ML Part 3.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "250dd1c1",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Importing the required libraries.\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import seaborn as sns\n",
    "\n",
    "from imblearn.over_sampling import SMOTE\n",
    "import lightgbm as lightgbm\n",
    "import xgboost as xgboost\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.tree import DecisionTreeClassifier\n",
    "from sklearn.ensemble import StackingClassifier\n",
    "from sklearn.decomposition import PCA\n",
    "\n",
    "from sklearn.metrics import roc_auc_score\n",
    "from sklearn import metrics\n",
    "from sklearn.model_selection import KFold\n",
    "from sklearn.model_selection import cross_val_score\n",
    "\n",
    "from collections import defaultdict\n",
    "import matplotlib.pyplot as plt\n",
    "from scipy.stats import spearmanr\n",
    "from scipy.cluster import hierarchy\n",
    "from scipy.spatial.distance import squareform\n",
    "from sklearn.inspection import permutation_importance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "e0906a7d",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Dataset1.\n",
    "customer= pd.read_csv('teleco.csv') "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "f9b3b715",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(100000, 100)"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "customer.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "54cffc06",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "rev_Mean        357\n",
       "mou_Mean        357\n",
       "totmrc_Mean     357\n",
       "da_Mean         357\n",
       "ovrmou_Mean     357\n",
       "               ... \n",
       "kid11_15       1732\n",
       "kid16_17       1732\n",
       "creditcd       1732\n",
       "eqpdays           1\n",
       "Customer_ID       0\n",
       "Length: 100, dtype: int64"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "customer.isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "e4cb992c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                  percent_missing\n",
      "rev_Mean                    0.357\n",
      "mou_Mean                    0.357\n",
      "totmrc_Mean                 0.357\n",
      "da_Mean                     0.357\n",
      "ovrmou_Mean                 0.357\n",
      "ovrrev_Mean                 0.357\n",
      "vceovr_Mean                 0.357\n",
      "datovr_Mean                 0.357\n",
      "roam_Mean                   0.357\n",
      "change_mou                  0.891\n",
      "change_rev                  0.891\n",
      "drop_vce_Mean               0.000\n",
      "drop_dat_Mean               0.000\n",
      "blck_vce_Mean               0.000\n",
      "blck_dat_Mean               0.000\n",
      "unan_vce_Mean               0.000\n",
      "unan_dat_Mean               0.000\n",
      "plcd_vce_Mean               0.000\n",
      "plcd_dat_Mean               0.000\n",
      "recv_vce_Mean               0.000\n",
      "recv_sms_Mean               0.000\n",
      "comp_vce_Mean               0.000\n",
      "comp_dat_Mean               0.000\n",
      "custcare_Mean               0.000\n",
      "ccrndmou_Mean               0.000\n",
      "cc_mou_Mean                 0.000\n",
      "inonemin_Mean               0.000\n",
      "threeway_Mean               0.000\n",
      "mou_cvce_Mean               0.000\n",
      "mou_cdat_Mean               0.000\n",
      "mou_rvce_Mean               0.000\n",
      "owylis_vce_Mean             0.000\n",
      "mouowylisv_Mean             0.000\n",
      "iwylis_vce_Mean             0.000\n",
      "mouiwylisv_Mean             0.000\n",
      "peak_vce_Mean               0.000\n",
      "peak_dat_Mean               0.000\n",
      "mou_peav_Mean               0.000\n",
      "mou_pead_Mean               0.000\n",
      "opk_vce_Mean                0.000\n",
      "opk_dat_Mean                0.000\n",
      "mou_opkv_Mean               0.000\n",
      "mou_opkd_Mean               0.000\n",
      "drop_blk_Mean               0.000\n",
      "attempt_Mean                0.000\n",
      "complete_Mean               0.000\n",
      "callfwdv_Mean               0.000\n",
      "callwait_Mean               0.000\n",
      "churn                       0.000\n",
      "months                      0.000\n",
      "uniqsubs                    0.000\n",
      "actvsubs                    0.000\n",
      "new_cell                    0.000\n",
      "crclscod                    0.000\n",
      "asl_flag                    0.000\n",
      "totcalls                    0.000\n",
      "totmou                      0.000\n",
      "totrev                      0.000\n",
      "adjrev                      0.000\n",
      "adjmou                      0.000\n",
      "adjqty                      0.000\n",
      "avgrev                      0.000\n",
      "avgmou                      0.000\n",
      "avgqty                      0.000\n",
      "avg3mou                     0.000\n",
      "avg3qty                     0.000\n",
      "avg3rev                     0.000\n",
      "avg6mou                     2.839\n",
      "avg6qty                     2.839\n",
      "avg6rev                     2.839\n",
      "prizm_social_one            7.388\n",
      "area                        0.040\n",
      "dualband                    0.001\n",
      "refurb_new                  0.001\n",
      "hnd_price                   0.847\n",
      "phones                      0.001\n",
      "models                      0.001\n",
      "hnd_webcap                 10.189\n",
      "truck                       1.732\n",
      "rv                          1.732\n",
      "ownrent                    33.706\n",
      "lor                        30.190\n",
      "dwlltype                   31.909\n",
      "marital                     1.732\n",
      "adults                     23.019\n",
      "infobase                   22.079\n",
      "income                     25.436\n",
      "numbcars                   49.366\n",
      "HHstatin                   37.923\n",
      "dwllsize                   38.308\n",
      "forgntvl                    1.732\n",
      "ethnic                      1.732\n",
      "kid0_2                      1.732\n",
      "kid3_5                      1.732\n",
      "kid6_10                     1.732\n",
      "kid11_15                    1.732\n",
      "kid16_17                    1.732\n",
      "creditcd                    1.732\n",
      "eqpdays                     0.001\n",
      "Customer_ID                 0.000\n"
     ]
    }
   ],
   "source": [
    "percent_missing = customer.isnull().sum() * 100 / len(customer)\n",
    "missing_value_df = pd.DataFrame({'percent_missing': percent_missing})\n",
    "print(missing_value_df.to_string())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "03fa2760",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/var/folders/j5/9fvnh0857k3_lqjpjrk029hw0000gn/T/ipykernel_749/2349601179.py:1: FutureWarning: Dropping of nuisance columns in DataFrame reductions (with 'numeric_only=None') is deprecated; in a future version this will raise TypeError.  Select only valid columns before calling the reduction.\n",
      "  customer.fillna(customer.mean(), inplace=True)\n"
     ]
    }
   ],
   "source": [
    "customer.fillna(customer.mean(), inplace=True)\n",
    "for column in customer.columns:\n",
    "    if customer[column].dtype == 'object':\n",
    "        customer[column].fillna(customer[column].mode()[0], inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "48c69eb0",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                  percent_missing\n",
      "rev_Mean                      0.0\n",
      "mou_Mean                      0.0\n",
      "totmrc_Mean                   0.0\n",
      "da_Mean                       0.0\n",
      "ovrmou_Mean                   0.0\n",
      "ovrrev_Mean                   0.0\n",
      "vceovr_Mean                   0.0\n",
      "datovr_Mean                   0.0\n",
      "roam_Mean                     0.0\n",
      "change_mou                    0.0\n",
      "change_rev                    0.0\n",
      "drop_vce_Mean                 0.0\n",
      "drop_dat_Mean                 0.0\n",
      "blck_vce_Mean                 0.0\n",
      "blck_dat_Mean                 0.0\n",
      "unan_vce_Mean                 0.0\n",
      "unan_dat_Mean                 0.0\n",
      "plcd_vce_Mean                 0.0\n",
      "plcd_dat_Mean                 0.0\n",
      "recv_vce_Mean                 0.0\n",
      "recv_sms_Mean                 0.0\n",
      "comp_vce_Mean                 0.0\n",
      "comp_dat_Mean                 0.0\n",
      "custcare_Mean                 0.0\n",
      "ccrndmou_Mean                 0.0\n",
      "cc_mou_Mean                   0.0\n",
      "inonemin_Mean                 0.0\n",
      "threeway_Mean                 0.0\n",
      "mou_cvce_Mean                 0.0\n",
      "mou_cdat_Mean                 0.0\n",
      "mou_rvce_Mean                 0.0\n",
      "owylis_vce_Mean               0.0\n",
      "mouowylisv_Mean               0.0\n",
      "iwylis_vce_Mean               0.0\n",
      "mouiwylisv_Mean               0.0\n",
      "peak_vce_Mean                 0.0\n",
      "peak_dat_Mean                 0.0\n",
      "mou_peav_Mean                 0.0\n",
      "mou_pead_Mean                 0.0\n",
      "opk_vce_Mean                  0.0\n",
      "opk_dat_Mean                  0.0\n",
      "mou_opkv_Mean                 0.0\n",
      "mou_opkd_Mean                 0.0\n",
      "drop_blk_Mean                 0.0\n",
      "attempt_Mean                  0.0\n",
      "complete_Mean                 0.0\n",
      "callfwdv_Mean                 0.0\n",
      "callwait_Mean                 0.0\n",
      "churn                         0.0\n",
      "months                        0.0\n",
      "uniqsubs                      0.0\n",
      "actvsubs                      0.0\n",
      "new_cell                      0.0\n",
      "crclscod                      0.0\n",
      "asl_flag                      0.0\n",
      "totcalls                      0.0\n",
      "totmou                        0.0\n",
      "totrev                        0.0\n",
      "adjrev                        0.0\n",
      "adjmou                        0.0\n",
      "adjqty                        0.0\n",
      "avgrev                        0.0\n",
      "avgmou                        0.0\n",
      "avgqty                        0.0\n",
      "avg3mou                       0.0\n",
      "avg3qty                       0.0\n",
      "avg3rev                       0.0\n",
      "avg6mou                       0.0\n",
      "avg6qty                       0.0\n",
      "avg6rev                       0.0\n",
      "prizm_social_one              0.0\n",
      "area                          0.0\n",
      "dualband                      0.0\n",
      "refurb_new                    0.0\n",
      "hnd_price                     0.0\n",
      "phones                        0.0\n",
      "models                        0.0\n",
      "hnd_webcap                    0.0\n",
      "truck                         0.0\n",
      "rv                            0.0\n",
      "ownrent                       0.0\n",
      "lor                           0.0\n",
      "dwlltype                      0.0\n",
      "marital                       0.0\n",
      "adults                        0.0\n",
      "infobase                      0.0\n",
      "income                        0.0\n",
      "numbcars                      0.0\n",
      "HHstatin                      0.0\n",
      "dwllsize                      0.0\n",
      "forgntvl                      0.0\n",
      "ethnic                        0.0\n",
      "kid0_2                        0.0\n",
      "kid3_5                        0.0\n",
      "kid6_10                       0.0\n",
      "kid11_15                      0.0\n",
      "kid16_17                      0.0\n",
      "creditcd                      0.0\n",
      "eqpdays                       0.0\n",
      "Customer_ID                   0.0\n"
     ]
    }
   ],
   "source": [
    "percent_missing = customer.isnull().sum() * 100 / len(customer)\n",
    "missing_value_df = pd.DataFrame({'percent_missing': percent_missing})\n",
    "print(missing_value_df.to_string())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "7f4b0ae5",
   "metadata": {},
   "outputs": [],
   "source": [
    "train_label=customer['churn']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "7b124b25",
   "metadata": {},
   "outputs": [],
   "source": [
    "train_label_v=train_label"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "0500bac4",
   "metadata": {},
   "outputs": [],
   "source": [
    "del customer['churn']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "3d25c8f9",
   "metadata": {},
   "outputs": [],
   "source": [
    "cat_customer= customer.select_dtypes(include=['object']).columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "f6c67b9b",
   "metadata": {},
   "outputs": [],
   "source": [
    "customer_cat=customer[cat_customer]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "26581431",
   "metadata": {},
   "outputs": [],
   "source": [
    "num_customer=customer.select_dtypes(include=['float64', 'int64']).columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "c8acb9ef",
   "metadata": {},
   "outputs": [],
   "source": [
    "customer_num=customer[num_customer]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "51b3f250",
   "metadata": {},
   "outputs": [],
   "source": [
    "del customer_num['change_rev']\n",
    "del customer_num['change_mou']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "868a9d18",
   "metadata": {},
   "outputs": [],
   "source": [
    "del customer_num['Customer_ID']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "78e9846f",
   "metadata": {},
   "outputs": [],
   "source": [
    "z=customer_cat\n",
    "#label encoder\n",
    "from sklearn.preprocessing import LabelEncoder\n",
    "label=LabelEncoder()\n",
    "z=z.astype(str)\n",
    "z[cat_customer] = z[cat_customer].apply(lambda col: label.fit_transform(col))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "ded1acf5",
   "metadata": {},
   "outputs": [],
   "source": [
    "frames = [customer_num, z]\n",
    "customer_updated=pd.concat(frames,axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "ba0d57ba",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "rev_Mean       0\n",
       "mou_Mean       0\n",
       "totmrc_Mean    0\n",
       "da_Mean        0\n",
       "ovrmou_Mean    0\n",
       "              ..\n",
       "kid3_5         0\n",
       "kid6_10        0\n",
       "kid11_15       0\n",
       "kid16_17       0\n",
       "creditcd       0\n",
       "Length: 96, dtype: int64"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "customer_updated.isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "d2831df7",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                  percent_missing\n",
      "rev_Mean                      0.0\n",
      "mou_Mean                      0.0\n",
      "totmrc_Mean                   0.0\n",
      "da_Mean                       0.0\n",
      "ovrmou_Mean                   0.0\n",
      "ovrrev_Mean                   0.0\n",
      "vceovr_Mean                   0.0\n",
      "datovr_Mean                   0.0\n",
      "roam_Mean                     0.0\n",
      "drop_vce_Mean                 0.0\n",
      "drop_dat_Mean                 0.0\n",
      "blck_vce_Mean                 0.0\n",
      "blck_dat_Mean                 0.0\n",
      "unan_vce_Mean                 0.0\n",
      "unan_dat_Mean                 0.0\n",
      "plcd_vce_Mean                 0.0\n",
      "plcd_dat_Mean                 0.0\n",
      "recv_vce_Mean                 0.0\n",
      "recv_sms_Mean                 0.0\n",
      "comp_vce_Mean                 0.0\n",
      "comp_dat_Mean                 0.0\n",
      "custcare_Mean                 0.0\n",
      "ccrndmou_Mean                 0.0\n",
      "cc_mou_Mean                   0.0\n",
      "inonemin_Mean                 0.0\n",
      "threeway_Mean                 0.0\n",
      "mou_cvce_Mean                 0.0\n",
      "mou_cdat_Mean                 0.0\n",
      "mou_rvce_Mean                 0.0\n",
      "owylis_vce_Mean               0.0\n",
      "mouowylisv_Mean               0.0\n",
      "iwylis_vce_Mean               0.0\n",
      "mouiwylisv_Mean               0.0\n",
      "peak_vce_Mean                 0.0\n",
      "peak_dat_Mean                 0.0\n",
      "mou_peav_Mean                 0.0\n",
      "mou_pead_Mean                 0.0\n",
      "opk_vce_Mean                  0.0\n",
      "opk_dat_Mean                  0.0\n",
      "mou_opkv_Mean                 0.0\n",
      "mou_opkd_Mean                 0.0\n",
      "drop_blk_Mean                 0.0\n",
      "attempt_Mean                  0.0\n",
      "complete_Mean                 0.0\n",
      "callfwdv_Mean                 0.0\n",
      "callwait_Mean                 0.0\n",
      "months                        0.0\n",
      "uniqsubs                      0.0\n",
      "actvsubs                      0.0\n",
      "totcalls                      0.0\n",
      "totmou                        0.0\n",
      "totrev                        0.0\n",
      "adjrev                        0.0\n",
      "adjmou                        0.0\n",
      "adjqty                        0.0\n",
      "avgrev                        0.0\n",
      "avgmou                        0.0\n",
      "avgqty                        0.0\n",
      "avg3mou                       0.0\n",
      "avg3qty                       0.0\n",
      "avg3rev                       0.0\n",
      "avg6mou                       0.0\n",
      "avg6qty                       0.0\n",
      "avg6rev                       0.0\n",
      "hnd_price                     0.0\n",
      "phones                        0.0\n",
      "models                        0.0\n",
      "truck                         0.0\n",
      "rv                            0.0\n",
      "lor                           0.0\n",
      "adults                        0.0\n",
      "income                        0.0\n",
      "numbcars                      0.0\n",
      "forgntvl                      0.0\n",
      "eqpdays                       0.0\n",
      "new_cell                      0.0\n",
      "crclscod                      0.0\n",
      "asl_flag                      0.0\n",
      "prizm_social_one              0.0\n",
      "area                          0.0\n",
      "dualband                      0.0\n",
      "refurb_new                    0.0\n",
      "hnd_webcap                    0.0\n",
      "ownrent                       0.0\n",
      "dwlltype                      0.0\n",
      "marital                       0.0\n",
      "infobase                      0.0\n",
      "HHstatin                      0.0\n",
      "dwllsize                      0.0\n",
      "ethnic                        0.0\n",
      "kid0_2                        0.0\n",
      "kid3_5                        0.0\n",
      "kid6_10                       0.0\n",
      "kid11_15                      0.0\n",
      "kid16_17                      0.0\n",
      "creditcd                      0.0\n"
     ]
    }
   ],
   "source": [
    "percent_missing = customer_updated.isnull().sum() * 100 / len(customer)\n",
    "missing_value_df = pd.DataFrame({'percent_missing': percent_missing})\n",
    "print(missing_value_df.to_string())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "24d39fdc",
   "metadata": {},
   "outputs": [],
   "source": [
    "#normalizing the values\n",
    "x_norm = (customer_updated - customer_updated.min(axis=0)) / (customer_updated.max(axis=0) - customer_updated.min(axis=0))\n",
    "customer_updated[:] = x_norm"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "6959ccdd",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['rev_Mean', 'mou_Mean', 'totmrc_Mean', 'da_Mean', 'ovrmou_Mean',\n",
       "       'ovrrev_Mean', 'vceovr_Mean', 'datovr_Mean', 'roam_Mean',\n",
       "       'drop_vce_Mean', 'drop_dat_Mean', 'blck_vce_Mean', 'blck_dat_Mean',\n",
       "       'unan_vce_Mean', 'unan_dat_Mean', 'plcd_vce_Mean', 'plcd_dat_Mean',\n",
       "       'recv_vce_Mean', 'recv_sms_Mean', 'comp_vce_Mean', 'comp_dat_Mean',\n",
       "       'custcare_Mean', 'ccrndmou_Mean', 'cc_mou_Mean', 'inonemin_Mean',\n",
       "       'threeway_Mean', 'mou_cvce_Mean', 'mou_cdat_Mean', 'mou_rvce_Mean',\n",
       "       'owylis_vce_Mean', 'mouowylisv_Mean', 'iwylis_vce_Mean',\n",
       "       'mouiwylisv_Mean', 'peak_vce_Mean', 'peak_dat_Mean', 'mou_peav_Mean',\n",
       "       'mou_pead_Mean', 'opk_vce_Mean', 'opk_dat_Mean', 'mou_opkv_Mean',\n",
       "       'mou_opkd_Mean', 'drop_blk_Mean', 'attempt_Mean', 'complete_Mean',\n",
       "       'callfwdv_Mean', 'callwait_Mean', 'months', 'uniqsubs', 'actvsubs',\n",
       "       'totcalls', 'totmou', 'totrev', 'adjrev', 'adjmou', 'adjqty', 'avgrev',\n",
       "       'avgmou', 'avgqty', 'avg3mou', 'avg3qty', 'avg3rev', 'avg6mou',\n",
       "       'avg6qty', 'avg6rev', 'hnd_price', 'phones', 'models', 'truck', 'rv',\n",
       "       'lor', 'adults', 'income', 'numbcars', 'forgntvl', 'eqpdays',\n",
       "       'new_cell', 'crclscod', 'asl_flag', 'prizm_social_one', 'area',\n",
       "       'dualband', 'refurb_new', 'hnd_webcap', 'ownrent', 'dwlltype',\n",
       "       'marital', 'infobase', 'HHstatin', 'dwllsize', 'ethnic', 'kid0_2',\n",
       "       'kid3_5', 'kid6_10', 'kid11_15', 'kid16_17', 'creditcd'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "customer_updated.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "3d22b227",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Method1: VIF (not the best method) performed on first dataset.\n",
    "from statsmodels.tools.tools import add_constant\n",
    "from statsmodels.stats.outliers_influence import variance_inflation_factor as vif\n",
    "\n",
    "def vify(cus):\n",
    "    v=cus\n",
    "    vif_data = pd.DataFrame()\n",
    "    vif_data[\"feature\"] = v.columns\n",
    "    vif_data[\"VIF\"] = [vif(v.values, i)for i in range(len(v.columns))]\n",
    "    return vif_data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "6795e0f8",
   "metadata": {},
   "outputs": [],
   "source": [
    "v_r1= vify(customer_updated)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "6593b201",
   "metadata": {},
   "outputs": [],
   "source": [
    "hcorr= v_r1[v_r1['VIF']>10].feature.to_list()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "88d97315",
   "metadata": {},
   "outputs": [],
   "source": [
    "v1=customer_updated.drop(columns=hcorr)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "8d81350f",
   "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>da_Mean</th>\n",
       "      <th>roam_Mean</th>\n",
       "      <th>recv_sms_Mean</th>\n",
       "      <th>threeway_Mean</th>\n",
       "      <th>mouowylisv_Mean</th>\n",
       "      <th>iwylis_vce_Mean</th>\n",
       "      <th>mouiwylisv_Mean</th>\n",
       "      <th>callfwdv_Mean</th>\n",
       "      <th>callwait_Mean</th>\n",
       "      <th>uniqsubs</th>\n",
       "      <th>...</th>\n",
       "      <th>infobase</th>\n",
       "      <th>HHstatin</th>\n",
       "      <th>dwllsize</th>\n",
       "      <th>ethnic</th>\n",
       "      <th>kid0_2</th>\n",
       "      <th>kid3_5</th>\n",
       "      <th>kid6_10</th>\n",
       "      <th>kid11_15</th>\n",
       "      <th>kid16_17</th>\n",
       "      <th>creditcd</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.001553</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.001567</td>\n",
       "      <td>0.005128</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.4</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.5625</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.001553</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.005051</td>\n",
       "      <td>0.013434</td>\n",
       "      <td>0.012195</td>\n",
       "      <td>0.002170</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.026646</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.4</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.0000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.4</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.5625</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000142</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.4</td>\n",
       "      <td>0.214286</td>\n",
       "      <td>0.8750</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.006000</td>\n",
       "      <td>0.001284</td>\n",
       "      <td>0.000215</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.4</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.3750</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>99995</th>\n",
       "      <td>0.005576</td>\n",
       "      <td>0.000349</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000932</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.4</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.8125</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>99996</th>\n",
       "      <td>0.005576</td>\n",
       "      <td>0.000349</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.005051</td>\n",
       "      <td>0.008262</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.5625</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>99997</th>\n",
       "      <td>0.005576</td>\n",
       "      <td>0.000349</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.4</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.8750</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>99998</th>\n",
       "      <td>0.005576</td>\n",
       "      <td>0.000349</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.015152</td>\n",
       "      <td>0.034890</td>\n",
       "      <td>0.032092</td>\n",
       "      <td>0.019303</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.4</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.8125</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>99999</th>\n",
       "      <td>0.005576</td>\n",
       "      <td>0.000349</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.055556</td>\n",
