GEh4cTEBBAz549adGiBXny5PFYLCplcuUwhq9cfEwplYDzV6/z7E8bGT5vO3VL5WLxgAc8miTWrFlDyZIlWbvW+t2nSULF5Y4tChGpC9QD8ojIy05PZQU81z5Wykut2XuWl2ds5vK1KN5uV5Fe9Yp79EpwCxcu5JFHHqFYsWIUKVLEY3GolC++rqf0QGZ7GefRtSvAI+4MSqnUJDLawWfLdvPd6gOUyZuZyU/VomLBrB6NaerUqfTq1YsqVaqwZMkSbUmoeN0xURhjVgGrRGSSMeZwMsakVKqx/+xV+k8LZvuJKzxWpyhvtqlIQHrPNshXrlzJY489RuPGjZk3bx5Zs3o2aamUz5XB7Gsi8ilQCfC/8aAx5kG3RaWUlzPGMP2fo7zz6w780/nw3eM1aVEpv6fDAqBhw4Z8/vnnvPDCC/j7+ye8gkrzXBnMngrsAkoA7wCHgH/cGJNSXu3StUhenPovQ+dspUax7CwZ+IDHk4TD4WDkyJEcOXIEX19fBg0apElCucyVFkUuY8wEERng1B21yt2BKeWN/tp/npdnhHA29Dqvty7Psw1L4uPjuQFrgOjoaJ555hl++uknMmTIwGuvvebReJT3cSVRRNn/T4rIQ8AJIPmu4K6UF4iKcfDFb3v4euV+iufKxNwX63Nf4WyeDouIiAi6du3K/PnzGTlyJK+++qqnQ1JeyJVE8Z6IZAMGY50/kRUY6M6glPImh8+H0T8ohM1HL9E1sAhvtatIpgyev9TLlStXaN++PStXrmTs2LG89NJLng5JeakEP83GmAX2zctAE7h5ZnaCRKQV1vQfvsAPxpiP4limMfAFkA44Z4zRuQOUVzDGMOff47w1bxu+PsK4HjV4qEoBT4d1k8PhICwsjKlTp9KjRw9Ph6O8WHwn3PkCXbDmeFpijNkmIm2BN4AAoHp8BdvrjwOaY80R9Y+IzDfG7HBaJjvwNdDKGHNERDx3iqpSd+FKRBTD5m5j/uYT1CqRk9Fdq1Eoe8q4POjx48fJmTMn2bNn56+//sLXV8+PVfcmvhbFBKAIsAEYIyKHgbrAUGPMLy6UXQvYZ4w5ACAiQUB7YIfTMj2AOcaYIwDGmDN3/QqUSmabDl9gQFAIJy9H8EqLsrzQuDS+Hh6wvmH37t00b96chg0bMnXqVE0SKknElygCgSrGGIeI+APngNLGmFMull0IOOp0/xhQO9YyZYF0IrIS6+zvL40xP8UuSET6AH0AihYt6mL1SiWt6BgH41bs58vf91AoRwAzn69LjaI5PB3WTf/++y+tWrVCRHjllVc8HY5KReJLFJHGGAeAMSZCRPbcRZIAiOsnlomj/ppAU6zurL9E5G9jzJ5bVjLmO+A7gMDAwNhlKOV2xy5eY2BQCBsPX6RT9UK8074SWfzTeTqsm1atWkW7du3ImTMny5cvp0yZMp4OSaUi8SWK8iKyxb4tQCn7vgDGGFMlgbKPYXVd3VAY69Da2MucM8aEAWEishqoCuxBqRRi/uYTvDl3K8bAF12r0aF6IU+HdIvr16/z+OOPU7hwYZYtW0bhwnr0ukpa8SWKCvdY9j9AGREpARwHumGNSTibB4wVET+sSQhrY12jWymPu3o9mhHztzNr0zGqF83Ol12rUzRXRk+HdZsMGTKwYMECChYsSO7cuT0djkqF4psU8J4mAjTGRItIX2Ap1uGxE40x20Xkefv5b40xO0VkCbAFcGAdQrvtXupVKimEHL3EgKBgjl64Rv8HS9O/aRn8fF2Z8Sb5fPXVV5w7d4533nmHKlUSauArlXhuPSvIGLMIWBTrsW9j3f8U+NSdcSjlqhiHYfzq/Xy+bA95s2QgqE9dapXI6emwbmGM4Z133uGdd96hY8eOxMTE6NFNyq08f/qoUinEycvhvDx9M38dOM9DVQrwQYf7yJYx5QxYg3US3cCBA/nqq6/o1asX33//vSYJ5XYuJQoRCQCKGmN2uzkepTxiybaTvDZ7K1ExDj55pAqP1izs0avP3Unv3r358ccfefnll/n000/x8UlZ3WEqdUowUYhIO+AzrMHmEiJSDRhpjHnYzbEp5XbXIqN5d8FOpm04QpXC2fiyW3VK5M7k6bDu6MEHH6R06dK8/vrrKTKRqdTJlRbFCKyzrFcCGGNCRKS4+0JSKnlsO36ZAUHBHDgXxvONSvFy87Kk90t5v9AvX77Mv//+S5MmTXjsscc8HY5Kg1xJFNHGmMv660WlFg6HYeLag3y8ZBc5M6Xnf8/Upn7plHlY6ZkzZ2jVqhV79+7l4MGDevir8ghXEsU2EekB+IpIGaA/sM69YSnlHmdCIxg8YzNr9p6jRcV8fNy5Cjkypfd0WHE6fPgwLVq04OjRo8yePVuThPIYVxJFP+BN4DrwM9Z5Ee+5Myil3OGPXacZMnMLYZHRvN+xMj1qFU2x/fw7d+6kRYsWhIaGsnz5curXd2lmf6XcwpVEUc4Y8yZWslDK60RExfDhop1M/uswFQpk5avu1SidN4unw4rXzz//TFRUFKtWraJq1aqeDkelcWJM/HPsicgKoAAwEwgyxmxPjsDuJDAw0GzcuNGTISgvsvtUKP2nBbP7dCjPNCjBq63KkcEv5Z53EBkZSfr06XE4HJw+fZoCBVLOhZCUdxORTcaYwMSsm+AhHsaYJkBj4CzwnYhsFZFhialMqeRijGHyukO0G/sn58MimfTU/QxvWzFFJ4l58+ZRvnx5Dh48iI+PjyYJlWK4dCygMeaUMWYM8DwQArzlzqCUuhfnr16n9+SNvD1/O/VL5WLJwIY0LpeyL544efJkOnfuTN68ecmaNaunw1HqFq6ccFcB6Ao8ApwHgoDBbo5LqURZvecsg2du5nJ4FCPaVeTJesVT7ID1DV988QWDBg2iWbNmzJ07l8yZM3s6JKVu4cpg9o/ANKCFMSb29SSUShGuR8fw6ZLd/PDnQcrkzcxPT9eiQoGU/8t88uTJDBo0iM6dOzN16lQyZMjg6ZCUuk2CicIYUyc5AlEqsfaduUr/acHsOHmFx+sU482HKuCfLuWORTjr1KkTJ0+eZMiQITq5n0qx7jhGISIz7P9bRWSL099WpyvfKeUxxhimbThC26/WcPJyON8/Eci7HSqn+CQRFRXFu+++S1hYGFmyZGHo0KGaJFSKFl+LYoD9v21yBKLU3bgYFsnQOVtYuv00DUrnZlSXquTL6u/psBJ07do1Hn30URYtWkTZsmXp2rWrp0NSKkHxXeHupH3zRWPMa87PicjHwGu3r6WU+63bf46Xp2/mfNh13mhTnt4NSuLjk7IHrAEuXbpEu3btWLt2LePHj9ckobyGK4fHNo/jsdZJHYhSCYmKcfDJkl30/GE9GdP7MvfF+vR5oJRXJInTp0/TuHFj1q9fz/Tp0+nTp4+nQ1LKZXdsUYjIC8CLQMlYYxJZgLXuDkwpZ4fOhTEgKJjNxy7T7f4ivNWuIhnTe88FGq9evcrVq1dZsGABLVq08HQ4St2V+L5pPwOLgQ+BoU6PhxpjLrg1KqVsxhhm/3uct+dtw8/Xh2961qD1fd5zxvKxY8coVKgQpUqVYufOnaRLl7IuraqUK+LrejLGmEPAS0Co0x8ikrKuNq9SpcvhUfQPCuGVmZupXCgbiwc09KoksWHDBqpWrcp771mTLWuSUN4qoRZFW2ATYADnjmADlHRjXCqN++fQBQYGhXDqSgRDWpbj+Ual8PWCsYgbfv/9d9q3b0++fPno2bOnp8NR6p7Ed9RTW/t/ieQLR6V10TEOvvpjH1/9sZfCOTIy6/m6VC+aw9Nh3ZU5c+bQvXt3ypYty7Jly3RyP+X1XJnrqT4QYowJE5HHgBrAF8aYI26PTqUpRy9cY+D0EDYdvkinGoV45+FKZPH3ru6aU6dO0bNnT2rWrMnChQvJkcO7kpxScXHlsJFvgKoiUhV4FZgATAEauTMwlbbM33yCN+dsBeDLbtVoX62QhyNKnPz587No0SJq1apFpkyZPB2OUknClfMooo11daP2wJfGmC+xDpFV6p5dvR7NyzNC6D8tmLL5s7BoQEOvSxLGGIYNG8a0adMAaNKkiSYJlaq4kihCReR14HFgoYj4At7VH6BSpJCjl3hozBp+CT7OgKZlmN6nDkVyZvR0WHclJiaGF154gffff58///zT0+Eo5RaudD11BXoATxtjTolIUeBT94alUrMYh+HbVfsZvXwP+bL6M/25utxf3PuOuI6MjOSJJ55g+vTpDB06lA8++MDTISnlFq5MM35KRKYC94tIW2CDMeYn94emUqMTl8IZND2E9Qcv8FCVAnzQ8T6yBXhfAzUqKor27duzZMkSPvnkE4YMGeLpkJRyG1eOeuqC1YJYiXUuxVciMsQYM8vNsalUZvHWkwyds5WoGAefPlKFR2oWTvFXn7uTdOnSUbNmTTp37kzv3r09HY5SbiXWOHU8C4hsBpobY87Y9/MAvxljqiZDfLcJDAw0Gzdu9ETVKpGuRUbz7oIdTNtwlCqFs/Flt+qUyO2dg70nT57k7NmzVKlSxdOhKHVXRGSTMSYwMeu6MkbhcyNJ2M7j2iC4Umw7fpn+QcEcPBfGC41LMahZWdL7eefH58CBAzRv3hxjDLt379YpOVSa4UqiWCIiS7Gumw3W4PYi94WkUgOHwzDhz4N8snQXuTJlYGrv2tQrldvTYSXatm3baNGiBRERESxevFiThEpTXBnMHiIinYAGWGMU3xlj5ro9MuW1zlyJYPDMzazZe46WlfLxUacq5MiU3tNhJdrff/9NmzZtCAgIYM2aNVSqVMnTISmVrOK7HkUZ4DOgFLAVeMUYczy5AlPe6bcdp3l19hauRUbzQcf76F6riNcOWN8watQocubMyfLlyylRQqc+U2lPfC2KicBPwGqgHfAV0Ck5glLeJyIqhg8W7eSnvw5TsUBWxnSvRum83n0Cf3R0NH5+fkyePJnQ0FDy5cvn6ZCU8oj4EkUWY8z39u3dIvJvcgSkvM+uU1foPy2YPaev0rtBCYa0KkcGP19Ph3VPvv/+e77//nt+++03smbNSsaM3nXGuFJJKb5E4S8i1fnvOhQBzveNMZo40jhjDJPXHeKDxbvI6p+OyU/XolHZPJ4O65598sknvPbaa7Ru3Ro/P++53KpS7hLft+Ak8LnT/VNO9w3woLuCUinfuavXGTJzMyt2n+XB8nn55JEq5M6cwdNh3RNjDK+//joff/wxXbt25aeffiJ9eu8dhFcqqcR34aImyRmI8h6r9pxl8IzNXImI4p2HK/FE3WJeP2ANMHLkSD7++GOef/55xo4di6+vd3efKZVUtF2tXHY9OoZPluxmwp8HKZsvM//rXYvy+bN6Oqwk06tXLwICAhgyZEiqSHxKJRXvPEVWJbt9Z0LpMG4dE/48yJN1izG/b4NUkSTCwsIYNWoUDoeDYsWK8eqrr2qSUCoWbVGoeBlj+HnDEd5dsIOM6f2Y8GQgTSukjsNEL1y4wEMPPcSGDRuoV68edevW9XRISqVIrsweK0BPoKQxZqR9PYr8xpgNbo9OedTFsEhem72FZTtO07BMbkY9WpW8Wf09HVaSOHHiBC1btmTPnj3MmjVLk4RS8XClRfE14MA6ymkkEArMBu53Y1zKw9btO8egGSFcCItk2EMVeLp+CXx8UkeXzP79+2nevDlnz55l0aJFNG3a1NMhKZWiuZIoahtjaohIMIAx5qKI6DGDqVRktIPPl+9h/Or9lMidiQlP3k/lQtk8HVaSOnbsGJGRkfz+++/UqlXL0+EoleK5kiii7OtkG7h5PQqHW6NSHnHwXBgDgoLZcuwy3WsVYXjbimRMn3qGsU6fPk2+fPlo1KgR+/btw98/dXSjKeVurhz1NAaYC+QVkfeBPwG9OHAqYoxh5sajPDRmDYfPX+ObnjX4sFOVVJUkli5dSqlSpZgxYwaAJgml7oIr04xPFZFNQFOs6Ts6GGN2uj0ylSwuh0fxxtytLNxykjolc/J5l2oUzB7g6bCS1PTp03n88cepVKkSjRo18nQ4SnkdV456KgpcA351fswYc8SdgSn3++fQBQYGhXD6SgRDWpbj+Ual8E0lA9Y3jB8/nhdeeIEGDRrw66+/ki1b6hpvUSo5uNK3sBBrfEIAf6AEsBvQq7d4qegYB2P+2MfYP/ZSJGdGZr1Qj2pFsns6rCS3adMmnn/+edq0acPMmTN1BlilEsmVrqf7nO+LSA3gObdFpNzq6IVrDAgK5t8jl+hcozDvtK9E5gypZyzCWc2aNZkzZw5t27bVS5cqdQ/uegoPe3pxPYfCC80LOU6bL9ew9/RVxnSvzqguVVNdkoiOjqZfv36sX78egI4dO2qSUOoeuTJG8bLTXR+gBnDWlcJFpBXwJeAL/GCM+egOy90P/A10NcbMcqVs5brQiCjenredOcHHqVksB190rUaRnKmvG+b69ev06NGDOXPmULhwYWrXru3pkJRKFVz5Oel8PctorDGL2QmtZJ97MQ5oDhwD/hGR+caYHXEs9zGw1NWgleuCj1xkQFAIxy5eY2CzMvRtUho/39Q3F+TVq1fp0KEDv//+O6NHj2bgwIGeDkmpVCPeRGHvxDMbY4YkouxawD5jzAG7rCCgPbAj1nL90ClBklyMw/DNyn2M/m0v+bP6M+O5ugQWz+npsNzi8uXLtGzZko0bNzJ58mSeeOIJT4ekVKpyx0QhIn7GmGh78DoxCgFHne4fA27pCxCRQkBHrHmkNFEkkROXwhk4PYQNBy/QrmpB3utQmWwBqbefPlOmTJQoUYLXX3+d9u3bezocpVKd+FoUG7DGI0JEZD4wEwi78aQxZk4CZcd1QL6Jdf8L4DVjTEx81wAQkT5AH4CiRYsmUG3atmjrSYbO3kKMwzDq0ap0qlEo1V5fYd++fWTKlIkCBQowbdo0T4ejVKrlyhhFTuA81q/+G+dTGCChRHEMKOJ0vzBwItYygUCQvSPLDbQRkWhjzC/OCxljvgO+AwgMDIydbBRwLTKad+bvYPrGo1QtnI0vu1WneO5Mng7LbUJCQmjZsiWVK1fm999/93Q4SqVq8SWKvPYRT9v4L0Hc4MrO+h+gjIiUAI4D3YAezgsYY0rcuC0ik4AFsZOEStjWY5cZEBTMwfNhvNi4FIOalyVdKhywvuHPP/+kbdu2ZMmShXHjxnk6HKVSvfgShS+QGde6kG5fwBrf6It1NJMvMNEYs11Enref/zYR8SonDofhhz8P8OnS3eTKlIGfe9ehbqlcng7LrRYvXkznzp0pUqQIy5cv165IpZKBGBP3Pl9E/jXGJHYg220CAwPNxo0bPR2Gx52+EsHgGZv5c985WlXKz0ed7yN7xtR9mZCYmBgCAwPx8fFh8eLF5M2b19MhKeU1RGSTMSYwMevG16JInSOgqcDyHad5ddZmIqIcfNjpPrrdXyTVDljf4HA48PX1ZdGiRWTMmFEn91MqGcWXKPT6kClMRFQM7y/cyZS/D1OpYFa+7Fad0nkzezostzLG8MEHH7B582amTZtGgQIFPB2SUmnOHUc8jTEXkjMQFb+dJ6/Q7qs/mfL3YZ5tWII5L9ZL9UnC4XAwePBghg0bhr+/Pw6HXlhRKU9IXTPCpULGGCatO8SHi3eRLSAdPz1diwfK5vF0WG4XHR3Ns88+y6RJk+jXrx9ffPEFPj6p90gupVIyTRQp2Lmr13ll5mZW7j5L0/J5+eSRKuTKnMHTYSWLG0lixIgRvPXWW6l+DEaplEwTRQq1cvcZXpm5mSsR0YxsX4nH6xRLUzvLZ599lsDAQF566SVPh6JUmqeJIoW5Hh3Dx4t3M3HtQcrly8LU3nUolz9LwiumAufOnWPBggX06tWLevXqUa9ePU+HpJRCE0WKsvd0KP2DQth58gq96hVnaOvy+Kfz9XRYyeLo0aO0aNGCw4cP07RpU4oUKZLwSkqpZKGJIgUwxjB1/RHeXbCDTBn8mPBkIE0r5PN0WMlmz549NG/enEuXLrFkyRJNEkqlMJooPOxCWCSvzd7C8h2naVgmN6O6VCVvFn9Ph5VsgoODadmyJQArVqygRo0UNxmAUmmeJgoPWrvvHC/PCOFCWCTDHqrA0/VL4OOTdgaswZoFNiAggOXLl1O2bFlPh6OUisMd53pKqVLDXE+R0Q5GLd/Nd6sPUDJ3Jr7sVp3KhdLWlBQXLlwgZ07rintXr14lc+bUffKgUp52L3M96RlMyezA2at0/mYd41cdoNv9Rfm1X4M0lySmTp1K8eLF2bBhA4AmCaVSOO16SibGGGZuPMaIX7eT3s+Hbx+rSavK+T0dVrIbO3Ys/fr1o0mTJlSoUMHT4SilXKCJIhlcvhbFG3O3snDrSeqWzMXnXatSIFuAp8NKVsYY3nvvPd566y3at29PUFAQ/v5pZ9BeKW+micLNNhy8wMCgYM6EXufVVuV47oFS+KaxAWuA2bNn89Zbb/Hkk0/yww8/4OenHz2lvIV+W90kKsbBmN/3Mm7FPormzMjsF+pRtUh2T4flMR07duSnn36iZ8+eOrmfUl5Gv7FucOT8NbqM/4uv/thHpxqFWdC/YZpMEuHh4Tz//PMcPXoUX19fHn/8cU0SSnkhbVEksV+CjzPsl22IwJju1Xm4akFPh+QRV65c4eGHH2b16tU0bNiQnj17ejokpVQiaaJIIqERUbw1bztzg48TWCwHX3SrRuEcGT0dlkecPXuWVq1asWXLFqZOnUr37t09HZJS6h5ookgC/x65yICgYI5fDGdQs7K81KQUfr5ps4vl2LFjNGvWjMOHDzNv3jzatGnj6ZCUUvdIE8U9iHEYvl6xjy9+30v+rP7MeK4ugcVzejosj8qcOTP58+fnhx9+oEGDBp4ORymVBDRRJNLxS+EMCgphw6ELPFy1IO91rExW/3SeDstjtm/fTsmSJcmePTsrVqxIUxdZUiq1S5v9I/do4ZaTtP5iNdtPXObzLlX5slu1NJ0kVq1aRd26dRk0aBCAJgmlUhltUdyFsOvRvPPrdmZsPEbVItkZ060axXJl8nRYHjV//ny6dOlCqVKlGD58uKfDUUq5gSYKF205dokBQSEcOh9G3yalGdCsDOnS6ID1DVOmTOGpp56iZs2aLFq0iFy5cnk6JKWUG2iiSIDDYfhuzQE+W7qbPFkyMO3ZOtQpqTvEy5cvM3jwYBo3bszcuXPJkiVtXNdbqbRIE0U8Tl2OYPDMENbuO0/ryvn5sNN9ZM+Y3tNhedSN65dky5aN1atXU6JECTJkyODhqJRS7qSJ4g6WbT/Fa7O3EBHl4KNO99H1/iJpfpDW4XAwYMAAcuTIwciRIylfvrynQ1JKJYO03ckeh/DIGN6cu5U+UzZRMHsAC/o3oFutomk+SURFRfHEE08wduxYrl27hrddGVEplXjaonCy48QV+gcFs+/MVfo8UJLBLcqSwc/X02F5XHh4OI8++igLFy7kgw8+YOjQoWk+cSqVlmiiwOp3/3HtIT5avItsGdMx5ZlaNCyTx9NhpQjGGNq2bcuKFSv45ptveP755z0dklIqmaX5RHE29DqvzNzMqj1naVYhLx93rkKuzDo4e4OI0KtXL/r06UPXrl09HY5SygPSdKJYsesMQ2ZtJjQimnfbV+KxOsW0S8V2+PBhduzYQevWrXn88cc9HY5SyoPSZKKIiIrh4yW7+HHtIcrnz8LPz9ahbD49D+CGHTt20KJFC6Kjo9m/fz+ZMqXts8+VSuvSXKLYczqU/tOC2XUqlF71ijO0dXn80+mA9Q3//PMPrVu3xs/Pj6VLl2qSUEqlnURhjOF/64/w3oIdZM7gx4+97qdJ+byeDitF+eOPP2jfvj25c+dm+fLllC5d2tMhKaVSgDSRKC6ERfLqrC38tvM0D5TNw2ePViFvFn9Ph5XiLF68mGLFirFs2TIKFkybl3BVSt1OvO3EqcDAQLNx40aXl/9z7zlenhHCpWtRvNa6PE/VK46Pjw5YO7ty5QpZs2bF4XBw9epVsmbN6umQlFJJTEQ2GWMCE7Nuqj0zOzLawYeLdvLYhPVk8fdj7kv1eKZBCU0SsXz++edUrFiRo0eP4uPjo0lCKXWbVNn1dODsVQYEhbD1+GV61C7K8IcqEpBeB6ydGWMYPnw477//Pp07dyZvXh2vUUrFLVUlCmMMMzYeZcT8HWRI58P4x2vSslJ+T4eV4jgcDvr27cs333zDM888w/jx4/H11USqlIpbqkkUl69F8frcLSzaeoq6JXMxums18mfTAeu4fPrpp3zzzTe8+uqrfPTRR3qSoVIqXqkiUaw/cJ5B00M4E3qd11qVp88DJfHVsYg7evHFF8mXLx+9evXydChKKS/g1YPZUTEOPlu6m27f/016Px9mv1CPFxqX0iQRh0uXLtG/f3/CwsLIkiWLJgmllMu8tkVx5Pw1+gcFE3L0Eo/WLMyIhyuRKYPXvhy3On36NC1btmTHjh107NiRJk2aeDokpZQX8do9a/+gYPafvcpX3avTrqqeHHYnhw4donnz5pw4cYIFCxZoklBK3TWvTRSnLkfQqlJ+TRLx2LFjB82bNyc8PJzffvuNunXrejokpZQX8toxivCoGD03IgHp06cnX758rF69WpOEUirRvLZFEREVQ4DO+hqn7du3U7FiRUqXLs2mTZv08Fel1D3xyhaFw2G4Hu0ggyaK