       "      <td>0.003944</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.4</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.3125</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>100000 rows × 41 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        da_Mean  roam_Mean  recv_sms_Mean  threeway_Mean  mouowylisv_Mean  \\\n",
       "0      0.001553   0.000000            0.0       0.000000         0.000000   \n",
       "1      0.001553   0.000000            0.0       0.005051         0.013434   \n",
       "2      0.000000   0.000000            0.0       0.000000         0.000000   \n",
       "3      0.000000   0.000000            0.0       0.000000         0.000142   \n",
       "4      0.000000   0.000000            0.0       0.000000         0.006000   \n",
       "...         ...        ...            ...            ...              ...   \n",
       "99995  0.005576   0.000349            0.0       0.000000         0.000932   \n",
       "99996  0.005576   0.000349            0.0       0.005051         0.008262   \n",
       "99997  0.005576   0.000349            0.0       0.000000         0.000000   \n",
       "99998  0.005576   0.000349            0.0       0.015152         0.034890   \n",
       "99999  0.005576   0.000349            0.0       0.055556         0.003944   \n",
       "\n",
       "       iwylis_vce_Mean  mouiwylisv_Mean  callfwdv_Mean  callwait_Mean  \\\n",
       "0             0.000000         0.000000            0.0       0.001567   \n",
       "1             0.012195         0.002170            0.0       0.026646   \n",
       "2             0.000000         0.000000            0.0       0.000000   \n",
       "3             0.000000         0.000000            0.0       0.000000   \n",
       "4             0.001284         0.000215            0.0       0.000000   \n",
       "...                ...              ...            ...            ...   \n",
       "99995         0.000000         0.000000            0.0       0.000000   \n",
       "99996         0.000000         0.000000            0.0       0.000000   \n",
       "99997         0.000000         0.000000            0.0       0.000000   \n",
       "99998         0.032092         0.019303            0.0       0.000000   \n",
       "99999         0.000000         0.000000            0.0       0.000000   \n",
       "\n",
       "       uniqsubs  ...  infobase  HHstatin  dwllsize  ethnic  kid0_2  kid3_5  \\\n",
       "0      0.005128  ...       0.0       0.4  0.000000  0.5625     0.0     0.0   \n",
       "1      0.000000  ...       0.0       0.4  0.000000  1.0000     0.0     0.0   \n",
       "2      0.000000  ...       0.0       0.4  0.000000  0.5625     0.0     1.0   \n",
       "3      0.000000  ...       0.0       0.4  0.214286  0.8750     1.0     0.0   \n",
       "4      0.000000  ...       0.0       0.4  1.000000  0.3750     0.0     0.0   \n",
       "...         ...  ...       ...       ...       ...     ...     ...     ...   \n",
       "99995  0.000000  ...       0.0       0.4  0.000000  0.8125     0.0     0.0   \n",
       "99996  0.000000  ...       0.0       1.0  0.000000  0.5625     0.0     0.0   \n",
       "99997  0.000000  ...       0.0       0.4  0.000000  0.8750     1.0     1.0   \n",
       "99998  0.000000  ...       0.0       0.4  0.000000  0.8125     0.0     0.0   \n",
       "99999  0.000000  ...       0.0       0.4  0.000000  0.3125     0.0     0.0   \n",
       "\n",
       "       kid6_10  kid11_15  kid16_17  creditcd  \n",
       "0          0.0       0.0       0.0       1.0  \n",
       "1          0.0       0.0       0.0       1.0  \n",
       "2          0.0       0.0       0.0       1.0  \n",
       "3          0.0       0.0       0.0       1.0  \n",
       "4          0.0       0.0       0.0       1.0  \n",
       "...        ...       ...       ...       ...  \n",
       "99995      0.0       1.0       0.0       1.0  \n",
       "99996      1.0       1.0       1.0       1.0  \n",
       "99997      0.0       0.0       0.0       0.0  \n",
       "99998      0.0       0.0       0.0       0.0  \n",
       "99999      0.0       0.0       0.0       0.0  \n",
       "\n",
       "[100000 rows x 41 columns]"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "v1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "55c57971",
   "metadata": {},
   "outputs": [],
   "source": [
    "customer_v=v1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "cc3bf497",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['da_Mean', 'roam_Mean', 'recv_sms_Mean', 'threeway_Mean',\n",
       "       'mouowylisv_Mean', 'iwylis_vce_Mean', 'mouiwylisv_Mean',\n",
       "       'callfwdv_Mean', 'callwait_Mean', 'uniqsubs', 'hnd_price', 'phones',\n",
       "       'models', 'truck', 'rv', 'lor', 'adults', 'income', 'numbcars',\n",
       "       'forgntvl', 'eqpdays', 'new_cell', 'crclscod', 'asl_flag',\n",
       "       'prizm_social_one', 'area', 'dualband', 'refurb_new', 'ownrent',\n",
       "       'dwlltype', 'marital', 'infobase', 'HHstatin', 'dwllsize', 'ethnic',\n",
       "       'kid0_2', 'kid3_5', 'kid6_10', 'kid11_15', 'kid16_17', 'creditcd'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "customer_v.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "5e38ca2d",
   "metadata": {},
   "outputs": [],
   "source": [
    "x_norm = (customer_v - customer_v.min(axis=0)) / (customer_v.max(axis=0) - customer_v.min(axis=0))\n",
    "customer_v[:] = x_norm"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "30bf6c7c",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "X_train, X_test, y_train, y_test = train_test_split(customer_v, train_label_v, test_size=0.3, random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "47eae755",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.5936000000000001\n"
     ]
    }
   ],
   "source": [
    "#Implementing the models after VIF.\n",
    "xg=xgboost.XGBClassifier()\n",
    "xg.fit(X_train,y_train)\n",
    "k_folds = KFold(n_splits = 10)\n",
    "print(cross_val_score(xg, X_train, y_train, cv=k_folds).mean())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "5c66cd26",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.5925946954822328"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_xg=xg.predict(X_test)\n",
    "fpr, tpr, thresholds = metrics.roc_curve(y_test, pred_xg)\n",
    "xg_m=metrics.auc(fpr, tpr)\n",
    "xg_m"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "81a40d3e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.5325\n"
     ]
    }
   ],
   "source": [
    "dt=DecisionTreeClassifier()\n",
    "dt.fit(X_train,y_train)\n",
    "k_folds = KFold(n_splits = 10)\n",
    "print(cross_val_score(dt, X_train, y_train, cv=k_folds).mean())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "8a583af1",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.5316757486239753"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred=dt.predict(X_test)\n",
    "fpr, tpr, thresholds = metrics.roc_curve(y_test, pred)\n",
    "dt_m=metrics.auc(fpr, tpr)\n",
    "dt_m"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "cb160ba4",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.5874\n"
     ]
    }
   ],
   "source": [
    "#random Forest\n",
    "rf=RandomForestClassifier()\n",
    "rf.fit(X_train,y_train)\n",
    "k_folds = KFold(n_splits = 10)\n",
    "print(cross_val_score(rf, X_train, y_train, cv=k_folds).mean())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "dcca771c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.5889912178691572"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\n",
    "pred_rf=rf.predict(X_test)\n",
    "fpr, tpr, thresholds = metrics.roc_curve(y_test, pred_rf)\n",
    "rf_m=metrics.auc(fpr, tpr)\n",
    "rf_m"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "9d93f289",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.6018\n"
     ]
    }
   ],
   "source": [
    "#lightbgm\n",
    "gbm=lightgbm.LGBMClassifier()\n",
    "gbm.fit(X_train,y_train)\n",
    "k_folds = KFold(n_splits = 10)\n",
    "print(cross_val_score(gbm, X_train, y_train, cv=k_folds).mean())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "39e1394a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.5983899757900744"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_gbm=gbm.predict(X_test)\n",
    "fpr, tpr, thresholds = metrics.roc_curve(y_test, pred_gbm)\n",
    "gbm_m=metrics.auc(fpr, tpr)\n",
    "gbm_m"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "c89baf0f",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Method2: hierarchal clustering (not a best method) on firstdataset.\n",
    "corr2=customer_updated.corr(method='spearman')\n",
    "data=customer_updated"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "id": "73c0005c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x576 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, (ax1, ax2) = plt.subplots(2, figsize=(12, 8))\n",
    "\n",
    "distance_matrix = 1 - np.abs(corr2)\n",
    "dist_linkage = hierarchy.ward(squareform(distance_matrix))\n",
    "dendro = hierarchy.dendrogram(\n",
    "    dist_linkage, labels=data.columns.tolist(), ax=ax1, leaf_rotation=90\n",
    ")\n",
    "dendro_idx = np.arange(0, len(dendro[\"ivl\"]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "id": "c269a645",
   "metadata": {},
   "outputs": [],
   "source": [
    "cluster_ids = hierarchy.fcluster(dist_linkage, 1, criterion=\"distance\")\n",
    "cluster_id_to_feature_ids = defaultdict(list)\n",
    "for idx, cluster_id in enumerate(cluster_ids):\n",
    "    cluster_id_to_feature_ids[cluster_id].append(idx)\n",
    "selected_features = [v[0] for v in cluster_id_to_feature_ids.values()]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "aca00b08",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[0, 4, 7, 8, 9, 10, 21, 44, 46, 47, 65, 67, 69, 71, 73, 76, 84, 86, 90, 93]"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "selected_features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "460f4861",
   "metadata": {},
   "outputs": [],
   "source": [
    "r=customer_updated.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "id": "d1da54cb",
   "metadata": {},
   "outputs": [],
   "source": [
    "r=r[selected_features]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "id": "ae64fa1f",
   "metadata": {},
   "outputs": [],
   "source": [
    "h=customer_updated[r]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "id": "42321d2b",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "X_train, X_test, y_train, y_test = train_test_split(h, train_label_v, test_size=0.3, random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "id": "3c2285fb",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.5852857142857143\n"
     ]
    }
   ],
   "source": [
    "#Implementing the models After hierarchial feature selection\n",
    "xg=xgboost.XGBClassifier()\n",
    "xg.fit(X_train,y_train)\n",
    "k_folds = KFold(n_splits = 10)\n",
    "print(cross_val_score(xg, X_train, y_train, cv=k_folds).mean())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "id": "805ced24",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.60      0.58      0.59     15031\n",
      "           1       0.59      0.61      0.60     14969\n",
      "\n",
      "    accuracy                           0.59     30000\n",
      "   macro avg       0.59      0.59      0.59     30000\n",
      "weighted avg       0.59      0.59      0.59     30000\n",
      "\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import classification_report\n",
    "pred_xg=xg.predict(X_test)\n",
    "fpr, tpr, thresholds = metrics.roc_curve(y_test, pred_xg)\n",
    "r=classification_report(y_test, pred_xg)\n",
    "xg_m=metrics.auc(fpr, tpr)\n",
    "print(r)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "id": "05f4ff65",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.5263714285714286\n"
     ]
    }
   ],
   "source": [
    "dt=DecisionTreeClassifier()\n",
    "dt.fit(X_train,y_train)\n",
    "k_folds = KFold(n_splits = 10)\n",
    "print(cross_val_score(dt, X_train, y_train, cv=k_folds).mean())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "id": "28a77c74",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.5288885500528737"
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred=dt.predict(X_test)\n",
    "fpr, tpr, thresholds = metrics.roc_curve(y_test, pred)\n",
    "dt_m=metrics.auc(fpr, tpr)\n",
    "dt_m"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "id": "ea9e4749",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.5748714285714286\n"
     ]
    }
   ],
   "source": [
    "#random Forest\n",
    "rf=RandomForestClassifier()\n",
    "rf.fit(X_train,y_train)\n",
    "k_folds = KFold(n_splits = 10)\n",
    "print(cross_val_score(rf, X_train, y_train, cv=k_folds).mean())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "id": "8d5a3974",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.5798356632092104"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_rf=rf.predict(X_test)\n",
    "fpr, tpr, thresholds = metrics.roc_curve(y_test, pred_rf)\n",
    "rf_m=metrics.auc(fpr, tpr)\n",
    "rf_m"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "id": "27dbd1ff",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.5901857142857143\n"
     ]
    }
   ],
   "source": [
    "gbm=lightgbm.LGBMClassifier()\n",
    "gbm.fit(X_train,y_train)\n",
    "k_folds = KFold(n_splits = 10)\n",
    "print(cross_val_score(gbm, X_train, y_train, cv=k_folds).mean())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "id": "5d979950",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.5965807969517594"
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_gbm=gbm.predict(X_test)\n",
    "fpr, tpr, thresholds = metrics.roc_curve(y_test, pred_gbm)\n",
    "gbm_m=metrics.auc(fpr, tpr)\n",
    "gbm_m"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "id": "a3bf6f9a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['rev_Mean', 'mou_Mean', 'totmrc_Mean', 'da_Mean', 'ovrmou_Mean',\n",
       "       'ovrrev_Mean', 'vceovr_Mean', 'datovr_Mean', 'roam_Mean',\n",
       "       'drop_vce_Mean', 'drop_dat_Mean', 'blck_vce_Mean', 'blck_dat_Mean',\n",
       "       'unan_vce_Mean', 'unan_dat_Mean', 'plcd_vce_Mean', 'plcd_dat_Mean',\n",
       "       'recv_vce_Mean', 'recv_sms_Mean', 'comp_vce_Mean', 'comp_dat_Mean',\n",
       "       'custcare_Mean', 'ccrndmou_Mean', 'cc_mou_Mean', 'inonemin_Mean',\n",
       "       'threeway_Mean', 'mou_cvce_Mean', 'mou_cdat_Mean', 'mou_rvce_Mean',\n",
       "       'owylis_vce_Mean', 'mouowylisv_Mean', 'iwylis_vce_Mean',\n",
       "       'mouiwylisv_Mean', 'peak_vce_Mean', 'peak_dat_Mean', 'mou_peav_Mean',\n",
       "       'mou_pead_Mean', 'opk_vce_Mean', 'opk_dat_Mean', 'mou_opkv_Mean',\n",
       "       'mou_opkd_Mean', 'drop_blk_Mean', 'attempt_Mean', 'complete_Mean',\n",
       "       'callfwdv_Mean', 'callwait_Mean', 'months', 'uniqsubs', 'actvsubs',\n",
       "       'totcalls', 'totmou', 'totrev', 'adjrev', 'adjmou', 'adjqty', 'avgrev',\n",
       "       'avgmou', 'avgqty', 'avg3mou', 'avg3qty', 'avg3rev', 'avg6mou',\n",
       "       'avg6qty', 'avg6rev', 'hnd_price', 'phones', 'models', 'truck', 'rv',\n",
       "       'lor', 'adults', 'income', 'numbcars', 'forgntvl', 'eqpdays',\n",
       "       'new_cell', 'crclscod', 'asl_flag', 'prizm_social_one', 'area',\n",
       "       'dualband', 'refurb_new', 'hnd_webcap', 'ownrent', 'dwlltype',\n",
       "       'marital', 'infobase', 'HHstatin', 'dwllsize', 'ethnic', 'kid0_2',\n",
       "       'kid3_5', 'kid6_10', 'kid11_15', 'kid16_17', 'creditcd'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "customer_updated.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "id": "35dcfac8",
   "metadata": {},
   "outputs": [],
   "source": [
    "p1 = customer_updated"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "id": "4227e144",
   "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>rev_Mean</th>\n",
       "      <th>mou_Mean</th>\n",
       "      <th>totmrc_Mean</th>\n",
       "      <th>da_Mean</th>\n",
       "      <th>ovrmou_Mean</th>\n",
       "      <th>ovrrev_Mean</th>\n",
       "      <th>vceovr_Mean</th>\n",
       "      <th>datovr_Mean</th>\n",
       "      <th>roam_Mean</th>\n",
       "      <th>drop_vce_Mean</th>\n",
       "      <th>...</th>\n",
       "      <th>infobase</th>\n",
       "      <th>HHstatin</th>\n",
       "      <th>dwllsize</th>\n",
       "      <th>ethnic</th>\n",
       "      <th>kid0_2</th>\n",
       "      <th>kid3_5</th>\n",
       "      <th>kid6_10</th>\n",
       "      <th>kid11_15</th>\n",
       "      <th>kid16_17</th>\n",
       "      <th>creditcd</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.007836</td>\n",
       "      <td>0.017961</td>\n",
       "      <td>0.113102</td>\n",
       "      <td>0.001553</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>0.002865</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.4</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.5625</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.016538</td>\n",
       "      <td>0.039548</td>\n",
       "      <td>0.147263</td>\n",
       "      <td>0.001553</td>\n",
       "      <td>0.005265</td>\n",
       "      <td>0.008255</td>\n",
       "      <td>0.010155</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.035817</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.4</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.0000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.006016</td>\n",
       "      <td>0.000840</td>\n",
       "      <td>0.100491</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>0.001433</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.4</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.5625</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.011474</td>\n",
       "      <td>0.000614</td>\n",
       "      <td>0.148579</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>0.0</td>\n",
       "      <td>0.4</td>\n",
       "      <td>0.214286</td>\n",
       "      <td>0.8750</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.015950</td>\n",
       "      <td>0.046736</td>\n",
       "      <td>0.226354</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>0.041547</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.4</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.3750</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>99995</th>\n",
       "      <td>0.016856</td>\n",
       "      <td>0.042072</td>\n",
       "      <td>0.167300</td>\n",
       "      <td>0.005576</td>\n",
       "      <td>0.009506</td>\n",
       "      <td>0.012300</td>\n",
       "      <td>0.014837</td>\n",
       "      <td>0.000617</td>\n",
       "      <td>0.000349</td>\n",
       "      <td>0.008596</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.4</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.8125</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>99996</th>\n",
       "      <td>0.016856</td>\n",
       "      <td>0.042072</td>\n",
       "      <td>0.167300</td>\n",
       "      <td>0.005576</td>\n",
       "      <td>0.009506</td>\n",
       "      <td>0.012300</td>\n",
       "      <td>0.014837</td>\n",
       "      <td>0.000617</td>\n",
       "      <td>0.000349</td>\n",
       "      <td>0.012894</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.5625</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>99997</th>\n",
       "      <td>0.016856</td>\n",
       "      <td>0.042072</td>\n",
       "      <td>0.167300</td>\n",
       "      <td>0.005576</td>\n",
       "      <td>0.009506</td>\n",
       "      <td>0.012300</td>\n",
       "      <td>0.014837</td>\n",
       "      <td>0.000617</td>\n",
       "      <td>0.000349</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.4</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.8750</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>99998</th>\n",
       "      <td>0.016856</td>\n",
       "      <td>0.042072</td>\n",
       "      <td>0.167300</td>\n",
       "      <td>0.005576</td>\n",
       "      <td>0.009506</td>\n",
       "      <td>0.012300</td>\n",
       "      <td>0.014837</td>\n",
       "      <td>0.000617</td>\n",
       "      <td>0.000349</td>\n",
       "      <td>0.080229</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.4</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.8125</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>99999</th>\n",
       "      <td>0.016856</td>\n",
       "      <td>0.042072</td>\n",
       "      <td>0.167300</td>\n",
       "      <td>0.005576</td>\n",
       "      <td>0.009506</td>\n",
       "      <td>0.012300</td>\n",
       "      <td>0.014837</td>\n",
       "      <td>0.000617</td>\n",
       "      <td>0.000349</td>\n",
       "      <td>0.024355</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.4</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.3125</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>100000 rows × 96 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       rev_Mean  mou_Mean  totmrc_Mean   da_Mean  ovrmou_Mean  ovrrev_Mean  \\\n",
       "0      0.007836  0.017961     0.113102  0.001553     0.000000     0.000000   \n",
       "1      0.016538  0.039548     0.147263  0.001553     0.005265     0.008255   \n",
       "2      0.006016  0.000840     0.100491  0.000000     0.000000     0.000000   \n",
       "3      0.011474  0.000614     0.148579  0.000000     0.000000     0.000000   \n",
       "4      0.015950  0.046736     0.226354  0.000000     0.000000     0.000000   \n",
       "...         ...       ...          ...       ...          ...          ...   \n",
       "99995  0.016856  0.042072     0.167300  0.005576     0.009506     0.012300   \n",
       "99996  0.016856  0.042072     0.167300  0.005576     0.009506     0.012300   \n",
       "99997  0.016856  0.042072     0.167300  0.005576     0.009506     0.012300   \n",
       "99998  0.016856  0.042072     0.167300  0.005576     0.009506     0.012300   \n",
       "99999  0.016856  0.042072     0.167300  0.005576     0.009506     0.012300   \n",
       "\n",
       "       vceovr_Mean  datovr_Mean  roam_Mean  drop_vce_Mean  ...  infobase  \\\n",
       "0         0.000000     0.000000   0.000000       0.002865  ...       0.0   \n",
       "1         0.010155     0.000000   0.000000       0.035817  ...       0.0   \n",
       "2         0.000000     0.000000   0.000000       0.001433  ...       0.0   \n",
       "3         0.000000     0.000000   0.000000       0.000000  ...       0.0   \n",
       "4         0.000000     0.000000   0.000000       0.041547  ...       0.0   \n",
       "...            ...          ...        ...            ...  ...       ...   \n",
       "99995     0.014837     0.000617   0.000349       0.008596  ...       0.0   \n",
       "99996     0.014837     0.000617   0.000349       0.012894  ...       0.0   \n",
       "99997     0.014837     0.000617   0.000349       0.000000  ...       0.0   \n",
       "99998     0.014837     0.000617   0.000349       0.080229  ...       0.0   \n",
       "99999     0.014837     0.000617   0.000349       0.024355  ...       0.0   \n",
       "\n",
       "       HHstatin  dwllsize  ethnic  kid0_2  kid3_5  kid6_10  kid11_15  \\\n",
       "0           0.4  0.000000  0.5625     0.0     0.0      0.0       0.0   \n",
       "1           0.4  0.000000  1.0000     0.0     0.0      0.0       0.0   \n",
       "2           0.4  0.000000  0.5625     0.0     1.0      0.0       0.0   \n",
       "3           0.4  0.214286  0.8750     1.0     0.0      0.0       0.0   \n",
       "4           0.4  1.000000  0.3750     0.0     0.0      0.0       0.0   \n",
       "...         ...       ...     ...     ...     ...      ...       ...   \n",
       "99995       0.4  0.000000  0.8125     0.0     0.0      0.0       1.0   \n",
       "99996       1.0  0.000000  0.5625     0.0     0.0      1.0       1.0   \n",
       "99997       0.4  0.000000  0.8750     1.0     1.0      0.0       0.0   \n",
       "99998       0.4  0.000000  0.8125     0.0     0.0      0.0       0.0   \n",
       "99999       0.4  0.000000  0.3125     0.0     0.0      0.0       0.0   \n",
       "\n",
       "       kid16_17  creditcd  \n",
       "0           0.0       1.0  \n",
       "1           0.0       1.0  \n",
       "2           0.0       1.0  \n",
       "3           0.0       1.0  \n",
       "4           0.0       1.0  \n",
       "...         ...       ...  \n",
       "99995       0.0       1.0  \n",
       "99996       1.0       1.0  \n",
       "99997       0.0       0.0  \n",
       "99998       0.0       0.0  \n",
       "99999       0.0       0.0  \n",
       "\n",
       "[100000 rows x 96 columns]"
      ]
     },
     "execution_count": 58,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "p1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "id": "5fd908f6",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Method3: PCA on first dataset. (not the best method) on first dataset."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "id": "1aabc43a",
   "metadata": {},
   "outputs": [],
   "source": [
    "pca_1 = PCA(n_components=90)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "id": "e9097d2a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "PCA(n_components=90)"
      ]
     },
     "execution_count": 61,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pca_1.fit(p1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "id": "78bfb6b1",
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pca1 = pca_1.transform(p1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "id": "69c8a508",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 1.33686005e-01, -7.98708405e-02, -1.39824830e-01, ...,\n",
       "         3.11337330e-07,  5.77680720e-05, -7.60370480e-07],\n",
       "       [ 5.93404796e-01,  9.51843296e-02,  7.63438136e-01, ...,\n",
       "        -1.29704024e-05,  1.60603927e-04,  1.04559570e-05],\n",
       "       [ 2.68290334e-01,  3.15053938e-01,  8.06870973e-01, ...,\n",
       "        -2.00911151e-05, -2.07561050e-05, -2.78173696e-06],\n",
       "       ...,\n",
       "       [-3.87387006e-01, -1.75147774e-01, -1.40189056e-01, ...,\n",
       "         1.46574362e-05, -2.17638592e-05,  8.18596557e-06],\n",
       "       [-7.75649250e-01, -2.74569429e-01, -1.47022724e-01, ...,\n",
       "        -1.04239993e-05, -8.35729776e-05,  1.42595377e-05],\n",
       "       [-4.52213654e-01, -1.70839769e-01, -2.32262976e-01, ...,\n",
       "        -3.68766458e-05, -7.34193476e-05,  8.03997851e-06]])"
      ]
     },