28yZM4caNWowevRoAE0SSql75pWJ4nq0A0BbFLFMnDiRRx99lJo1a/LUU095OhylVCrhlYkiPCoGAP90Xhm+W3z22Wc888wzNG/enOXLl5MjRw5Ph6SUSiXcuqcVkVYisltE9onI0Die7ykiW+y/dSJS1ZVyI+xEoS0Ky969e3njjTfo0qUL8+fP16vSKaWSlNsGs0XEFxgHNAeOAf+IyHxjzA6nxQ4CjYwxF0WkNfAdUDuhsm+0KNL6UU/GGESEMmXKsHbtWmrUqKGT+ymlkpw7WxS1gH3GmAPGmEggCGjvvIAxZp0x5qJ992+gsCsF32hRZPBLuzvFyMhIevbsSVBQEAD333+/JgmllFu4M1EUAo463T9mP3YnzwCLXSk4Io23KMLCwnj44YeZNm0aJ06c8HQ4SqlUzp3nUcR1XGac110VkSZYiaLBHZ7vA/QBKFq0KOGR1lFP/n5pbzD74sWLPPTQQ6xfv54ffviBZ555xtMhKaVSOXcmimNAEaf7hYHbfv6KSBXgB6C1MeZ8XAUZY77DGr8gMDDQpNUWRVhYGI0aNWL37t3MnDmTTp06eTokpVQa4M5E8Q9QRkRKAMeBbkAP5wVEpCgwB3jcGLPH1YL/Ozw2bSWKTJky0alTJxo0aECzZs08HY5SKo1wW6IwxkSLSF9gKeALTDTGbBeR5+3nvwXeAnIBX9tnEEcbYwITKjutHR67bds2oqOjqVatGiNGjPB0OEqpNMatcz0ZYxYBi2I99q3T7d5A77stNyINtSj+/vtv2rRpQ/HixXXeJqWUR3jlaHBElD2YncrPzF6+fDlNmzYlV65czJ49W5OEUsojvHJPmxbGKGbNmsVDDz1EmTJl+PPPPylRooSnQ1JKpVFemSgiomLw8xHS+Xpl+AkyxjBp0iRq1arFypUryZcvn6dDUkqlYV55PYrwVHwtivDwcAICApgxYwYAGTNm9HBESqm0zit/kkdExaS6a1EYYxg6dCgNGzbk6tWrZMyYUZOEUipF8NJE4SAgvVeGHqeYmBj69OnDxx9/TK1atQgICPB0SEopdZNX7m3DI1NP19P169fp1q0bP/zwA8OGDWPcuHE6uZ9SKkXxyjGKiOiYVHPEU//+/Zk1axajR49m4MCBng5HKaVu45WJIjwy9SSKN954g8aNG9O9e3dPh6KUUnHyyq6niGiHVyeKEydOMHz4cBwOB8WKFdMkoZRK0bwzUUTGEOClZ2Xv37+fBg0a8MUXX7Bnj8vzICqllMd45d7WW8+j2LJlCw0aNODKlSv88ccflC9f3tMhKaVUgrwyUUREed8Yxbp162jUqBG+vr6sWbOG+++/39MhKaWUS7wyUYR7YaK4fv06RYoUYe3atVSoUMHT4SillMu8MlFcj/Kewex9+/YB0KRJE4KDgylWrJiHI1JKqbvjlYkiMsbhFWMU48ePp1y5cixYsABAT6RTSnklr0sUDmOAlH0tCmMMH374Ic8//zytW7fmwQcf9HRISimVaCl3b3sHDitPEJA+Zf46N8YwZMgQ3njjDXr06MHcuXN1cj+llFfzwkRxo0WRMhPF77//zqhRo+jbty9TpkwhXbp0ng5JKaXuiddN4WHniRSbKJo1a8aKFSto1KiRXrpUKZUqeF+Lwu57SkmD2aGhoXTo0IENGzYA0LhxY00SSqlUw/sSBSlrMPv8+fM0bdqUBQsWsH//fk+Ho5RSSc77up4c1v+U0KI4fvw4LVq0YP/+/cyZM4eHH37Y0yEppVSS87pEkVIGs48fP079+vW5cOECS5YsoXHjxh6NRyml3CVl9N/cBUcKGczOly8fTZs2ZcWKFZoklFKpmte1KIzdovDUeRR//fUXJUqUIH/+/EyYMMEjMSilVHLy3haFX/KHvmjRIpo2bUq/fv2SvW6llPIUL0wUnmlRTJs2jfbt21OhQgW+/vrrZK1bKaU8yWsThb9f8iWKr7/+mp49e1K/fn1WrFhBnjx5kq1upZTyNK9LFMZAej8ffHyS54S2iIgIxo0bR9u2bVm8eDFZs2ZNlnqVUiql8LrBbIcxyTI+4XA4iImJwd/fn5UrV5I9e3adt0kplSZ5XYvCYdw/PhEdHc3TTz9Njx49cDgc5MmTR5OEUirN8rpEYYxx61nZERERPPLII0yePJkqVaronE1KqTTPC7ue3Hey3ZUrV+jQoQMrVqzgq6++om/fvm6pRymlvIkXJgrjtkTxyCOPsHr1av73v//Rs2dPt9Sh0o6oqCiOHTtGRESEp0NRaYi/vz+FCxdO0u5yr0sUxrhv5tgRI0bQr18/2rVr55byVdpy7NgxsmTJQvHixbULUyULYwznz5/n2LFjlChRIsnK9boxCkcSj1Hs2bPn5gl09erV0yShkkxERAS5cuXSJKGSjYiQK1euJG/Fel2LwmFMkh319O+//9KqVStEhG7dupEzZ84kKVepGzRJqOTmjs+cF7Yokuas7FWrVtG4cWMCAgJYs2aNJgmllLoDr0sUxhj877FF8euvv9KqVSsKFSrE2rVrKVu2bBJFp1TK4uvrS7Vq1ahcuTLt2rXj0qVLN5/bvn07Dz74IGXLlqVMmTK8++67N2dnBli8eDGBgYFUqFCB8uXL88orr3jgFcQvODiY3r17ezqMO7p+/Tpdu3aldOnS1K5dm0