     "execution_count": 63,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_pca1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "id": "375aeed4",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(100000, 90)"
      ]
     },
     "execution_count": 64,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_pca1.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "id": "181d7ab5",
   "metadata": {},
   "outputs": [],
   "source": [
    "dt = DecisionTreeClassifier()\n",
    "xg = xgboost.XGBRegressor()\n",
    "rf = RandomForestClassifier()\n",
    "gbm = lightgbm.LGBMClassifier()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "id": "445a27ae",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "X_train, X_test, y_train, y_test = train_test_split(y_pca1, train_label, test_size=0.3, random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "id": "b6ec765e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.5308714285714285\n"
     ]
    }
   ],
   "source": [
    "#1.desicion tree. \n",
    "dt.fit(X_train,y_train)\n",
    "k_folds = KFold(n_splits = 10)\n",
    "print(cross_val_score(dt, X_train, y_train, cv=k_folds).mean())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "id": "18fcb015",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.5263615437041933"
      ]
     },
     "execution_count": 68,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_dt = dt.predict(X_test)\n",
    "fpr_dt, tpr_dt, thresholds_dt = metrics.roc_curve(y_test, pred_dt)\n",
    "dt_m = metrics.auc(fpr_dt, tpr_dt)\n",
    "dt_m"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "id": "b18b25c2",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.5874428571428572\n"
     ]
    }
   ],
   "source": [
    "#2.Random Forest.\n",
    "rf.fit(X_train,y_train)\n",
    "k_folds = KFold(n_splits = 10)\n",
    "print(cross_val_score(rf, X_train,y_train, cv=k_folds).mean())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "id": "66349169",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.5867118481337158"
      ]
     },
     "execution_count": 70,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_rf = rf.predict(X_test)\n",
    "fpr_rf, tpr_rf, thresholds_rf = metrics.roc_curve(y_test, pred_rf)\n",
    "rf_m = metrics.auc(fpr_rf, tpr_rf)\n",
    "rf_m"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "id": "5be04ab2",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.027551847891131166\n"
     ]
    }
   ],
   "source": [
    "#3.XGBoost.\n",
    "xg.fit(X_train,y_train)\n",
    "k_folds = KFold(n_splits = 10)\n",
    "print(cross_val_score(xg, X_train,y_train, cv=k_folds).mean())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "id": "b5276204",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.6294182127595664"
      ]
     },
     "execution_count": 72,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_xg = xg.predict(X_test)\n",
    "fpr_xg, tpr_xg, thresholds_xg = metrics.roc_curve(y_test, pred_xg)\n",
    "xg_m = metrics.auc(fpr_xg, tpr_xg)\n",
    "xg_m"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "id": "d3bd3bbc",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.6056142857142859\n"
     ]
    }
   ],
   "source": [
    "#4.Light GBM.\n",
    "gbm.fit(X_train,y_train)\n",
    "k_folds = KFold(n_splits = 10)\n",
    "print(cross_val_score(gbm, X_train,y_train, cv=k_folds).mean())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "id": "a4d4e608",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.6075577571689095"
      ]
     },
     "execution_count": 74,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_gbm = gbm.predict(X_test)\n",
    "fpr_gbm, tpr_gbm, thresholds_gbm = metrics.roc_curve(y_test, pred_gbm)\n",
    "gbm_m = metrics.auc(fpr_gbm, tpr_gbm)\n",
    "gbm_m"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "eb72d0df",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "687b8ce2",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "id": "94dc1e7e",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Reading in the dataset 2.\n",
    "customer2 = pd.read_csv('second.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "id": "4b299853",
   "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>CustomerID</th>\n",
       "      <th>Churn</th>\n",
       "      <th>MonthlyRevenue</th>\n",
       "      <th>MonthlyMinutes</th>\n",
       "      <th>TotalRecurringCharge</th>\n",
       "      <th>DirectorAssistedCalls</th>\n",
       "      <th>OverageMinutes</th>\n",
       "      <th>RoamingCalls</th>\n",
       "      <th>PercChangeMinutes</th>\n",
       "      <th>PercChangeRevenues</th>\n",
       "      <th>...</th>\n",
       "      <th>ReferralsMadeBySubscriber</th>\n",
       "      <th>IncomeGroup</th>\n",
       "      <th>OwnsMotorcycle</th>\n",
       "      <th>AdjustmentsToCreditRating</th>\n",
       "      <th>HandsetPrice</th>\n",
       "      <th>MadeCallToRetentionTeam</th>\n",
       "      <th>CreditRating</th>\n",
       "      <th>PrizmCode</th>\n",
       "      <th>Occupation</th>\n",
       "      <th>MaritalStatus</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>3000002</td>\n",
       "      <td>Yes</td>\n",
       "      <td>24.00</td>\n",
       "      <td>219.0</td>\n",
       "      <td>22.0</td>\n",
       "      <td>0.25</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-157.0</td>\n",
       "      <td>-19.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "      <td>No</td>\n",
       "      <td>0</td>\n",
       "      <td>30</td>\n",
       "      <td>Yes</td>\n",
       "      <td>1-Highest</td>\n",
       "      <td>Suburban</td>\n",
       "      <td>Professional</td>\n",
       "      <td>No</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>3000010</td>\n",
       "      <td>Yes</td>\n",
       "      <td>16.99</td>\n",
       "      <td>10.0</td>\n",
       "      <td>17.0</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-4.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>5</td>\n",
       "      <td>No</td>\n",
       "      <td>0</td>\n",
       "      <td>30</td>\n",
       "      <td>No</td>\n",
       "      <td>4-Medium</td>\n",
       "      <td>Suburban</td>\n",
       "      <td>Professional</td>\n",
       "      <td>Yes</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3000014</td>\n",
       "      <td>No</td>\n",
       "      <td>38.00</td>\n",
       "      <td>8.0</td>\n",
       "      <td>38.0</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>6</td>\n",
       "      <td>No</td>\n",
       "      <td>0</td>\n",
       "      <td>Unknown</td>\n",
       "      <td>No</td>\n",
       "      <td>3-Good</td>\n",
       "      <td>Town</td>\n",
       "      <td>Crafts</td>\n",
       "      <td>Yes</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3000022</td>\n",
       "      <td>No</td>\n",
       "      <td>82.28</td>\n",
       "      <td>1312.0</td>\n",
       "      <td>75.0</td>\n",
       "      <td>1.24</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>157.0</td>\n",
       "      <td>8.1</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>6</td>\n",
       "      <td>No</td>\n",
       "      <td>0</td>\n",
       "      <td>10</td>\n",
       "      <td>No</td>\n",
       "      <td>4-Medium</td>\n",
       "      <td>Other</td>\n",
       "      <td>Other</td>\n",
       "      <td>No</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>3000026</td>\n",
       "      <td>Yes</td>\n",
       "      <td>17.14</td>\n",
       "      <td>0.0</td>\n",
       "      <td>17.0</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.2</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>9</td>\n",
       "      <td>No</td>\n",
       "      <td>1</td>\n",
       "      <td>10</td>\n",
       "      <td>No</td>\n",
       "      <td>1-Highest</td>\n",
       "      <td>Other</td>\n",
       "      <td>Professional</td>\n",
       "      <td>Yes</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>51042</th>\n",
       "      <td>3399958</td>\n",
       "      <td>Yes</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>6</td>\n",
       "      <td>No</td>\n",
       "      <td>0</td>\n",
       "      <td>60</td>\n",
       "      <td>No</td>\n",
       "      <td>1-Highest</td>\n",
       "      <td>Suburban</td>\n",
       "      <td>Other</td>\n",
       "      <td>Yes</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>51043</th>\n",
       "      <td>3399974</td>\n",
       "      <td>No</td>\n",
       "      <td>95.17</td>\n",
       "      <td>1745.0</td>\n",
       "      <td>85.0</td>\n",
       "      <td>0.99</td>\n",
       "      <td>45.0</td>\n",
       "      <td>4.7</td>\n",
       "      <td>122.0</td>\n",
       "      <td>15.9</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>9</td>\n",
       "      <td>No</td>\n",
       "      <td>1</td>\n",
       "      <td>60</td>\n",
       "      <td>No</td>\n",
       "      <td>3-Good</td>\n",
       "      <td>Other</td>\n",
       "      <td>Other</td>\n",
       "      <td>No</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>51044</th>\n",
       "      <td>3399978</td>\n",
       "      <td>Yes</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>7</td>\n",
       "      <td>No</td>\n",
       "      <td>1</td>\n",
       "      <td>80</td>\n",
       "      <td>No</td>\n",
       "      <td>5-Low</td>\n",
       "      <td>Other</td>\n",
       "      <td>Clerical</td>\n",
       "      <td>No</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>51045</th>\n",
       "      <td>3399990</td>\n",
       "      <td>No</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>9</td>\n",
       "      <td>No</td>\n",
       "      <td>0</td>\n",
       "      <td>30</td>\n",
       "      <td>No</td>\n",
       "      <td>5-Low</td>\n",
       "      <td>Other</td>\n",
       "      <td>Other</td>\n",
       "      <td>No</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>51046</th>\n",
       "      <td>3399994</td>\n",
       "      <td>No</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>No</td>\n",
       "      <td>1</td>\n",
       "      <td>60</td>\n",
       "      <td>Yes</td>\n",
       "      <td>5-Low</td>\n",
       "      <td>Other</td>\n",
       "      <td>Other</td>\n",
       "      <td>Unknown</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>51047 rows × 58 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       CustomerID Churn  MonthlyRevenue  MonthlyMinutes  TotalRecurringCharge  \\\n",
       "0         3000002   Yes           24.00           219.0                  22.0   \n",
       "1         3000010   Yes           16.99            10.0                  17.0   \n",
       "2         3000014    No           38.00             8.0                  38.0   \n",
       "3         3000022    No           82.28          1312.0                  75.0   \n",
       "4         3000026   Yes           17.14             0.0                  17.0   \n",
       "...           ...   ...             ...             ...                   ...   \n",
       "51042     3399958   Yes             NaN             NaN                   NaN   \n",
       "51043     3399974    No           95.17          1745.0                  85.0   \n",
       "51044     3399978   Yes             NaN             NaN                   NaN   \n",
       "51045     3399990    No             NaN             NaN                   NaN   \n",
       "51046     3399994    No             NaN             NaN                   NaN   \n",
       "\n",
       "       DirectorAssistedCalls  OverageMinutes  RoamingCalls  PercChangeMinutes  \\\n",
       "0                       0.25             0.0           0.0             -157.0   \n",
       "1                       0.00             0.0           0.0               -4.0   \n",
       "2                       0.00             0.0           0.0               -2.0   \n",
       "3                       1.24             0.0           0.0              157.0   \n",
       "4                       0.00             0.0           0.0                0.0   \n",
       "...                      ...             ...           ...                ...   \n",
       "51042                    NaN             NaN           NaN                NaN   \n",
       "51043                   0.99            45.0           4.7              122.0   \n",
       "51044                    NaN             NaN           NaN                NaN   \n",
       "51045                    NaN             NaN           NaN                NaN   \n",
       "51046                    NaN             NaN           NaN                NaN   \n",
       "\n",
       "       PercChangeRevenues  ...  ReferralsMadeBySubscriber  IncomeGroup  \\\n",
       "0                   -19.0  ...                          0            4   \n",
       "1                     0.0  ...                          0            5   \n",
       "2                     0.0  ...                          0            6   \n",
       "3                     8.1  ...                          0            6   \n",
       "4                    -0.2  ...                          0            9   \n",
       "...                   ...  ...                        ...          ...   \n",
       "51042                 NaN  ...                          0            6   \n",
       "51043                15.9  ...                          0            9   \n",
       "51044                 NaN  ...                          0            7   \n",
       "51045                 NaN  ...                          0            9   \n",
       "51046                 NaN  ...                          0            0   \n",
       "\n",
       "       OwnsMotorcycle  AdjustmentsToCreditRating  HandsetPrice  \\\n",
       "0                  No                          0            30   \n",
       "1                  No                          0            30   \n",
       "2                  No                          0       Unknown   \n",
       "3                  No                          0            10   \n",
       "4                  No                          1            10   \n",
       "...               ...                        ...           ...   \n",
       "51042              No                          0            60   \n",
       "51043              No                          1            60   \n",
       "51044              No                          1            80   \n",
       "51045              No                          0            30   \n",
       "51046              No                          1            60   \n",
       "\n",
       "       MadeCallToRetentionTeam  CreditRating  PrizmCode    Occupation  \\\n",
       "0                          Yes     1-Highest   Suburban  Professional   \n",
       "1                           No      4-Medium   Suburban  Professional   \n",
       "2                           No        3-Good       Town        Crafts   \n",
       "3                           No      4-Medium      Other         Other   \n",
       "4                           No     1-Highest      Other  Professional   \n",
       "...                        ...           ...        ...           ...   \n",
       "51042                       No     1-Highest   Suburban         Other   \n",
       "51043                       No        3-Good      Other         Other   \n",
       "51044                       No         5-Low      Other      Clerical   \n",
       "51045                       No         5-Low      Other         Other   \n",
       "51046                      Yes         5-Low      Other         Other   \n",
       "\n",
       "       MaritalStatus  \n",
       "0                 No  \n",
       "1                Yes  \n",
       "2                Yes  \n",
       "3                 No  \n",
       "4                Yes  \n",
       "...              ...  \n",
       "51042            Yes  \n",
       "51043             No  \n",
       "51044             No  \n",
       "51045             No  \n",
       "51046        Unknown  \n",
       "\n",
       "[51047 rows x 58 columns]"
      ]
     },
     "execution_count": 76,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "customer2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "id": "d631b026",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "CustomerID                     0\n",
       "Churn                          0\n",
       "MonthlyRevenue               156\n",
       "MonthlyMinutes               156\n",
       "TotalRecurringCharge         156\n",
       "DirectorAssistedCalls        156\n",
       "OverageMinutes               156\n",
       "RoamingCalls                 156\n",
       "PercChangeMinutes            367\n",
       "PercChangeRevenues           367\n",
       "DroppedCalls                   0\n",
       "BlockedCalls                   0\n",
       "UnansweredCalls                0\n",
       "CustomerCareCalls              0\n",
       "ThreewayCalls                  0\n",
       "ReceivedCalls                  0\n",
       "OutboundCalls                  0\n",
       "InboundCalls                   0\n",
       "PeakCallsInOut                 0\n",
       "OffPeakCallsInOut              0\n",
       "DroppedBlockedCalls            0\n",
       "CallForwardingCalls            0\n",
       "CallWaitingCalls               0\n",
       "MonthsInService                0\n",
       "UniqueSubs                     0\n",
       "ActiveSubs                     0\n",
       "ServiceArea                   24\n",
       "Handsets                       1\n",
       "HandsetModels                  1\n",
       "CurrentEquipmentDays           1\n",
       "AgeHH1                       909\n",
       "AgeHH2                       909\n",
       "ChildrenInHH                   0\n",
       "HandsetRefurbished             0\n",
       "HandsetWebCapable              0\n",
       "TruckOwner                     0\n",
       "RVOwner                        0\n",
       "Homeownership                  0\n",
       "BuysViaMailOrder               0\n",
       "RespondsToMailOffers           0\n",
       "OptOutMailings                 0\n",
       "NonUSTravel                    0\n",
       "OwnsComputer                   0\n",
       "HasCreditCard                  0\n",
       "RetentionCalls                 0\n",
       "RetentionOffersAccepted        0\n",
       "NewCellphoneUser               0\n",
       "NotNewCellphoneUser            0\n",
       "ReferralsMadeBySubscriber      0\n",
       "IncomeGroup                    0\n",
       "OwnsMotorcycle                 0\n",
       "AdjustmentsToCreditRating      0\n",
       "HandsetPrice                   0\n",
       "MadeCallToRetentionTeam        0\n",
       "CreditRating                   0\n",
       "PrizmCode                      0\n",
       "Occupation                     0\n",
       "MaritalStatus                  0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 77,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "customer2.isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "id": "5f7d7b62",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                           percent_missing\n",
      "CustomerID                           0.000\n",
      "Churn                                0.000\n",
      "MonthlyRevenue                       0.156\n",
      "MonthlyMinutes                       0.156\n",
      "TotalRecurringCharge                 0.156\n",
      "DirectorAssistedCalls                0.156\n",
      "OverageMinutes                       0.156\n",
      "RoamingCalls                         0.156\n",
      "PercChangeMinutes                    0.367\n",
      "PercChangeRevenues                   0.367\n",
      "DroppedCalls                         0.000\n",
      "BlockedCalls                         0.000\n",
      "UnansweredCalls                      0.000\n",
      "CustomerCareCalls                    0.000\n",
      "ThreewayCalls                        0.000\n",
      "ReceivedCalls                        0.000\n",
      "OutboundCalls                        0.000\n",
      "InboundCalls                         0.000\n",
      "PeakCallsInOut                       0.000\n",
      "OffPeakCallsInOut                    0.000\n",
      "DroppedBlockedCalls                  0.000\n",
      "CallForwardingCalls                  0.000\n",
      "CallWaitingCalls                     0.000\n",
      "MonthsInService                      0.000\n",
      "UniqueSubs                           0.000\n",
      "ActiveSubs                           0.000\n",
      "ServiceArea                          0.024\n",
      "Handsets                             0.001\n",
      "HandsetModels                        0.001\n",
      "CurrentEquipmentDays                 0.001\n",
      "AgeHH1                               0.909\n",
      "AgeHH2                               0.909\n",
      "ChildrenInHH                         0.000\n",
      "HandsetRefurbished                   0.000\n",
      "HandsetWebCapable                    0.000\n",
      "TruckOwner                           0.000\n",
      "RVOwner                              0.000\n",
      "Homeownership                        0.000\n",
      "BuysViaMailOrder                     0.000\n",
      "RespondsToMailOffers                 0.000\n",
      "OptOutMailings                       0.000\n",
      "NonUSTravel                          0.000\n",
      "OwnsComputer                         0.000\n",
      "HasCreditCard                        0.000\n",
      "RetentionCalls                       0.000\n",
      "RetentionOffersAccepted              0.000\n",
      "NewCellphoneUser                     0.000\n",
      "NotNewCellphoneUser                  0.000\n",
      "ReferralsMadeBySubscriber            0.000\n",
      "IncomeGroup                          0.000\n",
      "OwnsMotorcycle                       0.000\n",
      "AdjustmentsToCreditRating            0.000\n",
      "HandsetPrice                         0.000\n",
      "MadeCallToRetentionTeam              0.000\n",
      "CreditRating                         0.000\n",
      "PrizmCode                            0.000\n",
      "Occupation                           0.000\n",
      "MaritalStatus                        0.000\n"
     ]
    }
   ],
   "source": [
    "percent_missing = customer2.isnull().sum() * 100 / len(customer)\n",
    "missing_value_df = pd.DataFrame({'percent_missing': percent_missing})\n",
    "print(missing_value_df.to_string())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "id": "e16305ce",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/var/folders/j5/9fvnh0857k3_lqjpjrk029hw0000gn/T/ipykernel_749/1598678812.py:1: FutureWarning: Dropping of nuisance columns in DataFrame reductions (with 'numeric_only=None') is deprecated; in a future version this will raise TypeError.  Select only valid columns before calling the reduction.\n",
      "  customer2.fillna(customer2.mean(), inplace=True)\n"
     ]
    }
   ],
   "source": [
    "customer2.fillna(customer2.mean(), inplace=True)\n",
    "for column in customer2.columns:\n",
    "    if customer2[column].dtype == 'object':\n",
    "        customer2[column].fillna(customer2[column].mode()[0], inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "id": "007f2ae8",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                           percent_missing\n",
      "CustomerID                             0.0\n",
      "Churn                                  0.0\n",
      "MonthlyRevenue                         0.0\n",
      "MonthlyMinutes                         0.0\n",
      "TotalRecurringCharge                   0.0\n",
      "DirectorAssistedCalls                  0.0\n",
      "OverageMinutes                         0.0\n",
      "RoamingCalls                           0.0\n",
      "PercChangeMinutes                      0.0\n",
      "PercChangeRevenues                     0.0\n",
      "DroppedCalls                           0.0\n",
      "BlockedCalls                           0.0\n",
      "UnansweredCalls                        0.0\n",
      "CustomerCareCalls                      0.0\n",
      "ThreewayCalls                          0.0\n",
      "ReceivedCalls                          0.0\n",
      "OutboundCalls                          0.0\n",
      "InboundCalls                           0.0\n",
      "PeakCallsInOut                         0.0\n",
      "OffPeakCallsInOut                      0.0\n",
      "DroppedBlockedCalls                    0.0\n",
      "CallForwardingCalls                    0.0\n",
      "CallWaitingCalls                       0.0\n",
      "MonthsInService                        0.0\n",
      "UniqueSubs                             0.0\n",
      "ActiveSubs                             0.0\n",
      "ServiceArea                            0.0\n",
      "Handsets                               0.0\n",
      "HandsetModels                          0.0\n",
      "CurrentEquipmentDays                   0.0\n",
      "AgeHH1                                 0.0\n",
      "AgeHH2                                 0.0\n",
      "ChildrenInHH                           0.0\n",
      "HandsetRefurbished                     0.0\n",
      "HandsetWebCapable                      0.0\n",
      "TruckOwner                             0.0\n",
      "RVOwner                                0.0\n",
      "Homeownership                          0.0\n",
      "BuysViaMailOrder                       0.0\n",
      "RespondsToMailOffers                   0.0\n",
      "OptOutMailings                         0.0\n",
      "NonUSTravel                            0.0\n",
      "OwnsComputer                           0.0\n",
      "HasCreditCard                          0.0\n",
      "RetentionCalls                         0.0\n",
      "RetentionOffersAccepted                0.0\n",
      "NewCellphoneUser                       0.0\n",
      "NotNewCellphoneUser                    0.0\n",
      "ReferralsMadeBySubscriber              0.0\n",
      "IncomeGroup                            0.0\n",
      "OwnsMotorcycle                         0.0\n",
      "AdjustmentsToCreditRating              0.0\n",
      "HandsetPrice                           0.0\n",
      "MadeCallToRetentionTeam                0.0\n",
      "CreditRating                           0.0\n",
      "PrizmCode                              0.0\n",
      "Occupation                             0.0\n",
      "MaritalStatus                          0.0\n"