OHDsW5XGRkJH369KFs2bKUL1+e2bNnA7B69Wpq1KiBn58fs2bNurn82bNnadWqVXK8BMALE0VStChOnTpF5cqVWbNmDYULF06iyJRKeQICAggJCWHbtm3kzJmTcePGARAeHs7DDz/M0KFD2bNnD5s3b2bdunU3x+u2bdtG3759+d///sfOnTvZtm0bJUuWTNLYoqOj77mMDz744K5mc06KOu/GhAkTyJEjB/v27WPQoEG89tprcS73/vvvkzdvXvbs2cOOHTto1KgRAEWLFmXSpEn06NHjluXz5MlDgQIFWLt2rdtfA3jhGIUxhoD0ictvR44coWjRojz77LP06tVLz7ZWyeadX7ez48SVJC2zYsGsvN2uksvL161bly1btgDw888/U79+fVq0aAFAxowZGTt2LI0bN+all17ik08+4c0336R8+fIA+Pn58eKLL95W5tWrV+nXrx8bN25ERHj77bfp3LkzmTNn5urVqwDMmjWLBQsWMGnSJHr16kXOnDkJDg6mWrVqzJ07l5CQELJnzw5A6dKlWbt2LT4+Pjz//PMcOXIEgC+++IL69evfUndoaChbtmyhatWqAGzYsIGBAwcSHh5OQEAAP/74I+XKlWPSpEksXLiQiIgIwsLC+PXXX+nXrx9bt24lOjqaESNG0L59ew4dOsTjjz9OWFgYAGPHjqVevXoub9+4zJs3jxEjRgDW4fd9+/bFGHPbOMLEiRPZtWsXAD4+PuTOnRuA4sWL33wstg4dOjB16tTbtos7eF+i4O5bFMYYRo4cyccff8z69eu57777NEmoNCUmJobff/+dZ555BrC6nWrWrHnLMqVKleLq1atcuXKFbdu2MXjw4ATLfffdd8mWLRtbt24F4OLFiwmus2fPHn777Td8fX1xOBzMnTuXp556ivXr11O8eHHy5ctHjx49GDRoEA0aNODIkSO0bNmSnTt33lLOxo0bqVy58s375cuXZ/Xq1fj5+fHbb7/xxhtv3OzC+euvv9iyZQs5c+bkjTfe4MEHH2TixIlcunSJWrVq0axZM/Lmzcvy5cvx9/dn7969dO/enY0bN94Wf8OGDQkNDb3t8c8++4xmzZrd8tjx48cpUqQIYCXbbNmycf78+ZuJALjZHTh8+HBWrlxJqVKlGDt2LPny5Yt3OwYGBjJs2LB4l0kqXpco4O7menI4HAwaNIgxY8bQq1cvKlSo4MbIlIrb3fzyT0rh4eFUq1aNQ4cOUbNmTZo3bw4Q56/aG+7mqJnffvuNoKCgm/dz5MiR4DqPPvoovr7Wd7hr166MHDmSp556iqCgILp27Xqz3B07dtxc58qVK4SGhpIlS5abj508efKWKf8vX77Mk08+yd69exERoqKibj7XvHnzmwesLFu2jPnz5/PZZ58B1mHMR44coWDBgvTt25eQkBB8fX3Zs2dPnPGvWbMmwdd4g/OYzw2xt290dDTHjh2jfv36fP7553z++ee88sorTJkyJd6y8+bNy4kTJ1yO5V54ZaJw9czsqKgonnnmGaZMmcKgQYP47LPP4mzCKZVa3RijuHz5Mm3btmXcuHH079+fSpUqsXr16luWPXDgAJkzZyZLlixUqlSJTZs23ezWuZM7JRznx2If058pU6abt+vWrcu+ffs4e/Ysv/zyy81fyA6Hg7/++ouAgIB4X5tz2cOHD6dJkybMnTuXQ4cO3XIte+c6jTHMnj2bcuXK3VLeiBEjyJcvH5s3b8bhcODv7x9nvXfToihcuDBHjx6lcOHCREdHc/ny5duOsMyVKxcZM2akY8eOgJVIXbnMckRERLzbJyl55V7T1UQxYcIEpkyZwnvvvceoUaM0Sag0K1u2bIwZM4bPPvuMqKgoevbsyZ9//slvv/0GWC2P/v378+qrrwIwZMgQPvjgg5u/qh0OB59//vlt5bZo0YKxY8fevH+j6ylfvnzs3LnzZtfSnYgIHTt25OWXX6ZChQrkypUrznJDQkJuW7dChQrs27fv5v3Lly9TqFAhACZNmnTHOlu2bMlXX31189d+cHDwzfULFCiAj48PU6ZMISYmJs7116xZQ0hIyG1/sZMEwMMPP8zkyZMBa6zmwQcfvC2xigjt2rVj5cqVAPz+++9UrFjxjvHfsGfPnlu63tzKGONVf+nzlzYLNp8wroiOjjaLFy92aVmlktqOHTs8HYLJlCnTLffbtm1rfvrpJ2OMMVu2bDGNGjUyZcuWNaVKlTIjRowwDofj5rK//vqrqVGjhilfvrypUKGCeeWVV24rPzQ01DzxxBOmUqVKpkqVKmb27NnGGGNmzpxpSpYsaRo1amReeukl8+STTxpjjHnyySfNzJkzbynjn3/+MYCZNGnSzcfOnj1runTpYu677z5ToUIF89xzz8X5+ipXrmyuXLlijDFm3bp1pkyZMqZevXpm2LBhplixYsYYY3788Ufz0ksv3Vzn2rVrpk+fPqZy5cqmUqVK5qGHHjLGGLNnzx5z3333mdq1a5uhQ4fetu0SIzw83DzyyCOmVKlS5v777zf79++/+VzVqlVv3j506JBp2LChue+++8yDDz5oDh8+bIwxZsOGDaZQoUImY8aMJmfOnKZixYo31/n000/NmDFj4qw3rs8esNEkcr8rJo4+tJQsQ4EyZtEff9K0QtwDPWfOnOG5555jzJgxNweRlPKEnTt36piYm40ePZosWbKk6HMp3OWBBx5g3rx5cY4LxfXZE5FNxpjAxNTllX0xd5rr6ciRIzRs2JClS5eyd+/eZI5KKZXcXnjhBTJkyODpMJLd2bNnefnll106eCApeOVgdoY4EsWuXbto3rw5oaGhLFu2jAYNGnggMqVUcvL39+fxxx/3dBjJLk+ePHTo0CHZ6vPKRBG7RbFt2zaaNGmCr68vq1atSvBIDaWSi4nnMFSl3MEdwwne2fUU6zyKwoULU79+fdasWaNJQqUY/v7+nD9/3i1fXKXiYuzrUdzp0N7E8srB7EO7tlAgWwCrVq2iVq1ayXYssVJ3Q69wpzzhTle4u5fBbLd2PYlIK+BLwBf4wRjzUaznxX6+DXAN6GWM+TehcgPS+fLTTz/x9NNPM2TIED788EM3RK/UvUmXLl2SXmVMKU9xW9eTiPgC44DWQEWgu4jEPoukNVDG/usDfONK2T98O44nn3ySxo0b8+abbyZh1EoppWJz5xhFLWCfMeaAMSYSCALax1qmPfCTfT7I30B2ESkQX6ExoRcYMvhlOnXqxMKFC8mcObN7oldKKQW4N1EUAo463T9mP3a3y9wi5tolnn76aaZPn54mj59WSqnk5s4xiriOCYw9cu7KMohIH6yuKYDrEydO3DZx4sR7DC9VyA2c83QQKYRui//otviPbov/lEt4kbi5M1EcA5zn0CgMxJ4T15VlMMZ8B3wHICIbEztyn9rotviPbov/6Lb4j26L/4jI7RfXcJE7u57+AcqISAkRSQ90A+bHWmY+8IRY6gCXjTEn3RiTUkqpu+S2FoUxJlpE+gJLsQ6PnWiM2S4iz9vPfwsswjo0dh/W4bFPuSsepZRSiePW8yiMMYuwkoHzY9863TbAS3dZ7HdJEFpqodviP7ot/qPb4j+6Lf6T6G3hdWdmK6WUSl5eOdeTUkqp5JNiE4WItBKR3SKyT0SGxvG8iMgY+/ktIlLDE3EmBxe2RU97G2wRkXUikmpnRkxoWzgtd7+IxIjII8kZX3JyZVuISGMRCRGR7SKyKrljTC4ufEeyicivIrLZ3hapcjxURCaKyBkR2XaH5xO330zspfHc+Yc1+L0fKAmkBzYDFWMt0wZYjHUuRh1gvafj9uC2qAfksG+3Tsvbwmm5P7DGxx7xdNwe/FxkB3YARe37eT0dtwe3xRvAx/btPMAFIL2nY3fDtngAqAFsu8PzidpvptQWhVum//BSCW4LY8w6Y8xF++7fWOejpEaufC4A+gGzgTPJGVwyc2Vb9ADmGGOOABhjUuv2cGVbGCCLPRFpZqxEEZ28YbqfMWY11mu7k0TtN1NqonDL9B9e6m5f5zNYvxhSowS3hYgUAjoC35K6ufK5KAvkEJGVIrJJRJ5ItuiSlyvbYixQAeuE3q3AAGOMI3nCS1EStd9MqVe4S7LpP1IBl1+niDTBShSp9TqwrmyLL4DXjDExqfzKcq5sCz+gJtAUCAD+EpG/jTF73B1cMnNlW7QEQoAHgVLAchFZY4y54ubYUppE7TdTaqJIsuk/UgGXXqeIVAF+AFobY84nU2zJzZVtEQgE2UkiN9BGRKKNMb8kS4TJx9XvyDljTBgQJiKrgapAaksUrmyLp4CPjNVRv09EDgLlgQ3JE2KKkaj9ZkrtetLpP/6T4LYQkaLAHODxVPhr0VmC28IYU8IYU9wYUxyYBbyYCpMEuPYdmQc0FBE/EckI1AZ2JnOcycGVbXEEq2WFiOTDmiDvQLJGmTIkar+ZIlsURqf/uMnFbfEWkAv42v4lHW1S4URoLm6LNMGVbWGM2SkiS4AtgAPrKpNxHjbpzVz8XLwLTBKRrVjdL68ZY1LdrLIiMg1oDOQWkWPA20A6uLf9pp6ZrZRSKl4ptetJKaVUCqGJQimlVLw0USillIqXJgqllFLx0kShlFIqXpooUgh7ptMQp7/i8Sx7NQnqmyQiB+26/hWRuoko4wcRqWjffiPWc+vuNUa7nBvbZZs9+2f2BJavJiJtkqLuu4itoH3/fRE5mpj3R0TG2WXtEJFwp89Bks1+KyK9RMRhn5x547Ft8X3WElnPLe+BiDwc14yuiSi3l4ictbfLLhEZ5OI6BV1Y7lMROSUir9xrnKmRJoqUI9wYU83p71Ay1DnEGFMNGAqMv9uVjTG9jTE77LtvxHqu3r2HB/y3XSpjTXaW0BURq2EdJ56kRMQ3nthunNn6K9YEdXfNGPOS/V60AfY7fQ5mxVN/YhwD3kyisu6kGk7vgTFmvjHmoyQqe7q9neoDb4pIkQSW7wUkmCiMMUNI/fODJZomihRKRDKLyO/2r/2tInLbLKkiUkBEVjv94m5oP95CRP6y150pIpkTqG41UNpe92W7rG0iMtB+LJOILBRrLv9tItLVfnyliASKyEdAgB3HVPu5q/b/6bF+XU4Skc4i4mv/ivtHrHnxn3Nhs/yFPYGZiNQS69obwfb/cvZZuSOBrnYsXe3YJ9r1BN9hO4odyzZ7W994fY1FZIWI/Iw1kVy8jDF/J+XsALHrF5Hi4nSdARF5RURG2LdLicgSsSb/WyMi5e9Q7AKgkoiUi6O+OD83ItLG/gX/p1jXMlhgP+7qe9BLRMaKdU2IQyLiY6+fUawWWLq7iB8Ae5qafUABu6y37Pd4m4h