     ]
    }
   ],
   "source": [
    "percent_missing = customer2.isnull().sum() * 100 / len(customer)\n",
    "missing_value_df = pd.DataFrame({'percent_missing': percent_missing})\n",
    "print(missing_value_df.to_string())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "id": "e9b01da8",
   "metadata": {},
   "outputs": [],
   "source": [
    "del customer2['CustomerID']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "id": "ef2d584b",
   "metadata": {},
   "outputs": [],
   "source": [
    "del customer2['ServiceArea']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "id": "caf486db",
   "metadata": {},
   "outputs": [],
   "source": [
    "train_label2=customer2['Churn']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "id": "6500fd27",
   "metadata": {},
   "outputs": [],
   "source": [
    "del customer2['Churn']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "id": "ad597665",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0        Yes\n",
       "1        Yes\n",
       "2         No\n",
       "3         No\n",
       "4        Yes\n",
       "        ... \n",
       "51042    Yes\n",
       "51043     No\n",
       "51044    Yes\n",
       "51045     No\n",
       "51046     No\n",
       "Name: Churn, Length: 51047, dtype: object"
      ]
     },
     "execution_count": 85,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_label2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "id": "9d13fc6e",
   "metadata": {},
   "outputs": [],
   "source": [
    "train_label2=train_label2.replace('No',0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "id": "45b4a320",
   "metadata": {},
   "outputs": [],
   "source": [
    "train_label2=train_label2.replace('Yes',1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "id": "71e7e4d7",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0        1\n",
       "1        1\n",
       "2        0\n",
       "3        0\n",
       "4        1\n",
       "        ..\n",
       "51042    1\n",
       "51043    0\n",
       "51044    1\n",
       "51045    0\n",
       "51046    0\n",
       "Name: Churn, Length: 51047, dtype: int64"
      ]
     },
     "execution_count": 88,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_label2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "id": "8b733fef",
   "metadata": {},
   "outputs": [],
   "source": [
    "cat_customer2= customer2.select_dtypes(include=['object']).columns\n",
    "customer_cat2=customer2[cat_customer2]\n",
    "num_customer2=customer2.select_dtypes(include=['float64', 'int64']).columns\n",
    "customer_num2=customer2[num_customer2]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "id": "7c1912b7",
   "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>MonthlyRevenue</th>\n",
       "      <th>MonthlyMinutes</th>\n",
       "      <th>TotalRecurringCharge</th>\n",
       "      <th>DirectorAssistedCalls</th>\n",
       "      <th>OverageMinutes</th>\n",
       "      <th>RoamingCalls</th>\n",
       "      <th>PercChangeMinutes</th>\n",
       "      <th>PercChangeRevenues</th>\n",
       "      <th>DroppedCalls</th>\n",
       "      <th>BlockedCalls</th>\n",
       "      <th>...</th>\n",
       "      <th>Handsets</th>\n",
       "      <th>HandsetModels</th>\n",
       "      <th>CurrentEquipmentDays</th>\n",
       "      <th>AgeHH1</th>\n",
       "      <th>AgeHH2</th>\n",
       "      <th>RetentionCalls</th>\n",
       "      <th>RetentionOffersAccepted</th>\n",
       "      <th>ReferralsMadeBySubscriber</th>\n",
       "      <th>IncomeGroup</th>\n",
       "      <th>AdjustmentsToCreditRating</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>24.000000</td>\n",
       "      <td>219.000000</td>\n",
       "      <td>22.000000</td>\n",
       "      <td>0.250000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>-157.000000</td>\n",
       "      <td>-19.000000</td>\n",
       "      <td>0.7</td>\n",
       "      <td>0.7</td>\n",
       "      <td>...</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>361.0</td>\n",
       "      <td>62.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>16.990000</td>\n",
       "      <td>10.000000</td>\n",
       "      <td>17.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>-4.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.3</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1504.0</td>\n",
       "      <td>40.0</td>\n",
       "      <td>42.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>5</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>38.000000</td>\n",
       "      <td>8.000000</td>\n",
       "      <td>38.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>-2.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1812.0</td>\n",
       "      <td>26.0</td>\n",
       "      <td>26.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>6</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>82.280000</td>\n",
       "      <td>1312.000000</td>\n",
       "      <td>75.000000</td>\n",
       "      <td>1.240000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>157.000000</td>\n",
       "      <td>8.100000</td>\n",
       "      <td>52.0</td>\n",
       "      <td>7.7</td>\n",
       "      <td>...</td>\n",
       "      <td>9.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>458.0</td>\n",
       "      <td>30.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>6</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>17.140000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>17.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>-0.200000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>4.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>852.0</td>\n",
       "      <td>46.0</td>\n",
       "      <td>54.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>9</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>51042</th>\n",
       "      <td>58.834492</td>\n",
       "      <td>525.653416</td>\n",
       "      <td>46.830088</td>\n",
       "      <td>0.895229</td>\n",
       "      <td>40.027785</td>\n",
       "      <td>1.236244</td>\n",
       "      <td>-11.547908</td>\n",
       "      <td>-1.191985</td>\n",
       "      <td>9.3</td>\n",
       "      <td>2.7</td>\n",
       "      <td>...</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>526.0</td>\n",
       "      <td>68.0</td>\n",
       "      <td>64.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>6</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>51043</th>\n",
       "      <td>95.170000</td>\n",
       "      <td>1745.000000</td>\n",
       "      <td>85.000000</td>\n",
       "      <td>0.990000</td>\n",
       "      <td>45.000000</td>\n",
       "      <td>4.700000</td>\n",
       "      <td>122.000000</td>\n",
       "      <td>15.900000</td>\n",
       "      <td>16.7</td>\n",
       "      <td>0.7</td>\n",
       "      <td>...</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>464.0</td>\n",
       "      <td>48.0</td>\n",
       "      <td>48.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>9</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>51044</th>\n",
       "      <td>58.834492</td>\n",
       "      <td>525.653416</td>\n",
       "      <td>46.830088</td>\n",
       "      <td>0.895229</td>\n",
       "      <td>40.027785</td>\n",
       "      <td>1.236244</td>\n",
       "      <td>-11.547908</td>\n",
       "      <td>-1.191985</td>\n",
       "      <td>71.7</td>\n",
       "      <td>4.3</td>\n",
       "      <td>...</td>\n",
       "      <td>3.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>378.0</td>\n",
       "      <td>36.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>7</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>51045</th>\n",
       "      <td>58.834492</td>\n",
       "      <td>525.653416</td>\n",
       "      <td>46.830088</td>\n",
       "      <td>0.895229</td>\n",
       "      <td>40.027785</td>\n",
       "      <td>1.236244</td>\n",
       "      <td>-11.547908</td>\n",
       "      <td>-1.191985</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>433.0</td>\n",
       "      <td>32.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>9</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>51046</th>\n",
       "      <td>58.834492</td>\n",
       "      <td>525.653416</td>\n",
       "      <td>46.830088</td>\n",
       "      <td>0.895229</td>\n",
       "      <td>40.027785</td>\n",
       "      <td>1.236244</td>\n",
       "      <td>-11.547908</td>\n",
       "      <td>-1.191985</td>\n",
       "      <td>18.7</td>\n",
       "      <td>0.7</td>\n",
       "      <td>...</td>\n",
       "      <td>7.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>75.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>51047 rows × 34 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       MonthlyRevenue  MonthlyMinutes  TotalRecurringCharge  \\\n",
       "0           24.000000      219.000000             22.000000   \n",
       "1           16.990000       10.000000             17.000000   \n",
       "2           38.000000        8.000000             38.000000   \n",
       "3           82.280000     1312.000000             75.000000   \n",
       "4           17.140000        0.000000             17.000000   \n",
       "...               ...             ...                   ...   \n",
       "51042       58.834492      525.653416             46.830088   \n",
       "51043       95.170000     1745.000000             85.000000   \n",
       "51044       58.834492      525.653416             46.830088   \n",
       "51045       58.834492      525.653416             46.830088   \n",
       "51046       58.834492      525.653416             46.830088   \n",
       "\n",
       "       DirectorAssistedCalls  OverageMinutes  RoamingCalls  PercChangeMinutes  \\\n",
       "0                   0.250000        0.000000      0.000000        -157.000000   \n",
       "1                   0.000000        0.000000      0.000000          -4.000000   \n",
       "2                   0.000000        0.000000      0.000000          -2.000000   \n",
       "3                   1.240000        0.000000      0.000000         157.000000   \n",
       "4                   0.000000        0.000000      0.000000           0.000000   \n",
       "...                      ...             ...           ...                ...   \n",
       "51042               0.895229       40.027785      1.236244         -11.547908   \n",
       "51043               0.990000       45.000000      4.700000         122.000000   \n",
       "51044               0.895229       40.027785      1.236244         -11.547908   \n",
       "51045               0.895229       40.027785      1.236244         -11.547908   \n",
       "51046               0.895229       40.027785      1.236244         -11.547908   \n",
       "\n",
       "       PercChangeRevenues  DroppedCalls  BlockedCalls  ...  Handsets  \\\n",
       "0              -19.000000           0.7           0.7  ...       2.0   \n",
       "1                0.000000           0.3           0.0  ...       2.0   \n",
       "2                0.000000           0.0           0.0  ...       1.0   \n",
       "3                8.100000          52.0           7.7  ...       9.0   \n",
       "4               -0.200000           0.0           0.0  ...       4.0   \n",
       "...                   ...           ...           ...  ...       ...   \n",
       "51042           -1.191985           9.3           2.7  ...       2.0   \n",
       "51043           15.900000          16.7           0.7  ...       2.0   \n",
       "51044           -1.191985          71.7           4.3  ...       3.0   \n",
       "51045           -1.191985           0.0           0.0  ...       2.0   \n",
       "51046           -1.191985          18.7           0.7  ...       7.0   \n",
       "\n",
       "       HandsetModels  CurrentEquipmentDays  AgeHH1  AgeHH2  RetentionCalls  \\\n",
       "0                2.0                 361.0    62.0     0.0               1   \n",
       "1                1.0                1504.0    40.0    42.0               0   \n",
       "2                1.0                1812.0    26.0    26.0               0   \n",
       "3                4.0                 458.0    30.0     0.0               0   \n",
       "4                3.0                 852.0    46.0    54.0               0   \n",
       "...              ...                   ...     ...     ...             ...   \n",
       "51042            2.0                 526.0    68.0    64.0               0   \n",
       "51043            2.0                 464.0    48.0    48.0               0   \n",
       "51044            2.0                 378.0    36.0     0.0               0   \n",
       "51045            2.0                 433.0    32.0     0.0               0   \n",
       "51046            5.0                  75.0     0.0     0.0               2   \n",
       "\n",
       "       RetentionOffersAccepted  ReferralsMadeBySubscriber  IncomeGroup  \\\n",
       "0                            0                          0            4   \n",
       "1                            0                          0            5   \n",
       "2                            0                          0            6   \n",
       "3                            0                          0            6   \n",
       "4                            0                          0            9   \n",
       "...                        ...                        ...          ...   \n",
       "51042                        0                          0            6   \n",
       "51043                        0                          0            9   \n",
       "51044                        0                          0            7   \n",
       "51045                        0                          0            9   \n",
       "51046                        1                          0            0   \n",
       "\n",
       "       AdjustmentsToCreditRating  \n",
       "0                              0  \n",
       "1                              0  \n",
       "2                              0  \n",
       "3                              0  \n",
       "4                              1  \n",
       "...                          ...  \n",
       "51042                          0  \n",
       "51043                          1  \n",
       "51044                          1  \n",
       "51045                          0  \n",
       "51046                          1  \n",
       "\n",
       "[51047 rows x 34 columns]"
      ]
     },
     "execution_count": 90,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "customer_num2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "id": "d561be18",
   "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>ChildrenInHH</th>\n",
       "      <th>HandsetRefurbished</th>\n",
       "      <th>HandsetWebCapable</th>\n",
       "      <th>TruckOwner</th>\n",
       "      <th>RVOwner</th>\n",
       "      <th>Homeownership</th>\n",
       "      <th>BuysViaMailOrder</th>\n",
       "      <th>RespondsToMailOffers</th>\n",
       "      <th>OptOutMailings</th>\n",
       "      <th>NonUSTravel</th>\n",
       "      <th>...</th>\n",
       "      <th>HasCreditCard</th>\n",
       "      <th>NewCellphoneUser</th>\n",
       "      <th>NotNewCellphoneUser</th>\n",
       "      <th>OwnsMotorcycle</th>\n",
       "      <th>HandsetPrice</th>\n",
       "      <th>MadeCallToRetentionTeam</th>\n",
       "      <th>CreditRating</th>\n",
       "      <th>PrizmCode</th>\n",
       "      <th>Occupation</th>\n",
       "      <th>MaritalStatus</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>Yes</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>Known</td>\n",
       "      <td>Yes</td>\n",
       "      <td>Yes</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>...</td>\n",
       "      <td>Yes</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>30</td>\n",
       "      <td>Yes</td>\n",
       "      <td>1-Highest</td>\n",
       "      <td>Suburban</td>\n",
       "      <td>Professional</td>\n",
       "      <td>No</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Yes</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>Known</td>\n",
       "      <td>Yes</td>\n",
       "      <td>Yes</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>...</td>\n",
       "      <td>Yes</td>\n",
       "      <td>Yes</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>30</td>\n",
       "      <td>No</td>\n",
       "      <td>4-Medium</td>\n",
       "      <td>Suburban</td>\n",
       "      <td>Professional</td>\n",
       "      <td>Yes</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Yes</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>Unknown</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>...</td>\n",
       "      <td>Yes</td>\n",
       "      <td>Yes</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>Unknown</td>\n",
       "      <td>No</td>\n",
       "      <td>3-Good</td>\n",
       "      <td>Town</td>\n",
       "      <td>Crafts</td>\n",
       "      <td>Yes</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>Yes</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>Known</td>\n",
       "      <td>Yes</td>\n",
       "      <td>Yes</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>...</td>\n",
       "      <td>Yes</td>\n",
       "      <td>Yes</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>10</td>\n",
       "      <td>No</td>\n",
       "      <td>4-Medium</td>\n",
       "      <td>Other</td>\n",
       "      <td>Other</td>\n",
       "      <td>No</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>Known</td>\n",
       "      <td>Yes</td>\n",
       "      <td>Yes</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>...</td>\n",
       "      <td>Yes</td>\n",
       "      <td>No</td>\n",
       "      <td>Yes</td>\n",
       "      <td>No</td>\n",
       "      <td>10</td>\n",
       "      <td>No</td>\n",
       "      <td>1-Highest</td>\n",
       "      <td>Other</td>\n",
       "      <td>Professional</td>\n",
       "      <td>Yes</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>51042</th>\n",
       "      <td>No</td>\n",
       "      <td>Yes</td>\n",
       "      <td>Yes</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>Known</td>\n",
       "      <td>Yes</td>\n",
       "      <td>Yes</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>...</td>\n",
       "      <td>Yes</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>60</td>\n",
       "      <td>No</td>\n",
       "      <td>1-Highest</td>\n",
       "      <td>Suburban</td>\n",
       "      <td>Other</td>\n",
       "      <td>Yes</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>51043</th>\n",
       "      <td>Yes</td>\n",
       "      <td>No</td>\n",
       "      <td>Yes</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>Known</td>\n",
       "      <td>Yes</td>\n",
       "      <td>Yes</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>...</td>\n",
       "      <td>Yes</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>60</td>\n",
       "      <td>No</td>\n",
       "      <td>3-Good</td>\n",
       "      <td>Other</td>\n",
       "      <td>Other</td>\n",
       "      <td>No</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>51044</th>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>Yes</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>Known</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>...</td>\n",
       "      <td>Yes</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>80</td>\n",
       "      <td>No</td>\n",
       "      <td>5-Low</td>\n",
       "      <td>Other</td>\n",
       "      <td>Clerical</td>\n",
       "      <td>No</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>51045</th>\n",
       "      <td>Yes</td>\n",
       "      <td>No</td>\n",
       "      <td>Yes</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>Unknown</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>...</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>30</td>\n",
       "      <td>No</td>\n",
       "      <td>5-Low</td>\n",
       "      <td>Other</td>\n",
       "      <td>Other</td>\n",
       "      <td>No</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>51046</th>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>Yes</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>Unknown</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>...</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>60</td>\n",
       "      <td>Yes</td>\n",
       "      <td>5-Low</td>\n",
       "      <td>Other</td>\n",
       "      <td>Other</td>\n",
       "      <td>Unknown</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>51047 rows × 21 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      ChildrenInHH HandsetRefurbished HandsetWebCapable TruckOwner RVOwner  \\\n",
       "0               No                 No               Yes         No      No   \n",
       "1              Yes                 No                No         No      No   \n",
       "2              Yes                 No                No         No      No   \n",
       "3               No                 No               Yes         No      No   \n",
       "4               No                 No                No         No      No   \n",
       "...            ...                ...               ...        ...     ...   \n",
       "51042           No                Yes               Yes         No      No   \n",
       "51043          Yes                 No               Yes         No      No   \n",
       "51044           No                 No               Yes         No      No   \n",
       "51045          Yes                 No               Yes         No      No   \n",
       "51046           No                 No               Yes         No      No   \n",
       "\n",
       "      Homeownership BuysViaMailOrder RespondsToMailOffers OptOutMailings  \\\n",
       "0             Known              Yes                  Yes             No   \n",
       "1             Known              Yes                  Yes             No   \n",
       "2           Unknown               No                   No             No   \n",
       "3             Known              Yes                  Yes             No   \n",
       "4             Known              Yes                  Yes             No   \n",
       "...             ...              ...                  ...            ...   \n",
       "51042         Known              Yes                  Yes             No   \n",
       "51043         Known              Yes                  Yes             No   \n",
       "51044         Known               No                   No             No   \n",
       "51045       Unknown               No                   No             No   \n",
       "51046       Unknown               No                   No             No   \n",
       "\n",
       "      NonUSTravel  ... HasCreditCard NewCellphoneUser NotNewCellphoneUser  \\\n",
       "0              No  ...           Yes               No                  No   \n",
       "1              No  ...           Yes              Yes                  No   \n",
       "2              No  ...           Yes              Yes                  No   \n",
       "3              No  ...           Yes              Yes                  No   \n",
       "4              No  ...           Yes               No                 Yes   \n",
       "...           ...  ...           ...              ...                 ...   \n",
       "51042          No  ...           Yes               No                  No   \n",
       "51043          No  ...           Yes               No                  No   \n",
       "51044          No  ...           Yes               No                  No   \n",
       "51045          No  ...            No               No                  No   \n",
       "51046          No  ...            No               No                  No   \n",
       "\n",
       "      OwnsMotorcycle HandsetPrice MadeCallToRetentionTeam CreditRating  \\\n",
       "0                 No           30                     Yes    1-Highest   \n",
       "1                 No           30                      No     4-Medium   \n",
       "2                 No      Unknown                      No       3-Good   \n",
       "3                 No           10                      No     4-Medium   \n",
       "4                 No           10                      No    1-Highest   \n",
       "...              ...          ...                     ...          ...   \n",
       "51042             No           60                      No    1-Highest   \n",
       "51043             No           60                      No       3-Good   \n",
       "51044             No           80                      No        5-Low   \n",
       "51045             No           30                      No        5-Low   \n",
       "51046             No           60                     Yes        5-Low   \n",
       "\n",
       "      PrizmCode    Occupation MaritalStatus  \n",
       "0      Suburban  Professional            No  \n",
       "1      Suburban  Professional           Yes  \n",
       "2          Town        Crafts           Yes  \n",
       "3         Other         Other            No  \n",
       "4         Other  Professional           Yes  \n",
       "...         ...           ...           ...  \n",
       "51042  Suburban         Other           Yes  \n",
       "51043     Other         Other            No  \n",
       "51044     Other      Clerical            No  \n",
       "51045     Other         Other            No  \n",
       "51046     Other         Other       Unknown  \n",
       "\n",
       "[51047 rows x 21 columns]"
      ]
     },
     "execution_count": 91,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "customer_cat2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "id": "8847cdb3",
   "metadata": {},
   "outputs": [],
   "source": [
    "z2=customer_cat2\n",
    "#label encoder\n",
    "from sklearn.preprocessing import LabelEncoder\n",
    "label=LabelEncoder()\n",
    "z2=z2.astype(str)\n",
    "z2[cat_customer2] = z2[cat_customer2].apply(lambda col: label.fit_transform(col))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "id": "40a6c941",
   "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>ChildrenInHH</th>\n",
       "      <th>HandsetRefurbished</th>\n",
       "      <th>HandsetWebCapable</th>\n",
       "      <th>TruckOwner</th>\n",
       "      <th>RVOwner</th>\n",
       "      <th>Homeownership</th>\n",
       "      <th>BuysViaMailOrder</th>\n",
       "      <th>RespondsToMailOffers</th>\n",
       "      <th>OptOutMailings</th>\n",
       "      <th>NonUSTravel</th>\n",
       "      <th>...</th>\n",
       "      <th>HasCreditCard</th>\n",
       "      <th>NewCellphoneUser</th>\n",
       "      <th>NotNewCellphoneUser</th>\n",
       "      <th>OwnsMotorcycle</th>\n",
       "      <th>HandsetPrice</th>\n",
       "      <th>MadeCallToRetentionTeam</th>\n",
       "      <th>CreditRating</th>\n",
       "      <th>PrizmCode</th>\n",
       "      <th>Occupation</th>\n",
       "      <th>MaritalStatus</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>8</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>4</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>8</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "      <td>4</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>15</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>51042</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>13</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>51043</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>13</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>51044</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>14</td>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>51045</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>8</td>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>51046</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>13</td>\n",
       "      <td>1</td>\n",
       "      <td>4</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>51047 rows × 21 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       ChildrenInHH  HandsetRefurbished  HandsetWebCapable  TruckOwner  \\\n",
       "0                 0                   0                  1           0   \n",
       "1                 1                   0                  0           0   \n",
       "2                 1                   0                  0           0   \n",
       "3                 0                   0                  1           0   \n",
       "4                 0                   0                  0           0   \n",
       "...             ...                 ...                ...         ...   \n",
       "51042             0                   1                  1           0   \n",
       "51043             1                   0                  1           0   \n",
       "51044             0                   0                  1           0   \n",
       "51045             1                   0                  1           0   \n",
       "51046             0                   0                  1           0   \n",
       "\n",
       "       RVOwner  Homeownership  BuysViaMailOrder  RespondsToMailOffers  \\\n",
       "0            0              0                 1                     1   \n",
       "1            0              0                 1                     1   \n",
       "2            0              1                 0                     0   \n",
       "3            0              0                 1                     1   \n",
       "4            0              0                 1                     1   \n",
       "...        ...            ...               ...                   ...   \n",
       "51042        0              0                 1                     1   \n",
       "51043        0              0                 1                     1   \n",
       "51044        0              0                 0                     0   \n",
       "51045        0              1                 0                     0   \n",
       "51046        0              1                 0                     0   \n",
       "\n",
       "       OptOutMailings  NonUSTravel  ...  HasCreditCard  NewCellphoneUser  \\\n",
       "0                   0            0  ...              1                 0   \n",
       "1                   0            0  ...              1                 1   \n",
       "2                   0            0  ...              1                 1   \n",
       "3                   0            0  ...              1                 1   \n",
       "4                   0            0  ...              1                 0   \n",
       "...               ...          ...  ...            ...               ...   \n",
       "51042               0            0  ...              1                 0   \n",
       "51043               0            0  ...              1                 0   \n",
       "51044               0            0  ...              1                 0   \n",
       "51045               0            0  ...              0                 0   \n",
       "51046               0            0  ...              0                 0   \n",
       "\n",
       "       NotNewCellphoneUser  OwnsMotorcycle  HandsetPrice  \\\n",
       "0                        0               0             8   \n",
       "1                        0               0             8   \n",
       "2                        0               0            15   \n",
       "3                        0               0             0   \n",
       "4                        1               0             0   \n",
       "...                    ...             ...           ...   \n",
       "51042                    0               0            13   \n",
       "51043                    0               0            13   \n",
       "51044                    0               0            14   \n",
       "51045                    0               0             8   \n",
       "51046                    0               0            13   \n",
       "\n",
       "       MadeCallToRetentionTeam  CreditRating  PrizmCode  Occupation  \\\n",
       "0                            1             0          2           4   \n",
       "1                            0             3          2           4   \n",
       "2                            0             2          3           1   \n",
       "3                            0             3          0           3   \n",
       "4                            0             0          0           4   \n",
       "...                        ...           ...        ...         ...   \n",
       "51042                        0             0          2           3   \n",
       "51043                        0             2          0           3   \n",
       "51044                        0             4          0           0   \n",
       "51045                        0             4          0           3   \n",
       "51046                        1             4          0           3   \n",
       "\n",
       "       MaritalStatus  \n",
       "0                  0  \n",
       "1                  2  \n",
       "2                  2  \n",
       "3                  0  \n",
       "4                  2  \n",
       "...              ...  \n",
       "51042              2  \n",
       "51043              0  \n",
       "51044              0  \n",
       "51045              0  \n",
       "51046              1  \n",
       "\n",
       "[51047 rows x 21 columns]"
      ]
     },
     "execution_count": 93,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "z2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "id": "92a1f371",
   "metadata": {},
   "outputs": [],
   "source": [
    "frames2 = [customer_num2, z2]\n",
    "customer_updated2=pd.concat(frames2,axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "id": "7149a3af",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "MonthlyRevenue               0\n",
       "MonthlyMinutes               0\n",
       "TotalRecurringCharge         0\n",
       "DirectorAssistedCalls        0\n",
       "OverageMinutes               0\n",
       "RoamingCalls                 0\n",
       "PercChangeMinutes            0\n",
       "PercChangeRevenues           0\n",
       "DroppedCalls                 0\n",
       "BlockedCalls                 0\n",
       "UnansweredCalls              0\n",
       "CustomerCareCalls            0\n",
       "ThreewayCalls                0\n",
       "ReceivedCalls                0\n",
       "OutboundCalls                0\n",
       "InboundCalls                 0\n",
       "PeakCallsInOut               0\n",
       "OffPeakCallsInOut            0\n",
       "DroppedBlockedCalls          0\n",
       "CallForwardingCalls          0\n",
       "CallWaitingCalls             0\n",
       "MonthsInService              0\n",
       "UniqueSubs                   0\n",
       "ActiveSubs                   0\n",
       "Handsets                     0\n",
       "HandsetModels                0\n",
       "CurrentEquipmentDays         0\n",
       "AgeHH1                       0\n",
       "AgeHH2                       0\n",
       "RetentionCalls               0\n",
       "RetentionOffersAccepted      0\n",
       "ReferralsMadeBySubscriber    0\n",
       "IncomeGroup                  0\n",
       "AdjustmentsToCreditRating    0\n",
       "ChildrenInHH                 0\n",
       "HandsetRefurbished           0\n",
       "HandsetWebCapable            0\n",
       "TruckOwner                   0\n",
       "RVOwner                      0\n",
       "Homeownership                0\n",
       "BuysViaMailOrder             0\n",
       "RespondsToMailOffers         0\n",
       "OptOutMailings               0\n",
       "NonUSTravel                  0\n",
       "OwnsComputer                 0\n",
       "HasCreditCard                0\n",
       "NewCellphoneUser             0\n",
       "NotNewCellphoneUser          0\n",
       "OwnsMotorcycle               0\n",
       "HandsetPrice                 0\n",
       "MadeCallToRetentionTeam      0\n",
       "CreditRating                 0\n",
       "PrizmCode                    0\n",
       "Occupation                   0\n",
       "MaritalStatus                0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 95,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "customer_updated2.isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "id": "41135168",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    36336\n",
       "1    14711\n",
       "Name: Churn, dtype: int64"
      ]
     },
     "execution_count": 96,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_label2.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "id": "bb543712",
   "metadata": {},
   "outputs": [],
   "source": [
    "#applying SVM smote \n",
    "oversampler=SMOTE()\n",
    "customer_updated2,train_label2=oversampler.fit_resample(customer_updated2,train_label2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "id": "b1f34ee3",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1    36336\n",
       "0    36336\n",
       "Name: Churn, dtype: int64"
      ]
     },
     "execution_count": 98,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_label2.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "id": "58666c43",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Method2: on second dataset.\n",
    "corr2=customer_updated2.corr(method='spearman')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "id": "defc7347",
   "metadata": {},
   "outputs": [],
   "source": [
    "data2=customer_updated2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "id": "85284cfd",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x576 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(12, 8))\n",
    "\n",
    "distance_matrix = 1 - np.abs(corr2)\n",
    "dist_linkage = hierarchy.ward(squareform(distance_matrix))\n",
    "dendro = hierarchy.dendrogram(\n",
    "    dist_linkage, labels=data2.columns.tolist(), ax=ax1, leaf_rotation=90\n",
    ")\n",
    "dendro_idx = np.arange(0, len(dendro[\"ivl\"]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "id": "ef286922",
   "metadata": {},
   "outputs": [],
   "source": [
    "cluster_ids = hierarchy.fcluster(dist_linkage, 1, criterion=\"distance\")\n",
    "cluster_id_to_feature_ids = defaultdict(list)\n",
    "for idx, cluster_id in enumerate(cluster_ids):\n",
    "    cluster_id_to_feature_ids[cluster_id].append(idx)\n",
    "selected_features2 = [v[0] for v in cluster_id_to_feature_ids.values()]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "id": "0dc8535c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[0, 5, 6, 19, 21, 22, 24, 27, 29, 33, 34, 37]"
      ]
     },
     "execution_count": 103,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "selected_features2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "id": "d2d5239b",
   "metadata": {},
   "outputs": [],
   "source": [
    "r2=customer_updated2.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "id": "0911c12d",
   "metadata": {},
   "outputs": [],
   "source": [
    "r2=r2[selected_features2]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "id": "79e57067",
   "metadata": {},
   "outputs": [],
   "source": [
    "z2=customer_updated2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "id": "d42b7fbf",
   "metadata": {},
   "outputs": [],
   "source": [
    "z2=z2[r2]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "id": "bccc1861",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "X_train2, X_test2, y_train2, y_test2 = train_test_split(z2, train_label2, test_size=0.3, random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "id": "b91450cf",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.3797758064063115\n"
     ]
    }
   ],
   "source": [
    "#Implementing the models after Hirarchial feature selection on dataset 2.\n",
    "xg=xgboost.XGBRegressor()\n",
    "xg.fit(X_train2,y_train2)\n",
    "k_folds = KFold(n_splits = 10)\n",
    "print(cross_val_score(xg, X_train2, y_train2, cv=k_folds).mean())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "id": "d197a6e9",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8357970901469832"
      ]
     },
     "execution_count": 110,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics import classification_report\n",
    "pred_xg=xg.predict(X_test2)\n",
    "fpr, tpr, thresholds = metrics.roc_curve(y_test2, pred_xg)\n",
    "#r=classification_report(y_test2, pred_xg)\n",
    "xg_m=metrics.auc(fpr, tpr)\n",
    "xg_m"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 111,
   "id": "a2f94211",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.5318428571428572\n"
     ]
    }
   ],
   "source": [
    "#decision treee\n",
    "dt=DecisionTreeClassifier()\n",
    "dt.fit(X_train2,y_train2)\n",
    "k_folds = KFold(n_splits = 10)\n",
    "print(cross_val_score(dt, X_train, y_train, cv=k_folds).mean())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 112,
   "id": "51c08ba2",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.6754511540886704"
      ]
     },
     "execution_count": 112,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred2=dt.predict(X_test2)\n",
    "fpr, tpr, thresholds = metrics.roc_curve(y_test2, pred2)\n",
    "dt_m=metrics.auc(fpr, tpr)\n",
    "dt_m"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 113,
   "id": "d460ef1a",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.7327698053862787\n"
     ]
    }
   ],
   "source": [
    "#random Forest\n",
    "rf=RandomForestClassifier()\n",
    "rf.fit(X_train2,y_train2)\n",
    "k_folds = KFold(n_splits = 10)\n",
    "print(cross_val_score(rf, X_train2, y_train2, cv=k_folds).mean())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 114,
   "id": "52684a91",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7378061493435866"
      ]
     },
     "execution_count": 114,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_rf=rf.predict(X_test2)\n",
    "fpr, tpr, thresholds = metrics.roc_curve(y_test2, pred_rf)\n",
    "rf_m=metrics.auc(fpr, tpr)\n",
    "rf_m"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 115,
   "id": "2690be85",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.774071161784942\n"
     ]
    }
   ],
   "source": [
    "gbm=lightgbm.LGBMClassifier()\n",
    "gbm.fit(X_train2,y_train2)\n",
    "k_folds = KFold(n_splits = 10)\n",
    "print(cross_val_score(gbm, X_train2, y_train2, cv=k_folds).mean())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 116,
   "id": "953d3b6d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7758557798804185"
      ]
     },
     "execution_count": 116,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_gbm=gbm.predict(X_test2)\n",
    "fpr, tpr, thresholds = metrics.roc_curve(y_test2, pred_gbm)\n",
    "gbm_m=metrics.auc(fpr, tpr)\n",
    "gbm_m"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 117,
   "id": "8ae12533",
   "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>MonthlyRevenue</th>\n",
       "      <th>MonthlyMinutes</th>\n",
       "      <th>TotalRecurringCharge</th>\n",
       "      <th>DirectorAssistedCalls</th>\n",
       "      <th>OverageMinutes</th>\n",
       "      <th>RoamingCalls</th>\n",
       "      <th>PercChangeMinutes</th>\n",
       "      <th>PercChangeRevenues</th>\n",
       "      <th>DroppedCalls</th>\n",
       "      <th>BlockedCalls</th>\n",
       "      <th>...</th>\n",
       "      <th>HasCreditCard</th>\n",
       "      <th>NewCellphoneUser</th>\n",
       "      <th>NotNewCellphoneUser</th>\n",
       "      <th>OwnsMotorcycle</th>\n",
       "      <th>HandsetPrice</th>\n",
       "      <th>MadeCallToRetentionTeam</th>\n",
       "      <th>CreditRating</th>\n",
       "      <th>PrizmCode</th>\n",
       "      <th>Occupation</th>\n",
       "      <th>MaritalStatus</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>24.000000</td>\n",
       "      <td>219.000000</td>\n",
       "      <td>22.000000</td>\n",
       "      <td>0.250000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>-157.000000</td>\n",
       "      <td>-19.000000</td>\n",
       "      <td>0.700000</td>\n",
       "      <td>0.700000</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>8</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>4</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>16.990000</td>\n",
       "      <td>10.000000</td>\n",
       "      <td>17.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>-4.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.300000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>8</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "      <td>4</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>38.000000</td>\n",
       "      <td>8.000000</td>\n",
       "      <td>38.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>-2.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>15</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>82.280000</td>\n",
       "      <td>1312.000000</td>\n",
       "      <td>75.000000</td>\n",
       "      <td>1.240000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>157.000000</td>\n",
       "      <td>8.100000</td>\n",
       "      <td>52.000000</td>\n",
       "      <td>7.700000</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>17.140000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>17.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>-0.200000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>72667</th>\n",
       "      <td>76.406716</td>\n",
       "      <td>246.569562</td>\n",
       "      <td>30.403581</td>\n",
       "      <td>0.199552</td>\n",
       "      <td>124.726590</td>\n",
       "      <td>0.558747</td>\n",
       "      <td>125.017905</td>\n",
       "      <td>39.506267</td>\n",
       "      <td>2.703581</td>\n",
       "      <td>4.610331</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>15</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>72668</th>\n",
       "      <td>76.070898</td>\n",
       "      <td>1263.891883</td>\n",
       "      <td>58.674397</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>59.194294</td>\n",
       "      <td>0.260232</td>\n",
       "      <td>-322.530241</td>\n",
       "      <td>-20.362017</td>\n",
       "      <td>10.473488</td>\n",
       "      <td>3.172328</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>8</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>72669</th>\n",
       "      <td>37.296536</td>\n",
       "      <td>459.694376</td>\n",
       "      <td>38.842706</td>\n",
       "      <td>0.096068</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>45.296661</td>\n",
       "      <td>-0.323146</td>\n",
       "      <td>6.541564</td>\n",
       "      <td>6.378647</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>12</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>72670</th>\n",
       "      <td>67.966507</td>\n",
       "      <td>293.651909</td>\n",
       "      <td>27.652127</td>\n",
       "      <td>2.393919</td>\n",
       "      <td>101.565176</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>-288.652272</td>\n",
       "      <td>-47.091327</td>\n",
       "      <td>5.839142</td>\n",
       "      <td>0.460858</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>14</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>72671</th>\n",
       "      <td>40.695022</td>\n",
       "      <td>377.856819</td>\n",
       "      <td>32.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>20.000000</td>\n",
       "      <td>0.257892</td>\n",
       "      <td>25.008458</td>\n",
       "      <td>18.624241</td>\n",
       "      <td>1.380834</td>\n",
       "      <td>9.008851</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>14</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>72672 rows × 55 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       MonthlyRevenue  MonthlyMinutes  TotalRecurringCharge  \\\n",
       "0           24.000000      219.000000             22.000000   \n",
       "1           16.990000       10.000000             17.000000   \n",
       "2           38.000000        8.000000             38.000000   \n",
       "3           82.280000     1312.000000             75.000000   \n",
       "4           17.140000        0.000000             17.000000   \n",
       "...               ...             ...                   ...   \n",
       "72667       76.406716      246.569562             30.403581   \n",
       "72668       76.070898     1263.891883             58.674397   \n",
       "72669       37.296536      459.694376             38.842706   \n",
       "72670       67.966507      293.651909             27.652127   \n",
       "72671       40.695022      377.856819             32.000000   \n",
       "\n",
       "       DirectorAssistedCalls  OverageMinutes  RoamingCalls  PercChangeMinutes  \\\n",
       "0                   0.250000        0.000000      0.000000        -157.000000   \n",
       "1                   0.000000        0.000000      0.000000          -4.000000   \n",
       "2                   0.000000        0.000000      0.000000          -2.000000   \n",
       "3                   1.240000        0.000000      0.000000         157.000000   \n",
       "4                   0.000000        0.000000      0.000000           0.000000   \n",
       "...                      ...             ...           ...                ...   \n",
       "72667               0.199552      124.726590      0.558747         125.017905   \n",
       "72668               0.500000       59.194294      0.260232        -322.530241   \n",
       "72669               0.096068        0.000000      0.000000          45.296661   \n",
       "72670               2.393919      101.565176      0.000000        -288.652272   \n",
       "72671               0.000000       20.000000      0.257892          25.008458   \n",
       "\n",
       "       PercChangeRevenues  DroppedCalls  BlockedCalls  ...  HasCreditCard  \\\n",
       "0              -19.000000      0.700000      0.700000  ...              1   \n",
       "1                0.000000      0.300000      0.000000  ...              1   \n",
       "2                0.000000      0.000000      0.000000  ...              1   \n",
       "3                8.100000     52.000000      7.700000  ...              1   \n",
       "4               -0.200000      0.000000      0.000000  ...              1   \n",
       "...                   ...           ...           ...  ...            ...   \n",
       "72667           39.506267      2.703581      4.610331  ...              1   \n",
       "72668          -20.362017     10.473488      3.172328  ...              0   \n",
       "72669           -0.323146      6.541564      6.378647  ...              0   \n",
       "72670          -47.091327      5.839142      0.460858  ...              1   \n",
       "72671           18.624241      1.380834      9.008851  ...              0   \n",
       "\n",
       "       NewCellphoneUser  NotNewCellphoneUser  OwnsMotorcycle  HandsetPrice  \\\n",
       "0                     0                    0               0             8   \n",
       "1                     1                    0               0             8   \n",
       "2                     1                    0               0            15   \n",
       "3                     1                    0               0             0   \n",
       "4                     0                    1               0             0   \n",
       "...                 ...                  ...             ...           ...   \n",
       "72667                 0                    0               0            15   \n",
       "72668                 0                    0               0             8   \n",
       "72669                 0                    1               0            12   \n",
       "72670                 0                    0               0            14   \n",
       "72671                 0                    0               0            14   \n",
       "\n",
       "       MadeCallToRetentionTeam  CreditRating  PrizmCode  Occupation  \\\n",
       "0                            1             0          2           4   \n",
       "1                            0             3          2           4   \n",
       "2                            0             2          3           1   \n",
       "3                            0             3          0           3   \n",
       "4                            0             0          0           4   \n",
       "...                        ...           ...        ...         ...   \n",