8Z7+nj2BN6TLVjiVARGqKyCq7nqWSOmecTnrJOVe6/sU7j3wM1qRlIcBcrLPms9rP5cb6Utw4QfKq/X8w8KZ92xfIYi+7GshkP/4a8FYc9U3CvlYD8CiwHmsCua1AJqypmLcD1YHOwPdO62az/68EAp1jclrmRowdgcn27fRYM1cGAH2AYfbjGYCNQIk44rzq9PpmAq3s+1kBP/t2M2C2fbsXMNZp/Q+Ax+zb2bHmOcoUq47OwHK7jnxY0z0UwDrDNSyuuOJ6zQk97uLnoDj2tQRi1+/8nH3/FWCEfft3oIx9uzbwRxxl98KaRfUJp/dkm11unJ8bwN9+z27EMA1YcJfvwc37WNOKNLFvd8U6W/yu4rdvF8X6rvjb93M6LTcFaBfHZzQdsA7I41T/RKf1RgCveHpfkBL/UuQUHmlUuLGa1ACISDrgAxF5AGv6hUJYO7FTTuv8A0y0l/3FGBMiIo2AisBasabzSI/1Szwun4rIMOAs1qyzTYG5xppEDhGZAzQElgCficjHWDuJNXfxuhYDY0QkA9AKWG2MCReRFkAV+a8PPhtQBjgYa/0AEQnB2pltwtqh31h+soiUwZr9Mt0d6m8BPCz/9T37Y+1knOc8agBMM8bEAKfFuhLc/cAVYIMxJnZMySnB+u1f/vWAmfLfjLkZ4lnlZ6xumxJOj9Uh7s9NeeCAUwzTsJI8uP4eOJuOtYNegTUn09d3GX9XsWZJLgc8a4yJsB9vIiKvAhmBnFg/cn6NtW45oDLWzLFg/TBIjfPDJTlNFClXT6wrcdU0xkSJyCGsndxNxpjVdiJ5CJgiIp8CF4HlxpjuLtQxxNh94AAi0iyuhYwxe0SkJla/84cisswYM9KVF2GMiRCRlVjTPHfF2tGANd9OP2PM0gSKCDfGVBORbFjdJi8BY7Dm7llhjOko1mDsyjusL0BnY8zueOqIbz7ysATiuysishQr4W80xvR2YRXn+qO5tbv4xufBB7jk/EMjPsaaG2kUVqvhZmjE8bkRkerxFOXqe+BsPtZnKCdWC/YPrBasq/FPN8b0Fevgi4Uishi4BHyN1XI4anfH+cexrgDbjTF3feBGWqdjFClXNuCMnSSaAMViLyAixexlvgcmYF0C8W+gvojcGHPIKCJlXaxzNdDBXicTVrfRGrGOGrlmjPkf8JldT2xRdssmLkFYk481xJq4Dfv/CzfWEZGydp1xMsZcBvoDr9jrZAOO20/3clo0FKsL7oalQD+xf0LeYce3GuuXqq+I5MG6nKRbpp82xrQ01iC1K0kittNAXhHJZbfQ2tplXgEOisijcHPMJaHrpk/C6i7KY9+/0+dmF1BS/jsyqqtTGa6+BzcZY65ibdsvsVqnMYmJ3xjzF1YX0wD+Swrn7NaJ85FizrHsBvLYSQZ7bKRSfPUoiyaKlGsqECgiG7FaF7viWKYxECIiwVj97F8aY85ifWmnicgWrB1AvAODNxhj/sXagWzAGrP4wRgTDNwHbLC7gN4E3otj9e+ALWIPZseyDGvn+5uxLlUJ1rUzdgD/ijVAO54EWrh2LJuxuiw+wfpluharC+GGFUDFGwOpWL9609mxbbPvxzYXa4bVzVi/cF81xpyKY7l4icgnYs3YmVFEjtm/bJOMMSYKa6B4PVbryvkz0RN4RkQ2Y3W7xHWJWOeyIrFaZnnt+3F+bowx4cCLwBIR+RMrWV22i3H1PYhtOvCY/T9R8ds+xvoBEgN8jzW+9gtWl+wNk4Bv7c+uL1YS+diuJwSry0slQGePVSqRROSqMSahI8q8nohkNsZctVtl44C9xpjRno4rqdmJ/aox5jNPx5LSaItCqcS7Ik4n3KViz9q/yLdjdTfd9Tk3KZ09vvcYSTwmlVpoi0IppVS8tEWhlFIqXpoolFJKxUsThVJKqXhpolBKKRUvTRRKKaXipYlCKaVUvP4PRmV9sJ8iRZsAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "(array([0.        , 0.00806156, 1.        ]),\n",
       " array([0.        , 0.23255814, 1.        ]),\n",
       " array([2, 1, 0], dtype=int64))"
      ]
     },
     "execution_count": 86,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "draw_roc(y_pred_final.churn, y_pred_final.predicted)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'0.89'"
      ]
     },
     "execution_count": 87,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#draw_roc(y_pred_final.Churn, y_pred_final.predicted)\n",
    "\"{:2.2f}\".format(metrics.roc_auc_score(y_pred_final.churn, y_pred_final.Churn_Prob))"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}