       "72667                        0             0          0           3   \n",
       "72668                        0             2          3           3   \n",
       "72669                        0             1          1           3   \n",
       "72670                        0             2          1           2   \n",
       "72671                        0             1          0           3   \n",
       "\n",
       "       MaritalStatus  \n",
       "0                  0  \n",
       "1                  2  \n",
       "2                  2  \n",
       "3                  0  \n",
       "4                  2  \n",
       "...              ...  \n",
       "72667              0  \n",
       "72668              0  \n",
       "72669              1  \n",
       "72670              0  \n",
       "72671              1  \n",
       "\n",
       "[72672 rows x 55 columns]"
      ]
     },
     "execution_count": 117,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "customer_updated2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 118,
   "id": "31623c00",
   "metadata": {},
   "outputs": [],
   "source": [
    "p2 = customer_updated2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 119,
   "id": "041f6150",
   "metadata": {},
   "outputs": [],
   "source": [
    "pca_2 = PCA(n_components=51)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 120,
   "id": "dbcb8b92",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "PCA(n_components=51)"
      ]
     },
     "execution_count": 120,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pca_2.fit(p2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 121,
   "id": "6eeaa3af",
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pca2 = pca_2.transform(p2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 122,
   "id": "e4cfd6c9",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(72672, 51)"
      ]
     },
     "execution_count": 122,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_pca2.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 123,
   "id": "19427da3",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "X_train, X_test, y_train, y_test = train_test_split(y_pca2, train_label2, test_size=0.3, random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 124,
   "id": "b905dc9e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.6304501670925889\n"
     ]
    }
   ],
   "source": [
    "#1.desicion tree. \n",
    "dt.fit(X_train,y_train)\n",
    "k_folds = KFold(n_splits = 10)\n",
    "print(cross_val_score(dt, X_train, y_train, cv=k_folds).mean())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 125,
   "id": "12565114",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.6354857544773406"
      ]
     },
     "execution_count": 125,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_dt = dt.predict(X_test)\n",
    "fpr_dt, tpr_dt, thresholds_dt = metrics.roc_curve(y_test, pred_dt)\n",
    "dt_m = metrics.auc(fpr_dt, tpr_dt)\n",
    "dt_m"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 126,
   "id": "9f7a8739",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.7383723216040889\n"
     ]
    }
   ],
   "source": [
    "#2.Random Forest.\n",
    "rf.fit(X_train,y_train)\n",
    "k_folds = KFold(n_splits = 10)\n",
    "print(cross_val_score(rf, X_train,y_train, cv=k_folds).mean())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 127,
   "id": "105e927f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.740700683748026"
      ]
     },
     "execution_count": 127,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_rf = rf.predict(X_test)\n",
    "fpr_rf, tpr_rf, thresholds_rf = metrics.roc_curve(y_test, pred_rf)\n",
    "rf_m = metrics.auc(fpr_rf, tpr_rf)\n",
    "rf_m"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 128,
   "id": "12dd0a1d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.2634059777432035\n"
     ]
    }
   ],
   "source": [
    "#3.XGBoost.\n",
    "xg.fit(X_train,y_train)\n",
    "k_folds = KFold(n_splits = 10)\n",
    "print(cross_val_score(xg, X_train,y_train, cv=k_folds).mean())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 129,
   "id": "641aa2dc",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8026886735620872"
      ]
     },
     "execution_count": 129,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_xg = xg.predict(X_test)\n",
    "fpr_xg, tpr_xg, thresholds_xg = metrics.roc_curve(y_test, pred_xg)\n",
    "xg_m = metrics.auc(fpr_xg, tpr_xg)\n",
    "xg_m"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 130,
   "id": "eb9d81ab",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.7287202673481423\n"
     ]
    }
   ],
   "source": [
    "#4.Light GBM.\n",
    "gbm.fit(X_train,y_train)\n",
    "k_folds = KFold(n_splits = 10)\n",
    "print(cross_val_score(gbm, X_train,y_train, cv=k_folds).mean())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 131,
   "id": "d17f7d42",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7282947873269069"
      ]
     },
     "execution_count": 131,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_gbm = gbm.predict(X_test)\n",
    "fpr_gbm, tpr_gbm, thresholds_gbm = metrics.roc_curve(y_test, pred_gbm)\n",
    "gbm_m = metrics.auc(fpr_gbm, tpr_gbm)\n",
    "gbm_m"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "409ece6f",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6efe3941",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c17065da",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 132,
   "id": "ef84377a",
   "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>rev_Mean</th>\n",
       "      <th>mou_Mean</th>\n",
       "      <th>totmrc_Mean</th>\n",
       "      <th>da_Mean</th>\n",
       "      <th>ovrmou_Mean</th>\n",
       "      <th>ovrrev_Mean</th>\n",
       "      <th>vceovr_Mean</th>\n",
       "      <th>datovr_Mean</th>\n",
       "      <th>roam_Mean</th>\n",
       "      <th>drop_vce_Mean</th>\n",
       "      <th>...</th>\n",
       "      <th>infobase</th>\n",
       "      <th>HHstatin</th>\n",
       "      <th>dwllsize</th>\n",
       "      <th>ethnic</th>\n",
       "      <th>kid0_2</th>\n",
       "      <th>kid3_5</th>\n",
       "      <th>kid6_10</th>\n",
       "      <th>kid11_15</th>\n",
       "      <th>kid16_17</th>\n",
       "      <th>creditcd</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.556517</td>\n",
       "      <td>0.017520</td>\n",
       "      <td>-0.02246</td>\n",
       "      <td>-0.021778</td>\n",
       "      <td>-0.021308</td>\n",
       "      <td>-0.039820</td>\n",
       "      <td>0.015799</td>\n",
       "      <td>0.010703</td>\n",
       "      <td>-0.011647</td>\n",
       "      <td>...</td>\n",
       "      <td>0.001484</td>\n",
       "      <td>0.009292</td>\n",
       "      <td>0.001218</td>\n",
       "      <td>-0.030494</td>\n",
       "      <td>-0.003983</td>\n",
       "      <td>-0.002123</td>\n",
       "      <td>0.001138</td>\n",
       "      <td>0.002035</td>\n",
       "      <td>-0.002917</td>\n",
       "      <td>-0.022795</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.017520</td>\n",
       "      <td>-0.164801</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>-0.02246</td>\n",
       "      <td>0.045300</td>\n",
       "      <td>-0.107643</td>\n",
       "      <td>-0.112533</td>\n",
       "      <td>0.015799</td>\n",
       "      <td>0.010703</td>\n",
       "      <td>-0.060797</td>\n",
       "      <td>...</td>\n",
       "      <td>0.001484</td>\n",
       "      <td>0.009292</td>\n",
       "      <td>0.001218</td>\n",
       "      <td>-0.400201</td>\n",
       "      <td>-0.003983</td>\n",
       "      <td>-0.002123</td>\n",
       "      <td>0.001138</td>\n",
       "      <td>0.002035</td>\n",
       "      <td>-0.002917</td>\n",
       "      <td>-0.022795</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>-0.015816</td>\n",
       "      <td>0.751490</td>\n",
       "      <td>0.094113</td>\n",
       "      <td>0.04202</td>\n",
       "      <td>-0.021778</td>\n",
       "      <td>-0.021308</td>\n",
       "      <td>-0.039820</td>\n",
       "      <td>0.015799</td>\n",
       "      <td>0.010703</td>\n",
       "      <td>-0.028216</td>\n",
       "      <td>...</td>\n",
       "      <td>0.001484</td>\n",
       "      <td>0.009292</td>\n",
       "      <td>0.001218</td>\n",
       "      <td>-0.030494</td>\n",
       "      <td>-0.003983</td>\n",
       "      <td>0.043076</td>\n",
       "      <td>0.001138</td>\n",
       "      <td>0.002035</td>\n",
       "      <td>-0.002917</td>\n",
       "      <td>-0.022795</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>-1.081092</td>\n",
       "      <td>0.199842</td>\n",
       "      <td>0.017520</td>\n",
       "      <td>0.04202</td>\n",
       "      <td>-0.021778</td>\n",
       "      <td>-0.021308</td>\n",
       "      <td>-0.039820</td>\n",
       "      <td>0.015799</td>\n",
       "      <td>0.010703</td>\n",
       "      <td>0.175511</td>\n",
       "      <td>...</td>\n",
       "      <td>0.001484</td>\n",
       "      <td>0.009292</td>\n",
       "      <td>0.052330</td>\n",
       "      <td>0.018434</td>\n",
       "      <td>0.095326</td>\n",
       "      <td>-0.002123</td>\n",
       "      <td>0.001138</td>\n",
       "      <td>0.002035</td>\n",
       "      <td>-0.002917</td>\n",
       "      <td>-0.022795</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>-0.270162</td>\n",
       "      <td>-0.339154</td>\n",
       "      <td>-0.136630</td>\n",
       "      <td>0.04202</td>\n",
       "      <td>-0.021778</td>\n",
       "      <td>-0.021308</td>\n",
       "      <td>-0.039820</td>\n",
       "      <td>0.015799</td>\n",
       "      <td>0.010703</td>\n",
       "      <td>0.063701</td>\n",
       "      <td>...</td>\n",
       "      <td>0.001484</td>\n",
       "      <td>0.009292</td>\n",
       "      <td>-0.037101</td>\n",
       "      <td>0.066336</td>\n",
       "      <td>-0.003983</td>\n",
       "      <td>-0.002123</td>\n",
       "      <td>0.001138</td>\n",
       "      <td>0.002035</td>\n",
       "      <td>-0.002917</td>\n",
       "      <td>-0.022795</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 96 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   rev_Mean  mou_Mean  totmrc_Mean  da_Mean  ovrmou_Mean  ovrrev_Mean  \\\n",
       "0  0.000000  0.556517     0.017520 -0.02246    -0.021778    -0.021308   \n",
       "1  0.017520 -0.164801     0.000000 -0.02246     0.045300    -0.107643   \n",
       "2 -0.015816  0.751490     0.094113  0.04202    -0.021778    -0.021308   \n",
       "3 -1.081092  0.199842     0.017520  0.04202    -0.021778    -0.021308   \n",
       "4 -0.270162 -0.339154    -0.136630  0.04202    -0.021778    -0.021308   \n",
       "\n",
       "   vceovr_Mean  datovr_Mean  roam_Mean  drop_vce_Mean  ...  infobase  \\\n",
       "0    -0.039820     0.015799   0.010703      -0.011647  ...  0.001484   \n",
       "1    -0.112533     0.015799   0.010703      -0.060797  ...  0.001484   \n",
       "2    -0.039820     0.015799   0.010703      -0.028216  ...  0.001484   \n",
       "3    -0.039820     0.015799   0.010703       0.175511  ...  0.001484   \n",
       "4    -0.039820     0.015799   0.010703       0.063701  ...  0.001484   \n",
       "\n",
       "   HHstatin  dwllsize    ethnic    kid0_2    kid3_5   kid6_10  kid11_15  \\\n",
       "0  0.009292  0.001218 -0.030494 -0.003983 -0.002123  0.001138  0.002035   \n",
       "1  0.009292  0.001218 -0.400201 -0.003983 -0.002123  0.001138  0.002035   \n",
       "2  0.009292  0.001218 -0.030494 -0.003983  0.043076  0.001138  0.002035   \n",
       "3  0.009292  0.052330  0.018434  0.095326 -0.002123  0.001138  0.002035   \n",
       "4  0.009292 -0.037101  0.066336 -0.003983 -0.002123  0.001138  0.002035   \n",
       "\n",
       "   kid16_17  creditcd  \n",
       "0 -0.002917 -0.022795  \n",
       "1 -0.002917 -0.022795  \n",
       "2 -0.002917 -0.022795  \n",
       "3 -0.002917 -0.022795  \n",
       "4 -0.002917 -0.022795  \n",
       "\n",
       "[5 rows x 96 columns]"
      ]
     },
     "execution_count": 132,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Method 3: Weight of evidence (best method ) for first dataset\n",
    "from mlencoders.weight_of_evidence_encoder import WeightOfEvidenceEncoder\n",
    "\n",
    "enc = WeightOfEvidenceEncoder(cols=customer_updated.columns)\n",
    "customer_updated = enc.fit_transform(customer_updated, train_label)\n",
    "customer_updated.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 133,
   "id": "36d2be05",
   "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>rev_Mean</th>\n",
       "      <th>mou_Mean</th>\n",
       "      <th>totmrc_Mean</th>\n",
       "      <th>da_Mean</th>\n",
       "      <th>ovrmou_Mean</th>\n",
       "      <th>ovrrev_Mean</th>\n",
       "      <th>vceovr_Mean</th>\n",
       "      <th>datovr_Mean</th>\n",
       "      <th>roam_Mean</th>\n",
       "      <th>drop_vce_Mean</th>\n",
       "      <th>...</th>\n",
       "      <th>infobase</th>\n",
       "      <th>HHstatin</th>\n",
       "      <th>dwllsize</th>\n",
       "      <th>ethnic</th>\n",
       "      <th>kid0_2</th>\n",
       "      <th>kid3_5</th>\n",
       "      <th>kid6_10</th>\n",
       "      <th>kid11_15</th>\n",
       "      <th>kid16_17</th>\n",
       "      <th>creditcd</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.556517</td>\n",
       "      <td>0.017520</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.015799</td>\n",
       "      <td>0.010703</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.001484</td>\n",
       "      <td>0.009292</td>\n",
       "      <td>0.001218</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.001138</td>\n",
       "      <td>0.002035</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.01752</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.0453</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.015799</td>\n",
       "      <td>0.010703</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.001484</td>\n",
       "      <td>0.009292</td>\n",
       "      <td>0.001218</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.001138</td>\n",
       "      <td>0.002035</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.751490</td>\n",
       "      <td>0.094113</td>\n",
       "      <td>0.04202</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.015799</td>\n",
       "      <td>0.010703</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.001484</td>\n",
       "      <td>0.009292</td>\n",
       "      <td>0.001218</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.043076</td>\n",
       "      <td>0.001138</td>\n",
       "      <td>0.002035</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.199842</td>\n",
       "      <td>0.017520</td>\n",
       "      <td>0.04202</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.015799</td>\n",
       "      <td>0.010703</td>\n",
       "      <td>0.175511</td>\n",
       "      <td>...</td>\n",
       "      <td>0.001484</td>\n",
       "      <td>0.009292</td>\n",
       "      <td>0.052330</td>\n",
       "      <td>0.018434</td>\n",
       "      <td>0.095326</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.001138</td>\n",
       "      <td>0.002035</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.04202</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.015799</td>\n",
       "      <td>0.010703</td>\n",
       "      <td>0.063701</td>\n",
       "      <td>...</td>\n",
       "      <td>0.001484</td>\n",
       "      <td>0.009292</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.066336</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.001138</td>\n",
       "      <td>0.002035</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 96 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   rev_Mean  mou_Mean  totmrc_Mean  da_Mean  ovrmou_Mean  ovrrev_Mean  \\\n",
       "0   0.00000  0.556517     0.017520  0.00000       0.0000          0.0   \n",
       "1   0.01752  0.000000     0.000000  0.00000       0.0453          0.0   \n",
       "2   0.00000  0.751490     0.094113  0.04202       0.0000          0.0   \n",
       "3   0.00000  0.199842     0.017520  0.04202       0.0000          0.0   \n",
       "4   0.00000  0.000000     0.000000  0.04202       0.0000          0.0   \n",
       "\n",
       "   vceovr_Mean  datovr_Mean  roam_Mean  drop_vce_Mean  ...  infobase  \\\n",
       "0          0.0     0.015799   0.010703       0.000000  ...  0.001484   \n",
       "1          0.0     0.015799   0.010703       0.000000  ...  0.001484   \n",
       "2          0.0     0.015799   0.010703       0.000000  ...  0.001484   \n",
       "3          0.0     0.015799   0.010703       0.175511  ...  0.001484   \n",
       "4          0.0     0.015799   0.010703       0.063701  ...  0.001484   \n",
       "\n",
       "   HHstatin  dwllsize    ethnic    kid0_2    kid3_5   kid6_10  kid11_15  \\\n",
       "0  0.009292  0.001218  0.000000  0.000000  0.000000  0.001138  0.002035   \n",
       "1  0.009292  0.001218  0.000000  0.000000  0.000000  0.001138  0.002035   \n",
       "2  0.009292  0.001218  0.000000  0.000000  0.043076  0.001138  0.002035   \n",
       "3  0.009292  0.052330  0.018434  0.095326  0.000000  0.001138  0.002035   \n",
       "4  0.009292  0.000000  0.066336  0.000000  0.000000  0.001138  0.002035   \n",
       "\n",
       "   kid16_17  creditcd  \n",
       "0       0.0       0.0  \n",
       "1       0.0       0.0  \n",
       "2       0.0       0.0  \n",
       "3       0.0       0.0  \n",
       "4       0.0       0.0  \n",
       "\n",
       "[5 rows x 96 columns]"
      ]
     },
     "execution_count": 133,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "customer_updated[customer_updated < 0]=0\n",
    "customer_updated.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 134,
   "id": "8e5ab8b4",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    50438\n",
       "1    49562\n",
       "Name: churn, dtype: int64"
      ]
     },
     "execution_count": 134,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_label.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 135,
   "id": "addb25f1",
   "metadata": {},
   "outputs": [],
   "source": [
    "#applying SVM smote \n",
    "oversampler=SMOTE()\n",
    "customer_updated,train_label=oversampler.fit_resample(customer_updated,train_label)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 136,
   "id": "e220f20f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1    50438\n",
       "0    50438\n",
       "Name: churn, dtype: int64"
      ]
     },
     "execution_count": 136,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_label.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 137,
   "id": "76c17c24",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(100876, 57)"
      ]
     },
     "execution_count": 137,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#feature Selection\n",
    "from sklearn.feature_selection import SelectPercentile, chi2\n",
    "customer_updated = SelectPercentile(chi2, percentile=60).fit_transform(customer_updated, train_label)\n",
    "customer_updated.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 138,
   "id": "eafe42b5",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "X_train, X_test, y_train, y_test = train_test_split(customer_updated, train_label, test_size=0.3, random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 139,
   "id": "d57a0faf",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.tree import DecisionTreeClassifier\n",
    "from sklearn.metrics import roc_auc_score\n",
    "from sklearn.model_selection import KFold\n",
    "from sklearn.model_selection import cross_val_score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 140,
   "id": "29b39462",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.8118334880116391\n"
     ]
    }
   ],
   "source": [
    "xg=xgboost.XGBClassifier()\n",
    "xg.fit(X_train,y_train)\n",
    "k_folds = KFold(n_splits = 10)\n",
    "print(cross_val_score(xg, X_train, y_train, cv=k_folds).mean())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 141,
   "id": "2869d6cd",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8137283310890652"
      ]
     },
     "execution_count": 141,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_xg=xg.predict(X_test)\n",
    "fpr_x1, tpr_x1, thresholds_x1 = metrics.roc_curve(y_test, pred_xg)\n",
    "xg_m=metrics.auc(fpr_x1, tpr_x1)\n",
    "xg_m"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 142,
   "id": "e3054f21",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.6820414195333291\n"
     ]
    }
   ],
   "source": [
    "dt=DecisionTreeClassifier()\n",
    "dt.fit(X_train,y_train)\n",
    "k_folds = KFold(n_splits = 10)\n",
    "print(cross_val_score(dt, X_train, y_train, cv=k_folds).mean())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 143,
   "id": "eef80f2b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.6844271205783846"
      ]
     },
     "execution_count": 143,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred=dt.predict(X_test)\n",
    "fpr_d1, tpr_d1, thresholds_d1 = metrics.roc_curve(y_test, pred)\n",
    "dt_m=metrics.auc(fpr_d1, tpr_d1)\n",
    "dt_m"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 144,
   "id": "2ff30ff9",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.7988331263535856\n"
     ]
    }
   ],
   "source": [
    "#random Forest\n",
    "rf=RandomForestClassifier()\n",
    "rf.fit(X_train,y_train)\n",
    "k_folds = KFold(n_splits = 10)\n",
    "print(cross_val_score(rf, X_train, y_train, cv=k_folds).mean())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 145,
   "id": "31207286",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8002121249254094"
      ]
     },
     "execution_count": 145,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_rf=rf.predict(X_test)\n",
    "fpr_r1, tpr_r1, thresholds_r1 = metrics.roc_curve(y_test, pred_rf)\n",
    "rf_m=metrics.auc(fpr_r1, tpr_r1)\n",
    "rf_m"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 146,
   "id": "84783ee8",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.8144675494620633\n"
     ]
    }
   ],
   "source": [
    "#lightbgm\n",
    "gbm=lightgbm.LGBMClassifier()\n",
    "gbm.fit(X_train,y_train)\n",
    "k_folds = KFold(n_splits = 10)\n",
    "print(cross_val_score(gbm, X_train, y_train, cv=k_folds).mean())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 147,
   "id": "85dd8479",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8125740555222799"
      ]
     },
     "execution_count": 147,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_gbm=gbm.predict(X_test)\n",
    "fpr_l1, tpr_l1, thresholds_l1 = metrics.roc_curve(y_test, pred_gbm)\n",
    "gbm_m=metrics.auc(fpr_l1, tpr_l1)\n",
    "gbm_m"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 148,
   "id": "8fe9e775",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "StackingClassifier(estimators=[('xg',\n",
       "                                XGBClassifier(base_score=None, booster=None,\n",
       "                                              callbacks=None,\n",
       "                                              colsample_bylevel=None,\n",
       "                                              colsample_bynode=None,\n",
       "                                              colsample_bytree=None,\n",
       "                                              early_stopping_rounds=None,\n",
       "                                              enable_categorical=False,\n",
       "                                              eval_metric=None,\n",
       "                                              feature_types=None, gamma=None,\n",
       "                                              gpu_id=None, grow_policy=None,\n",
       "                                              importance_type=None,\n",
       "                                              interaction_constraints=None,\n",
       "                                              learning_r...\n",
       "                                              max_cat_to_onehot=None,\n",
       "                                              max_delta_step=None,\n",
       "                                              max_depth=None, max_leaves=None,\n",
       "                                              min_child_weight=None,\n",
       "                                              missing=nan,\n",
       "                                              monotone_constraints=None,\n",
       "                                              n_estimators=100, n_jobs=None,\n",
       "                                              num_parallel_tree=None,\n",
       "                                              predictor=None, random_state=None, ...)),\n",
       "                               ('dt', DecisionTreeClassifier()),\n",
       "                               ('rf', RandomForestClassifier()),\n",
       "                               ('gbm', LGBMClassifier())],\n",
       "                   final_estimator=LogisticRegression())"
      ]
     },
     "execution_count": 148,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Implementing the stacking Ensemble for the first dataset.\n",
    "from sklearn.ensemble import StackingClassifier\n",
    "estimators = [('xg',xgboost.XGBClassifier()),('dt',DecisionTreeClassifier()),('rf',RandomForestClassifier()),('gbm',lightgbm.LGBMClassifier())]\n",
    "sc=StackingClassifier(estimators=estimators, final_estimator=LogisticRegression())\n",
    "sc.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 149,
   "id": "2367217a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.5"
      ]
     },
     "execution_count": 149,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_sc=sc.predict(X_test)\n",
    "fpr_s1, tpr_s1, thresholds_s1 = metrics.roc_curve(y_test, pred_sc)\n",
    "sc_m=metrics.auc(fpr_s1, fpr_s1)\n",
    "sc_m"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 150,
   "id": "60a43f29",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.81      0.82      0.81     15056\n",
      "           1       0.82      0.82      0.82     15207\n",
      "\n",
      "    accuracy                           0.82     30263\n",
      "   macro avg       0.82      0.82      0.82     30263\n",
      "weighted avg       0.82      0.82      0.82     30263\n",
      "\n"
     ]
    }
   ],
   "source": [
    "r=classification_report(y_test, pred_sc)\n",
    "print(r)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 151,
   "id": "5ef51d62",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[(('xg',\n",
       "   XGBClassifier(base_score=None, booster=None, callbacks=None,\n",
       "                 colsample_bylevel=None, colsample_bynode=None,\n",
       "                 colsample_bytree=None, early_stopping_rounds=None,\n",
       "                 enable_categorical=False, eval_metric=None, feature_types=None,\n",
       "                 gamma=None, gpu_id=None, grow_policy=None, importance_type=None,\n",
       "                 interaction_constraints=None, learning_rate=None, max_bin=None,\n",
       "                 max_cat_threshold=None, max_cat_to_onehot=None,\n",
       "                 max_delta_step=None, max_depth=None, max_leaves=None,\n",
       "                 min_child_weight=None, missing=nan, monotone_constraints=None,\n",
       "                 n_estimators=100, n_jobs=None, num_parallel_tree=None,\n",
       "                 predictor=None, random_state=None, ...)),\n",
       "  ('dt', DecisionTreeClassifier())),\n",
       " (('xg',\n",
       "   XGBClassifier(base_score=None, booster=None, callbacks=None,\n",
       "                 colsample_bylevel=None, colsample_bynode=None,\n",
       "                 colsample_bytree=None, early_stopping_rounds=None,\n",
       "                 enable_categorical=False, eval_metric=None, feature_types=None,\n",
       "                 gamma=None, gpu_id=None, grow_policy=None, importance_type=None,\n",
       "                 interaction_constraints=None, learning_rate=None, max_bin=None,\n",
       "                 max_cat_threshold=None, max_cat_to_onehot=None,\n",
       "                 max_delta_step=None, max_depth=None, max_leaves=None,\n",
       "                 min_child_weight=None, missing=nan, monotone_constraints=None,\n",
       "                 n_estimators=100, n_jobs=None, num_parallel_tree=None,\n",
       "                 predictor=None, random_state=None, ...)),\n",
       "  ('rf', RandomForestClassifier())),\n",
       " (('xg',\n",
       "   XGBClassifier(base_score=None, booster=None, callbacks=None,\n",
       "                 colsample_bylevel=None, colsample_bynode=None,\n",
       "                 colsample_bytree=None, early_stopping_rounds=None,\n",
       "                 enable_categorical=False, eval_metric=None, feature_types=None,\n",
       "                 gamma=None, gpu_id=None, grow_policy=None, importance_type=None,\n",
       "                 interaction_constraints=None, learning_rate=None, max_bin=None,\n",
       "                 max_cat_threshold=None, max_cat_to_onehot=None,\n",
       "                 max_delta_step=None, max_depth=None, max_leaves=None,\n",
       "                 min_child_weight=None, missing=nan, monotone_constraints=None,\n",
       "                 n_estimators=100, n_jobs=None, num_parallel_tree=None,\n",
       "                 predictor=None, random_state=None, ...)),\n",
       "  ('gbm', LGBMClassifier())),\n",
       " (('dt', DecisionTreeClassifier()), ('rf', RandomForestClassifier())),\n",
       " (('dt', DecisionTreeClassifier()), ('gbm', LGBMClassifier())),\n",
       " (('rf', RandomForestClassifier()), ('gbm', LGBMClassifier()))]"
      ]
     },
     "execution_count": 151,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Implementing different combinations of stacking base models.\n",
    "from itertools import combinations\n",
    "r=2\n",
    "com=list(combinations(estimators, r))\n",
    "com"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 152,
   "id": "25878909",
   "metadata": {},
   "outputs": [],
   "source": [
    "m=[]\n",
    "for c in com:\n",
    "    sc=StackingClassifier(estimators=c, final_estimator=LogisticRegression())\n",
    "    sc.fit(X_train, y_train)\n",
    "    pred_sc=sc.predict(X_test)\n",
    "    fpr, tpr, thresholds = metrics.roc_curve(y_test, pred_sc)\n",
    "    sc_m=metrics.auc(fpr, tpr)\n",
    "    m.append(sc_m)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 153,
   "id": "bdd4e670",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('xg', 'dt'),\n",
       " ('xg', 'rf'),\n",
       " ('xg', 'gbm'),\n",
       " ('dt', 'rf'),\n",
       " ('dt', 'gbm'),\n",
       " ('rf', 'gbm')]"
      ]
     },
     "execution_count": 153,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "com2=('xg','dt','rf','gbm')\n",
    "com2=list(combinations(com2, r))\n",
    "com2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 154,
   "id": "6517ab7b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[0.8146111337995458,\n",
       " 0.8147678512789883,\n",
       " 0.8157937138582146,\n",
       " 0.7997989112276792,\n",
       " 0.8118118892161008,\n",
       " 0.812765067712049]"
      ]
     },
     "execution_count": 154,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "m"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 155,
   "id": "b93a8fce",
   "metadata": {},
   "outputs": [],
   "source": [
    "res = \"\\n\".join(\"{} {}\".format(x, y) for x, y in zip(com2, m))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 156,
   "id": "2e546c1d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "('xg', 'dt') 0.8146111337995458\n",
      "('xg', 'rf') 0.8147678512789883\n",
      "('xg', 'gbm') 0.8157937138582146\n",
      "('dt', 'rf') 0.7997989112276792\n",
      "('dt', 'gbm') 0.8118118892161008\n",
      "('rf', 'gbm') 0.812765067712049\n"
     ]
    }
   ],
   "source": [
    "print(res)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 157,
   "id": "a386f7b3",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[(('xg',\n",
       "   XGBClassifier(base_score=None, booster=None, callbacks=None,\n",
       "                 colsample_bylevel=None, colsample_bynode=None,\n",
       "                 colsample_bytree=None, early_stopping_rounds=None,\n",
       "                 enable_categorical=False, eval_metric=None, feature_types=None,\n",
       "                 gamma=None, gpu_id=None, grow_policy=None, importance_type=None,\n",
       "                 interaction_constraints=None, learning_rate=None, max_bin=None,\n",
       "                 max_cat_threshold=None, max_cat_to_onehot=None,\n",
       "                 max_delta_step=None, max_depth=None, max_leaves=None,\n",
       "                 min_child_weight=None, missing=nan, monotone_constraints=None,\n",
       "                 n_estimators=100, n_jobs=None, num_parallel_tree=None,\n",
       "                 predictor=None, random_state=None, ...)),\n",
       "  ('dt', DecisionTreeClassifier()),\n",
       "  ('rf', RandomForestClassifier())),\n",
       " (('xg',\n",
       "   XGBClassifier(base_score=None, booster=None, callbacks=None,\n",
       "                 colsample_bylevel=None, colsample_bynode=None,\n",
       "                 colsample_bytree=None, early_stopping_rounds=None,\n",
       "                 enable_categorical=False, eval_metric=None, feature_types=None,\n",
       "                 gamma=None, gpu_id=None, grow_policy=None, importance_type=None,\n",
       "                 interaction_constraints=None, learning_rate=None, max_bin=None,\n",
       "                 max_cat_threshold=None, max_cat_to_onehot=None,\n",
       "                 max_delta_step=None, max_depth=None, max_leaves=None,\n",
       "                 min_child_weight=None, missing=nan, monotone_constraints=None,\n",
       "                 n_estimators=100, n_jobs=None, num_parallel_tree=None,\n",
       "                 predictor=None, random_state=None, ...)),\n",
       "  ('dt', DecisionTreeClassifier()),\n",
       "  ('gbm', LGBMClassifier())),\n",
       " (('xg',\n",
       "   XGBClassifier(base_score=None, booster=None, callbacks=None,\n",
       "                 colsample_bylevel=None, colsample_bynode=None,\n",
       "                 colsample_bytree=None, early_stopping_rounds=None,\n",
       "                 enable_categorical=False, eval_metric=None, feature_types=None,\n",
       "                 gamma=None, gpu_id=None, grow_policy=None, importance_type=None,\n",
       "                 interaction_constraints=None, learning_rate=None, max_bin=None,\n",
       "                 max_cat_threshold=None, max_cat_to_onehot=None,\n",
       "                 max_delta_step=None, max_depth=None, max_leaves=None,\n",
       "                 min_child_weight=None, missing=nan, monotone_constraints=None,\n",
       "                 n_estimators=100, n_jobs=None, num_parallel_tree=None,\n",
       "                 predictor=None, random_state=None, ...)),\n",
       "  ('rf', RandomForestClassifier()),\n",
       "  ('gbm', LGBMClassifier())),\n",
       " (('dt', DecisionTreeClassifier()),\n",
       "  ('rf', RandomForestClassifier()),\n",
       "  ('gbm', LGBMClassifier()))]"
      ]
     },
     "execution_count": 157,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "r=3\n",
    "com=list(combinations(estimators, r))\n",
    "com"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 158,
   "id": "bf5a90bf",
   "metadata": {},
   "outputs": [],
   "source": [
    "q=[]\n",
    "for c in com:\n",
    "    sc=StackingClassifier(estimators=c, final_estimator=LogisticRegression())\n",
    "    sc.fit(X_train, y_train)\n",
    "    pred_sc=sc.predict(X_test)\n",
    "    fpr, tpr, thresholds = metrics.roc_curve(y_test, pred_sc)\n",
    "    sc_m=metrics.auc(fpr, tpr)\n",
    "    q.append(sc_m)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 159,
   "id": "f857b325",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[0.8149000291723419, 0.8156930965324641, 0.8155516854478687, 0.813685696369904]"
      ]
     },
     "execution_count": 159,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "q"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 160,
   "id": "f1530f97",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "('xg', 'dt', 'rf') 0.8149000291723419\n",
      "('xg', 'dt', 'gbm') 0.8156930965324641\n",
      "('xg', 'rf', 'gbm') 0.8155516854478687\n",
      "('dt', 'rf', 'gbm') 0.813685696369904\n"
     ]
    }
   ],
   "source": [
    "com2=('xg','dt','rf','gbm')\n",
    "com2=list(combinations(com2, r))\n",
    "res = \"\\n\".join(\"{} {}\".format(x, y) for x, y in zip(com2, q))\n",
    "print(res)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 161,
   "id": "85ef955f",
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "from sklearn.metrics import roc_curve"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 162,
   "id": "8cb587c9",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Plotting the Roc Curve.\n",
    "roc_auc_d1 = metrics.auc(fpr_d1, tpr_d1)\n",
    "roc_auc_r1 = metrics.auc(fpr_r1, tpr_r1)\n",
    "roc_auc_x1 = metrics.auc(fpr_x1, tpr_x1)\n",
    "roc_auc_l1 = metrics.auc(fpr_l1, tpr_l1)\n",
    "roc_auc_s1 = metrics.auc(fpr_s1, tpr_s1)\n",
    "\n",
    "plt.plot(fpr_d1, tpr_d1, \"r:\", linewidth=2, label=\"Decision Tree (AUC = %.2f)\" % roc_auc_d1)\n",
    "plt.plot(fpr_r1, tpr_r1, \"g--\", linewidth=2, label=\"Random Forest (AUC = %.2f)\" % roc_auc_r1)\n",
    "plt.plot(fpr_x1, tpr_x1, \"b-\", linewidth=2, label=\"XGBoost (AUC = %.2f)\" % roc_auc_x1)\n",
    "plt.plot(fpr_l1, tpr_l1, \"m-.\", linewidth=2, label=\"LightBGM (AUC = %.2f)\" % roc_auc_l1)\n",
    "plt.plot(fpr_s1, tpr_s1, \"k-\", linewidth=2, label=\"Stack Ensemble (AUC = %.2f)\" % roc_auc_s1)\n",
    "\n",
    "\n",
    "plt.legend(loc=\"lower right\")\n",
    "plt.xlabel(\"False Positive Rate\")\n",
    "plt.ylabel(\"True Positive Rate \")\n",
    "plt.title(\"ROC Curve Comparison of Multiple Models\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 203,
   "id": "93c5bcfb",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "StackingClassifier(estimators=[('xg',\n",
       "                                XGBClassifier(base_score=None, booster=None,\n",
       "                                              callbacks=None,\n",
       "                                              colsample_bylevel=None,\n",
       "                                              colsample_bynode=None,\n",
       "                                              colsample_bytree=None,\n",
       "                                              early_stopping_rounds=None,\n",
       "                                              enable_categorical=False,\n",
       "                                              eval_metric=None,\n",
       "                                              feature_types=None, gamma=None,\n",
       "                                              gpu_id=None, grow_policy=None,\n",
       "                                              importance_type=None,\n",
       "                                              interaction_constraints=None,\n",
       "                                              learning_r...n=None,\n",
       "                                              max_cat_threshold=None,\n",
       "                                              max_cat_to_onehot=None,\n",
       "                                              max_delta_step=None,\n",
       "                                              max_depth=None, max_leaves=None,\n",
       "                                              min_child_weight=None,\n",
       "                                              missing=nan,\n",
       "                                              monotone_constraints=None,\n",
       "                                              n_estimators=100, n_jobs=None,\n",
       "                                              num_parallel_tree=None,\n",
       "                                              predictor=None, random_state=None, ...)),\n",
       "                               ('dt', DecisionTreeClassifier()),\n",
       "                               ('gbm', LGBMClassifier())],\n",
       "                   final_estimator=LogisticRegression())"
      ]
     },
     "execution_count": 203,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Stacking for the best combinations.\n",
    "estimators = [('xg',xgboost.XGBClassifier()),('dt',DecisionTreeClassifier()),('gbm',lightgbm.LGBMClassifier())]\n",
    "sc_c1=StackingClassifier(estimators=estimators, final_estimator=LogisticRegression())\n",
    "sc_c1.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 204,
   "id": "72afe6b7",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.5"
      ]
     },
     "execution_count": 204,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_sc_c1=sc_c1.predict(X_test)\n",
    "fpr_sc1, tpr_sc1, thresholds_sc1 = metrics.roc_curve(y_test, pred_sc_c1)\n",
    "sc_mc1=metrics.auc(fpr_sc1, fpr_sc1)\n",
    "sc_mc1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 212,
   "id": "4c86e9e8",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.81      0.82      0.82     15056\n",
      "           1       0.82      0.81      0.82     15207\n",
      "\n",
      "    accuracy                           0.82     30263\n",
      "   macro avg       0.82      0.82      0.82     30263\n",
      "weighted avg       0.82      0.82      0.82     30263\n",
      "\n"
     ]
    }
   ],
   "source": [
    "r_c1=classification_report(y_test, pred_sc_c1)\n",
    "print(r_c1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 207,
   "id": "73e914cf",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Plotting the Roc Curve.\n",
    "roc_auc_c1 = metrics.auc(fpr_sc1, tpr_sc1)\n",
    "plt.plot(fpr_sc1, tpr_sc1, \"k-\", linewidth=2, label=\"Stack Ensemble Best Combination 1 (AUC = %.2f)\" % roc_auc_c1)\n",
    "\n",
    "plt.legend(loc=\"lower right\")\n",
    "plt.xlabel(\"False Positive Rate\")\n",
    "plt.ylabel(\"True Positive Rate \")\n",
    "plt.title(\"ROC Curve Comparison of Multiple Models\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "9bc273b7",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "bb0be07b",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "3514f248",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e930fbde",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b0eb6e9e",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 163,
   "id": "5f1cf3ea",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Second datset for validating the approach of WOE and stacking.\n",
    "customer2= pd.read_csv('second.csv') "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 164,
   "id": "a2a062e9",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/var/folders/j5/9fvnh0857k3_lqjpjrk029hw0000gn/T/ipykernel_749/1598678812.py:1: FutureWarning: Dropping of nuisance columns in DataFrame reductions (with 'numeric_only=None') is deprecated; in a future version this will raise TypeError.  Select only valid columns before calling the reduction.\n",
      "  customer2.fillna(customer2.mean(), inplace=True)\n"
     ]
    }
   ],
   "source": [
    "customer2.fillna(customer2.mean(), inplace=True)\n",
    "for column in customer2.columns:\n",
    "    if customer2[column].dtype == 'object':\n",
    "        customer2[column].fillna(customer2[column].mode()[0], inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 165,
   "id": "01ad8b62",
   "metadata": {},
   "outputs": [],
   "source": [
    "del customer2['ServiceArea']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 166,
   "id": "f7625bf5",
   "metadata": {},
   "outputs": [],
   "source": [
    "train_label2=customer2['Churn']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 167,
   "id": "c79fc5a5",
   "metadata": {},
   "outputs": [],
   "source": [
    "del customer2['Churn']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 168,
   "id": "54dea984",
   "metadata": {},
   "outputs": [],
   "source": [
    "train_label2=train_label2.replace('No',0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 169,
   "id": "6a123137",
   "metadata": {},
   "outputs": [],
   "source": [
    "train_label2=train_label2.replace('Yes',1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 170,
   "id": "d0999942",
   "metadata": {},
   "outputs": [],
   "source": [
    "cat_customer2= customer2.select_dtypes(include=['object']).columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 171,
   "id": "b8b918ae",
   "metadata": {},
   "outputs": [],
   "source": [
    "customer_cat2=customer2[cat_customer2]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 172,
   "id": "afd1d46f",
   "metadata": {},
   "outputs": [],
   "source": [
    "num_customer2=customer2.select_dtypes(include=['float64', 'int64']).columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 173,
   "id": "fb1d130f",
   "metadata": {},
   "outputs": [],
   "source": [
    "customer_num2=customer2[num_customer2]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 174,
   "id": "1b43a19c",
   "metadata": {},
   "outputs": [],
   "source": [
    "z2=customer_cat2\n",
    "#label encoder\n",
    "from sklearn.preprocessing import LabelEncoder\n",
    "label=LabelEncoder()\n",
    "z2=z2.astype(str)\n",
    "z2[cat_customer2] = z2[cat_customer2].apply(lambda col: label.fit_transform(col))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 175,
   "id": "c79b830e",
   "metadata": {},
   "outputs": [],
   "source": [
    "frames2 = [customer_num2, z2]\n",
    "customer_updated2=pd.concat(frames2,axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 176,
   "id": "b3cbabe9",
   "metadata": {},
   "outputs": [],
   "source": [
    "#normalizing the values\n",
    "x_norm = (customer_updated - customer_updated.min(axis=0)) / (customer_updated.max(axis=0) - customer_updated.min(axis=0))\n",
    "customer_updated[:] = x_norm"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 177,
   "id": "1d3d5edc",
   "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>CustomerID</th>\n",
       "      <th>MonthlyRevenue</th>\n",
       "      <th>MonthlyMinutes</th>\n",
       "      <th>TotalRecurringCharge</th>\n",
       "      <th>DirectorAssistedCalls</th>\n",
       "      <th>OverageMinutes</th>\n",
       "      <th>RoamingCalls</th>\n",
       "      <th>PercChangeMinutes</th>\n",
       "      <th>PercChangeRevenues</th>\n",
       "      <th>DroppedCalls</th>\n",
       "      <th>...</th>\n",
       "      <th>HasCreditCard</th>\n",
       "      <th>NewCellphoneUser</th>\n",
       "      <th>NotNewCellphoneUser</th>\n",
       "      <th>OwnsMotorcycle</th>\n",
       "      <th>HandsetPrice</th>\n",
       "      <th>MadeCallToRetentionTeam</th>\n",
       "      <th>CreditRating</th>\n",
       "      <th>PrizmCode</th>\n",
       "      <th>Occupation</th>\n",
       "      <th>MaritalStatus</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.211066</td>\n",
       "      <td>0.316427</td>\n",
       "      <td>0.136343</td>\n",
       "      <td>0.005571</td>\n",
       "      <td>-0.021213</td>\n",
       "      <td>0.0061</td>\n",
       "      <td>0.344598</td>\n",
       "      <td>0.434210</td>\n",
       "      <td>-0.073113</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.013214</td>\n",
       "      <td>0.008279</td>\n",
       "      <td>-0.005946</td>\n",
       "      <td>-0.001871</td>\n",
       "      <td>-0.089412</td>\n",
       "      <td>0.705279</td>\n",
       "      <td>0.096502</td>\n",
       "      <td>-0.033245</td>\n",
       "      <td>-0.031427</td>\n",
       "      <td>-0.085188</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.051298</td>\n",
       "      <td>0.658080</td>\n",
       "      <td>-0.032963</td>\n",
       "      <td>0.045701</td>\n",
       "      <td>-0.021213</td>\n",
       "      <td>0.0061</td>\n",
       "      <td>-0.031125</td>\n",
       "      <td>0.090737</td>\n",
       "      <td>0.007682</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.013214</td>\n",
       "      <td>-0.035063</td>\n",
       "      <td>-0.005946</td>\n",
       "      <td>-0.001871</td>\n",
       "      <td>-0.089412</td>\n",
       "      <td>-0.028163</td>\n",
       "      <td>-0.135768</td>\n",
       "      <td>-0.033245</td>\n",
       "      <td>-0.031427</td>\n",
       "      <td>-0.013617</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>-0.102025</td>\n",
       "      <td>-0.588099</td>\n",
       "      <td>0.045701</td>\n",
       "      <td>-0.021213</td>\n",
       "      <td>0.0061</td>\n",
       "      <td>-0.115330</td>\n",
       "      <td>0.090737</td>\n",
       "      <td>0.188690</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.013214</td>\n",
       "      <td>-0.035063</td>\n",
       "      <td>-0.005946</td>\n",
       "      <td>-0.001871</td>\n",
       "      <td>0.065183</td>\n",
       "      <td>-0.028163</td>\n",
       "      <td>0.104594</td>\n",
       "      <td>0.056476</td>\n",
       "      <td>-0.038029</td>\n",
       "      <td>-0.013617</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.211066</td>\n",
       "      <td>0.211066</td>\n",
       "      <td>-0.099873</td>\n",
       "      <td>-0.126072</td>\n",
       "      <td>-0.021213</td>\n",
       "      <td>0.0061</td>\n",
       "      <td>-0.782185</td>\n",
       "      <td>0.035176</td>\n",
       "      <td>0.498748</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.013214</td>\n",
       "      <td>-0.035063</td>\n",
       "      <td>-0.005946</td>\n",
       "      <td>-0.001871</td>\n",
       "      <td>-0.118379</td>\n",
       "      <td>-0.028163</td>\n",
       "      <td>-0.135768</td>\n",
       "      <td>-0.009552</td>\n",
       "      <td>0.011057</td>\n",
       "      <td>-0.085188</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.194399</td>\n",
       "      <td>1.357169</td>\n",
       "      <td>-0.032963</td>\n",
       "      <td>0.045701</td>\n",
       "      <td>-0.021213</td>\n",
       "      <td>0.0061</td>\n",
       "      <td>0.775924</td>\n",
       "      <td>-0.094446</td>\n",
       "      <td>0.188690</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.013214</td>\n",
       "      <td>0.008279</td>\n",
       "      <td>0.036867</td>\n",
       "      <td>-0.001871</td>\n",
       "      <td>-0.118379</td>\n",
       "      <td>-0.028163</td>\n",
       "      <td>0.096502</td>\n",
       "      <td>-0.009552</td>\n",
       "      <td>-0.031427</td>\n",
       "      <td>-0.013617</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 56 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   CustomerID  MonthlyRevenue  MonthlyMinutes  TotalRecurringCharge  \\\n",
       "0         0.0        0.211066        0.316427              0.136343   \n",
       "1         0.0       -0.051298        0.658080             -0.032963   \n",
       "2         0.0        0.000000       -0.102025             -0.588099   \n",
       "3         0.0        0.211066        0.211066             -0.099873   \n",
       "4         0.0       -0.194399        1.357169             -0.032963   \n",
       "\n",
       "   DirectorAssistedCalls  OverageMinutes  RoamingCalls  PercChangeMinutes  \\\n",
       "0               0.005571       -0.021213        0.0061           0.344598   \n",
       "1               0.045701       -0.021213        0.0061          -0.031125   \n",
       "2               0.045701       -0.021213        0.0061          -0.115330   \n",
       "3              -0.126072       -0.021213        0.0061          -0.782185   \n",
       "4               0.045701       -0.021213        0.0061           0.775924   \n",
       "\n",
       "   PercChangeRevenues  DroppedCalls  ...  HasCreditCard  NewCellphoneUser  \\\n",
       "0            0.434210     -0.073113  ...      -0.013214          0.008279   \n",
       "1            0.090737      0.007682  ...      -0.013214         -0.035063   \n",
       "2            0.090737      0.188690  ...      -0.013214         -0.035063   \n",
       "3            0.035176      0.498748  ...      -0.013214         -0.035063   \n",
       "4           -0.094446      0.188690  ...      -0.013214          0.008279   \n",
       "\n",
       "   NotNewCellphoneUser  OwnsMotorcycle  HandsetPrice  MadeCallToRetentionTeam  \\\n",
       "0            -0.005946       -0.001871     -0.089412                 0.705279   \n",
       "1            -0.005946       -0.001871     -0.089412                -0.028163   \n",
       "2            -0.005946       -0.001871      0.065183                -0.028163   \n",
       "3            -0.005946       -0.001871     -0.118379                -0.028163   \n",
       "4             0.036867       -0.001871     -0.118379                -0.028163   \n",
       "\n",
       "   CreditRating  PrizmCode  Occupation  MaritalStatus  \n",
       "0      0.096502  -0.033245   -0.031427      -0.085188  \n",
       "1     -0.135768  -0.033245   -0.031427      -0.013617  \n",
       "2      0.104594   0.056476   -0.038029      -0.013617  \n",
       "3     -0.135768  -0.009552    0.011057      -0.085188  \n",
       "4      0.096502  -0.009552   -0.031427      -0.013617  \n",
       "\n",
       "[5 rows x 56 columns]"
      ]
     },
     "execution_count": 177,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "enc = WeightOfEvidenceEncoder(cols=customer_updated2.columns)\n",
    "customer_updated2 = enc.fit_transform(customer_updated2, train_label2)\n",
    "customer_updated2.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 178,
   "id": "c32180ef",
   "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>CustomerID</th>\n",
       "      <th>MonthlyRevenue</th>\n",
       "      <th>MonthlyMinutes</th>\n",
       "      <th>TotalRecurringCharge</th>\n",
       "      <th>DirectorAssistedCalls</th>\n",
       "      <th>OverageMinutes</th>\n",
       "      <th>RoamingCalls</th>\n",
       "      <th>PercChangeMinutes</th>\n",
       "      <th>PercChangeRevenues</th>\n",
       "      <th>DroppedCalls</th>\n",
       "      <th>...</th>\n",
       "      <th>HasCreditCard</th>\n",
       "      <th>NewCellphoneUser</th>\n",
       "      <th>NotNewCellphoneUser</th>\n",
       "      <th>OwnsMotorcycle</th>\n",
       "      <th>HandsetPrice</th>\n",
       "      <th>MadeCallToRetentionTeam</th>\n",
       "      <th>CreditRating</th>\n",
       "      <th>PrizmCode</th>\n",
       "      <th>Occupation</th>\n",
       "      <th>MaritalStatus</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.211066</td>\n",
       "      <td>0.316427</td>\n",
       "      <td>0.136343</td>\n",
       "      <td>0.005571</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0061</td>\n",
       "      <td>0.344598</td>\n",
       "      <td>0.434210</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.008279</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.705279</td>\n",
       "      <td>0.096502</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.658080</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.045701</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0061</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.090737</td>\n",
       "      <td>0.007682</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</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>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.045701</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0061</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.090737</td>\n",
       "      <td>0.188690</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.065183</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.104594</td>\n",
       "      <td>0.056476</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.211066</td>\n",
       "      <td>0.211066</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0061</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.035176</td>\n",
       "      <td>0.498748</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.011057</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.357169</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.045701</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0061</td>\n",
       "      <td>0.775924</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.188690</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.008279</td>\n",
       "      <td>0.036867</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.096502</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 56 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   CustomerID  MonthlyRevenue  MonthlyMinutes  TotalRecurringCharge  \\\n",
       "0         0.0        0.211066        0.316427              0.136343   \n",
       "1         0.0        0.000000        0.658080              0.000000   \n",
       "2         0.0        0.000000        0.000000              0.000000   \n",
       "3         0.0        0.211066        0.211066              0.000000   \n",
       "4         0.0        0.000000        1.357169              0.000000   \n",
       "\n",
       "   DirectorAssistedCalls  OverageMinutes  RoamingCalls  PercChangeMinutes  \\\n",
       "0               0.005571             0.0        0.0061           0.344598   \n",
       "1               0.045701             0.0        0.0061           0.000000   \n",
       "2               0.045701             0.0        0.0061           0.000000   \n",
       "3               0.000000             0.0        0.0061           0.000000   \n",
       "4               0.045701             0.0        0.0061           0.775924   \n",
       "\n",
       "   PercChangeRevenues  DroppedCalls  ...  HasCreditCard  NewCellphoneUser  \\\n",
       "0            0.434210      0.000000  ...            0.0          0.008279   \n",
       "1            0.090737      0.007682  ...            0.0          0.000000   \n",
       "2            0.090737      0.188690  ...            0.0          0.000000   \n",
       "3            0.035176      0.498748  ...            0.0          0.000000   \n",
       "4            0.000000      0.188690  ...            0.0          0.008279   \n",
       "\n",
       "   NotNewCellphoneUser  OwnsMotorcycle  HandsetPrice  MadeCallToRetentionTeam  \\\n",
       "0             0.000000             0.0      0.000000                 0.705279   \n",
       "1             0.000000             0.0      0.000000                 0.000000   \n",
       "2             0.000000             0.0      0.065183                 0.000000   \n",
       "3             0.000000             0.0      0.000000                 0.000000   \n",
       "4             0.036867             0.0      0.000000                 0.000000   \n",
       "\n",
       "   CreditRating  PrizmCode  Occupation  MaritalStatus  \n",
       "0      0.096502   0.000000    0.000000            0.0  \n",
       "1      0.000000   0.000000    0.000000            0.0  \n",
       "2      0.104594   0.056476    0.000000            0.0  \n",
       "3      0.000000   0.000000    0.011057            0.0  \n",
       "4      0.096502   0.000000    0.000000            0.0  \n",
       "\n",
       "[5 rows x 56 columns]"
      ]
     },
     "execution_count": 178,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "customer_updated2[customer_updated2 < 0]=0\n",
    "customer_updated2.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 179,
   "id": "bb0ed290",
   "metadata": {},
   "outputs": [],
   "source": [
    "del customer_updated2['CustomerID']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 180,
   "id": "d1aa8411",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    36336\n",
       "1    14711\n",
       "Name: Churn, dtype: int64"
      ]
     },
     "execution_count": 180,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_label2.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 181,
   "id": "8309dcb7",
   "metadata": {},
   "outputs": [],
   "source": [
    "#applying SVM smote \n",
    "oversampler=SMOTE()\n",
    "customer_updated2,train_label2=oversampler.fit_resample(customer_updated2,train_label2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 182,
   "id": "ae63556b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1    36336\n",
       "0    36336\n",
       "Name: Churn, dtype: int64"
      ]
     },
     "execution_count": 182,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_label2.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 183,
   "id": "a2660e8b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(72672, 33)"
      ]
     },
     "execution_count": 183,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "customer_updated2 = SelectPercentile(chi2, percentile=60).fit_transform(customer_updated2, train_label2)\n",
    "customer_updated2.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 184,
   "id": "d2ee63df",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "X_train2, X_test2, y_train2, y_test2 = train_test_split(customer_updated2, train_label2, test_size=0.3, random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 185,
   "id": "9a0daa1b",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.8420680165126793\n"
     ]
    }
   ],
   "source": [
    "xg=xgboost.XGBClassifier()\n",
    "xg.fit(X_train2,y_train2)\n",
    "k_folds = KFold(n_splits = 10)\n",
    "print(cross_val_score(xg, X_train2, y_train2, cv=k_folds).mean())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 186,
   "id": "15af30ab",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8426384078452165"
      ]
     },
     "execution_count": 186,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics import classification_report\n",
    "pred_xg=xg.predict(X_test2)\n",
    "fpr_x2, tpr_x2, thresholds_x2 = metrics.roc_curve(y_test2, pred_xg)\n",
    "#r=classification_report(y_test2, pred_xg)\n",
    "xg_m=metrics.auc(fpr_x2, tpr_x2)\n",
    "xg_m"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 187,
   "id": "dee16873",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.7467859249066248\n"
     ]
    }
   ],
   "source": [
    "dt=DecisionTreeClassifier()\n",
    "dt.fit(X_train2,y_train2)\n",
    "k_folds = KFold(n_splits = 10)\n",
    "print(cross_val_score(dt, X_train2, y_train2, cv=k_folds).mean())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 188,
   "id": "8fd23855",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7513162936134751"
      ]
     },
     "execution_count": 188,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred=dt.predict(X_test2)\n",
    "fpr_d2, tpr_d2, thresholds_d2 = metrics.roc_curve(y_test2, pred)\n",
    "dt_m=metrics.auc(fpr_d2, tpr_d2)\n",
    "dt_m"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 189,
   "id": "c1e30faa",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.8418910949479065\n"
     ]
    }
   ],
   "source": [
    "#random Forest\n",
    "rf=RandomForestClassifier()\n",
    "rf.fit(X_train2,y_train2)\n",
    "k_folds = KFold(n_splits = 10)\n",
    "print(cross_val_score(rf, X_train2, y_train2, cv=k_folds).mean())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 190,
   "id": "e7d93f1a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8442086644942899"
      ]
     },
     "execution_count": 190,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_rf=rf.predict(X_test2)\n",
    "fpr_r2, tpr_r2, thresholds_r2 = metrics.roc_curve(y_test2, pred_rf)\n",
    "rf_m=metrics.auc(fpr_r2, tpr_r2)\n",
    "rf_m"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 191,
   "id": "7932f512",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.8475132691173581\n"
     ]
    }
   ],
   "source": [
    "gbm=lightgbm.LGBMClassifier()\n",
    "gbm.fit(X_train2,y_train2)\n",
    "k_folds = KFold(n_splits = 10)\n",
    "print(cross_val_score(gbm, X_train2, y_train2, cv=k_folds).mean())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 192,
   "id": "cd21ea5b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8466485172710054"
      ]
     },
     "execution_count": 192,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_gbm=gbm.predict(X_test2)\n",
    "fpr_l2, tpr_l2, thresholds_l2 = metrics.roc_curve(y_test2, pred_gbm)\n",
    "gbm_m=metrics.auc(fpr_l2, tpr_l2)\n",
    "gbm_m"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 193,
   "id": "c5ddf01d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "StackingClassifier(estimators=[('xg',\n",
       "                                XGBClassifier(base_score=None, booster=None,\n",
       "                                              callbacks=None,\n",
       "                                              colsample_bylevel=None,\n",
       "                                              colsample_bynode=None,\n",
       "                                              colsample_bytree=None,\n",
       "                                              early_stopping_rounds=None,\n",
       "                                              enable_categorical=False,\n",
       "                                              eval_metric=None,\n",
       "                                              feature_types=None, gamma=None,\n",
       "                                              gpu_id=None, grow_policy=None,\n",
       "                                              importance_type=None,\n",
       "                                              interaction_constraints=None,\n",
       "                                              learning_r...\n",
       "                                              max_cat_to_onehot=None,\n",
       "                                              max_delta_step=None,\n",
       "                                              max_depth=None, max_leaves=None,\n",
       "                                              min_child_weight=None,\n",
       "                                              missing=nan,\n",
       "                                              monotone_constraints=None,\n",
       "                                              n_estimators=100, n_jobs=None,\n",
       "                                              num_parallel_tree=None,\n",
       "                                              predictor=None, random_state=None, ...)),\n",
       "                               ('dt', DecisionTreeClassifier()),\n",
       "                               ('rf', RandomForestClassifier()),\n",
       "                               ('gbm', LGBMClassifier())],\n",
       "                   final_estimator=LogisticRegression())"
      ]
     },
     "execution_count": 193,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.ensemble import StackingClassifier\n",
    "from sklearn.ensemble import StackingClassifier\n",
    "estimators = [('xg',xgboost.XGBClassifier()),('dt',DecisionTreeClassifier()),('rf',RandomForestClassifier()),('gbm',lightgbm.LGBMClassifier())]\n",
    "sc=StackingClassifier(estimators=estimators, final_estimator=LogisticRegression())\n",
    "sc.fit(X_train2, y_train2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 194,
   "id": "ab6989fc",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.8480525607063868\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.83      0.88      0.85     10851\n",
      "           1       0.87      0.82      0.84     10951\n",
      "\n",
      "    accuracy                           0.85     21802\n",
      "   macro avg       0.85      0.85      0.85     21802\n",
      "weighted avg       0.85      0.85      0.85     21802\n",
      "\n"
     ]
    }
   ],
   "source": [
    "pred_sc=sc.predict(X_test2)\n",
    "fpr_s2, tpr_s2, thresholds_s2 = metrics.roc_curve(y_test2, pred_sc)\n",
    "sc_m=metrics.auc(fpr_s2, tpr_s2)\n",
    "r=classification_report(y_test2, pred_sc)\n",
    "print(sc_m)\n",
    "print(r)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 195,
   "id": "6af87026",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[(('xg',\n",
       "   XGBClassifier(base_score=None, booster=None, callbacks=None,\n",
       "                 colsample_bylevel=None, colsample_bynode=None,\n",
       "                 colsample_bytree=None, early_stopping_rounds=None,\n",
       "                 enable_categorical=False, eval_metric=None, feature_types=None,\n",
       "                 gamma=None, gpu_id=None, grow_policy=None, importance_type=None,\n",
       "                 interaction_constraints=None, learning_rate=None, max_bin=None,\n",
       "                 max_cat_threshold=None, max_cat_to_onehot=None,\n",
       "                 max_delta_step=None, max_depth=None, max_leaves=None,\n",
       "                 min_child_weight=None, missing=nan, monotone_constraints=None,\n",
       "                 n_estimators=100, n_jobs=None, num_parallel_tree=None,\n",
       "                 predictor=None, random_state=None, ...)),\n",
       "  ('dt', DecisionTreeClassifier())),\n",
       " (('xg',\n",
       "   XGBClassifier(base_score=None, booster=None, callbacks=None,\n",
       "                 colsample_bylevel=None, colsample_bynode=None,\n",
       "                 colsample_bytree=None, early_stopping_rounds=None,\n",
       "                 enable_categorical=False, eval_metric=None, feature_types=None,\n",
       "                 gamma=None, gpu_id=None, grow_policy=None, importance_type=None,\n",
       "                 interaction_constraints=None, learning_rate=None, max_bin=None,\n",
       "                 max_cat_threshold=None, max_cat_to_onehot=None,\n",
       "                 max_delta_step=None, max_depth=None, max_leaves=None,\n",
       "                 min_child_weight=None, missing=nan, monotone_constraints=None,\n",
       "                 n_estimators=100, n_jobs=None, num_parallel_tree=None,\n",
       "                 predictor=None, random_state=None, ...)),\n",
       "  ('rf', RandomForestClassifier())),\n",
       " (('xg',\n",
       "   XGBClassifier(base_score=None, booster=None, callbacks=None,\n",
       "                 colsample_bylevel=None, colsample_bynode=None,\n",
       "                 colsample_bytree=None, early_stopping_rounds=None,\n",
       "                 enable_categorical=False, eval_metric=None, feature_types=None,\n",
       "                 gamma=None, gpu_id=None, grow_policy=None, importance_type=None,\n",
       "                 interaction_constraints=None, learning_rate=None, max_bin=None,\n",
       "                 max_cat_threshold=None, max_cat_to_onehot=None,\n",
       "                 max_delta_step=None, max_depth=None, max_leaves=None,\n",
       "                 min_child_weight=None, missing=nan, monotone_constraints=None,\n",
       "                 n_estimators=100, n_jobs=None, num_parallel_tree=None,\n",
       "                 predictor=None, random_state=None, ...)),\n",
       "  ('gbm', LGBMClassifier())),\n",
       " (('dt', DecisionTreeClassifier()), ('rf', RandomForestClassifier())),\n",
       " (('dt', DecisionTreeClassifier()), ('gbm', LGBMClassifier())),\n",
       " (('rf', RandomForestClassifier()), ('gbm', LGBMClassifier()))]"
      ]
     },
     "execution_count": 195,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "r=2\n",
    "com=list(combinations(estimators, r))\n",
    "com"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 196,
   "id": "979fddea",
   "metadata": {},
   "outputs": [],
   "source": [
    "m=[]\n",
    "for c in com:\n",
    "    sc=StackingClassifier(estimators=c, final_estimator=LogisticRegression())\n",
    "    sc.fit(X_train2, y_train2)\n",
    "    pred_sc=sc.predict(X_test2)\n",
    "    fpr, tpr, thresholds = metrics.roc_curve(y_test2, pred_sc)\n",
    "    sc_m=metrics.auc(fpr, tpr)\n",
    "    m.append(sc_m)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 197,
   "id": "f9474c74",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('xg', 'dt'),\n",
       " ('xg', 'rf'),\n",
       " ('xg', 'gbm'),\n",
       " ('dt', 'rf'),\n",
       " ('dt', 'gbm'),\n",
       " ('rf', 'gbm')]"
      ]
     },
     "execution_count": 197,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "r=2\n",
    "com2=('xg','dt','rf','gbm')\n",
    "com2=list(combinations(com2, r))\n",
    "com2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 198,
   "id": "8ee523a6",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "('xg', 'dt') 0.8420240307565219\n",
      "('xg', 'rf') 0.8472046301105481\n",
      "('xg', 'gbm') 0.8458334152786104\n",
      "('dt', 'rf') 0.8442911525668236\n",
      "('dt', 'gbm') 0.8460152054584584\n",
      "('rf', 'gbm') 0.847632382353238\n"
     ]
    }
   ],
   "source": [
    "res = \"\\n\".join(\"{} {}\".format(x, y) for x, y in zip(com2, m))\n",
    "print(res)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 199,
   "id": "edf00a98",
   "metadata": {},
   "outputs": [],
   "source": [
    "r=3\n",
    "com=list(combinations(estimators, r))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 200,
   "id": "604e6421",
   "metadata": {},
   "outputs": [],
   "source": [
    "q=[]\n",
    "for c in com:\n",
    "    sc=StackingClassifier(estimators=c, final_estimator=LogisticRegression())\n",
    "    sc.fit(X_train, y_train)\n",
    "    pred_sc=sc.predict(X_test)\n",
    "    fpr, tpr, thresholds = metrics.roc_curve(y_test, pred_sc)\n",
    "    sc_m=metrics.auc(fpr, tpr)\n",
    "    q.append(sc_m)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 201,
   "id": "3e9d71f1",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "('xg', 'dt', 'rf') 0.8143367892198535\n",
      "('xg', 'dt', 'gbm') 0.8156602169375408\n",
      "('xg', 'rf', 'gbm') 0.8156868311526928\n",
      "('dt', 'rf', 'gbm') 0.8133236910689168\n"
     ]
    }
   ],
   "source": [
    "r=3\n",
    "com2=('xg','dt','rf','gbm')\n",
    "com2=list(combinations(com2, r))\n",
    "res = \"\\n\".join(\"{} {}\".format(x, y) for x, y in zip(com2, q))\n",
    "print(res)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 202,
   "id": "a82dbacd",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Plotting the Roc Curve.\n",
    "roc_auc_d2 = metrics.auc(fpr_d2, tpr_d2)\n",
    "roc_auc_r2 = metrics.auc(fpr_r2, tpr_r2)\n",
    "roc_auc_x2 = metrics.auc(fpr_x2, tpr_x2)\n",
    "roc_auc_l2 = metrics.auc(fpr_l2, tpr_l2)\n",
    "roc_auc_s2 = metrics.auc(fpr_s2, tpr_s2)\n",
    "\n",
    "plt.plot(fpr_d2, tpr_d2, \"r:\", linewidth=2, label=\"Decision Tree (AUC = %.2f)\" % roc_auc_d2)\n",
    "plt.plot(fpr_r2, tpr_r2, \"g--\", linewidth=2, label=\"Random Forest (AUC = %.2f)\" % roc_auc_r2)\n",
    "plt.plot(fpr_x2, tpr_x2, \"b-\", linewidth=2, label=\"XGBoost (AUC = %.2f)\" % roc_auc_x2)\n",
    "plt.plot(fpr_l2, tpr_l2, \"m-.\", linewidth=2, label=\"LightBGM (AUC = %.2f)\" % roc_auc_l2)\n",
    "plt.plot(fpr_s2, tpr_s2, \"k-\", linewidth=2, label=\"Stack Ensemble (AUC = %.2f)\" % roc_auc_s2)\n",
    "\n",
    "\n",
    "plt.legend(loc=\"lower right\")\n",
    "plt.xlabel(\"False Positive Rate\")\n",
    "plt.ylabel(\"True Positive Rate \")\n",
    "plt.title(\"ROC Curve Comparison of Multiple Models\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 208,
   "id": "ace5733f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "StackingClassifier(estimators=[('xg',\n",
       "                                XGBClassifier(base_score=None, booster=None,\n",
       "                                              callbacks=None,\n",
       "                                              colsample_bylevel=None,\n",
       "                                              colsample_bynode=None,\n",
       "                                              colsample_bytree=None,\n",
       "                                              early_stopping_rounds=None,\n",
       "                                              enable_categorical=False,\n",
       "                                              eval_metric=None,\n",
       "                                              feature_types=None, gamma=None,\n",
       "                                              gpu_id=None, grow_policy=None,\n",
       "                                              importance_type=None,\n",
       "                                              interaction_constraints=None,\n",
       "                                              learning_rate=None, max_bin=None,\n",
       "                                              max_cat_threshold=None,\n",
       "                                              max_cat_to_onehot=None,\n",
       "                                              max_delta_step=None,\n",
       "                                              max_depth=None, max_leaves=None,\n",
       "                                              min_child_weight=None,\n",
       "                                              missing=nan,\n",
       "                                              monotone_constraints=None,\n",
       "                                              n_estimators=100, n_jobs=None,\n",
       "                                              num_parallel_tree=None,\n",
       "                                              predictor=None, random_state=None, ...)),\n",
       "                               ('rf', RandomForestClassifier())],\n",
       "                   final_estimator=LogisticRegression())"
      ]
     },
     "execution_count": 208,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Stacking for the best combinations.\n",
    "estimators2 = [('xg',xgboost.XGBClassifier()),('rf',RandomForestClassifier())]\n",
    "sc_c2=StackingClassifier(estimators=estimators2, final_estimator=LogisticRegression())\n",
    "sc_c2.fit(X_train2, y_train2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 209,
   "id": "0d511a52",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.5"
      ]
     },
     "execution_count": 209,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_sc_c2=sc_c2.predict(X_test2)\n",
    "fpr_sc2, tpr_sc2, thresholds_sc2 = metrics.roc_curve(y_test2, pred_sc_c2)\n",
    "sc_mc2=metrics.auc(fpr_sc2, fpr_sc2)\n",
    "sc_mc2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 210,
   "id": "6115f4f0",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.83      0.88      0.85     10851\n",
      "           1       0.87      0.82      0.84     10951\n",
      "\n",
      "    accuracy                           0.85     21802\n",
      "   macro avg       0.85      0.85      0.85     21802\n",
      "weighted avg       0.85      0.85      0.85     21802\n",
      "\n"
     ]
    }
   ],
   "source": [
    "r_c2=classification_report(y_test2, pred_sc_c2)\n",
    "print(r_c2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 211,
   "id": "b387d89d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Plotting the Roc Curve.\n",
    "roc_auc_c2 = metrics.auc(fpr_sc2, tpr_sc2)\n",
    "plt.plot(fpr_sc2, tpr_sc2, \"k-\", linewidth=2, label=\"Stack Ensemble Best Combination 2(AUC = %.2f)\" % roc_auc_c2)\n",
    "\n",
    "plt.legend(loc=\"lower right\")\n",
    "plt.xlabel(\"False Positive Rate\")\n",
    "plt.ylabel(\"True Positive Rate \")\n",
    "plt.title(\"ROC Curve Comparison of Multiple Models\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "4b651048",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "5172c1d2",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.9.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}