loan_data_set.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "from sklearn.preprocessing import RobustScaler\n",
    "from sklearn.metrics import roc_curve, auc\n",
    "from sklearn.metrics import roc_auc_score\n",
    "from sklearn.cluster import KMeans\n",
    "from sklearn import datasets\n",
    "from io import StringIO\n",
    "from sklearn.tree import export_graphviz\n",
    "from sklearn.model_selection import train_test_split, cross_val_score\n",
    "from sklearn import tree\n",
    "from sklearn import metrics\n",
    "from sklearn.metrics import precision_score, recall_score, f1_score, roc_auc_score, accuracy_score, classification_report,confusion_matrix\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "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>Loan_ID</th>\n",
       "      <th>Gender</th>\n",
       "      <th>Married</th>\n",
       "      <th>Dependents</th>\n",
       "      <th>Education</th>\n",
       "      <th>Self_Employed</th>\n",
       "      <th>ApplicantIncome</th>\n",
       "      <th>CoapplicantIncome</th>\n",
       "      <th>LoanAmount</th>\n",
       "      <th>Loan_Amount_Term</th>\n",
       "      <th>Credit_History</th>\n",
       "      <th>Property_Area</th>\n",
       "      <th>Loan_Status</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>LP001002</td>\n",
       "      <td>Male</td>\n",
       "      <td>No</td>\n",
       "      <td>0</td>\n",
       "      <td>Graduate</td>\n",
       "      <td>No</td>\n",
       "      <td>5849</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>360.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>Urban</td>\n",
       "      <td>Y</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>LP001003</td>\n",
       "      <td>Male</td>\n",
       "      <td>Yes</td>\n",
       "      <td>1</td>\n",
       "      <td>Graduate</td>\n",
       "      <td>No</td>\n",
       "      <td>4583</td>\n",
       "      <td>1508.0</td>\n",
       "      <td>128.0</td>\n",
       "      <td>360.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>Rural</td>\n",
       "      <td>N</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>LP001005</td>\n",
       "      <td>Male</td>\n",
       "      <td>Yes</td>\n",
       "      <td>0</td>\n",
       "      <td>Graduate</td>\n",
       "      <td>Yes</td>\n",
       "      <td>3000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>66.0</td>\n",
       "      <td>360.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>Urban</td>\n",
       "      <td>Y</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>LP001006</td>\n",
       "      <td>Male</td>\n",
       "      <td>Yes</td>\n",
       "      <td>0</td>\n",
       "      <td>Not Graduate</td>\n",
       "      <td>No</td>\n",
       "      <td>2583</td>\n",
       "      <td>2358.0</td>\n",
       "      <td>120.0</td>\n",
       "      <td>360.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>Urban</td>\n",
       "      <td>Y</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>LP001008</td>\n",
       "      <td>Male</td>\n",
       "      <td>No</td>\n",
       "      <td>0</td>\n",
       "      <td>Graduate</td>\n",
       "      <td>No</td>\n",
       "      <td>6000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>141.0</td>\n",
       "      <td>360.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>Urban</td>\n",
       "      <td>Y</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    Loan_ID Gender Married Dependents     Education Self_Employed  \\\n",
       "0  LP001002   Male      No          0      Graduate            No   \n",
       "1  LP001003   Male     Yes          1      Graduate            No   \n",
       "2  LP001005   Male     Yes          0      Graduate           Yes   \n",
       "3  LP001006   Male     Yes          0  Not Graduate            No   \n",
       "4  LP001008   Male      No          0      Graduate            No   \n",
       "\n",
       "   ApplicantIncome  CoapplicantIncome  LoanAmount  Loan_Amount_Term  \\\n",
       "0             5849                0.0         NaN             360.0   \n",
       "1             4583             1508.0       128.0             360.0   \n",
       "2             3000                0.0        66.0             360.0   \n",
       "3             2583             2358.0       120.0             360.0   \n",
       "4             6000                0.0       141.0             360.0   \n",
       "\n",
       "   Credit_History Property_Area Loan_Status  \n",
       "0             1.0         Urban           Y  \n",
       "1             1.0         Rural           N  \n",
       "2             1.0         Urban           Y  \n",
       "3             1.0         Urban           Y  \n",
       "4             1.0         Urban           Y  "
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#load data file\n",
    "data = pd.read_csv(r\"C:\\Users\\Lt col Haider\\OneDrive\\IBA\\semester 4\\Foundations to Data Science\\project\\loan_data_set.csv\")\n",
    "data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(614, 13)"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Data Cleaning"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Loan_ID is a unique variable so drop it\n",
    "data =  data.drop(columns = ['Loan_ID'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "#re-label categorical variables\n",
    "data.Loan_Status.replace({'Y': 0, 'N':1}, inplace=True)\n",
    "data.Gender.replace({'Male':1, 'Female':0}, inplace=True)\n",
    "data.Married.replace({'Yes': 1, 'No':0}, inplace=True)\n",
    "data.Education.replace({'Graduate': 1, 'Not Graduate':0}, inplace=True)\n",
    "data.Self_Employed.replace({'Yes': 1, 'No':0}, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "#create dummy variables for 'Property_Area' and 'Dependents'\n",
    "data = data.join(pd.get_dummies(data.Dependents, prefix='Dependents'))\n",
    "data.drop(columns= ['Dependents'], inplace=True)\n",
    "data = data.join(pd.get_dummies(data.Property_Area, prefix='Property_Area'))\n",
    "data.drop(columns= ['Property_Area'], inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "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>Gender</th>\n",
       "      <th>Married</th>\n",
       "      <th>Education</th>\n",
       "      <th>Self_Employed</th>\n",
       "      <th>ApplicantIncome</th>\n",
       "      <th>CoapplicantIncome</th>\n",
       "      <th>LoanAmount</th>\n",
       "      <th>Loan_Amount_Term</th>\n",
       "      <th>Credit_History</th>\n",
       "      <th>Loan_Status</th>\n",
       "      <th>Dependents_0</th>\n",
       "      <th>Dependents_1</th>\n",
       "      <th>Dependents_2</th>\n",
       "      <th>Dependents_3+</th>\n",
       "      <th>Property_Area_Rural</th>\n",
       "      <th>Property_Area_Semiurban</th>\n",
       "      <th>Property_Area_Urban</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5849</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>360.0</td>\n",
       "      <td>1.0</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>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4583</td>\n",
       "      <td>1508.0</td>\n",
       "      <td>128.0</td>\n",
       "      <td>360.0</td>\n",
       "      <td>1.0</td>\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",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1</td>\n",
       "      <td>1.0</td>\n",
       "      <td>3000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>66.0</td>\n",
       "      <td>360.0</td>\n",
       "      <td>1.0</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>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2583</td>\n",
       "      <td>2358.0</td>\n",
       "      <td>120.0</td>\n",
       "      <td>360.0</td>\n",
       "      <td>1.0</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>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0.0</td>\n",
       "      <td>6000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>141.0</td>\n",
       "      <td>360.0</td>\n",
       "      <td>1.0</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>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Gender  Married  Education  Self_Employed  ApplicantIncome  \\\n",
       "0     1.0      0.0          1            0.0             5849   \n",
       "1     1.0      1.0          1            0.0             4583   \n",
       "2     1.0      1.0          1            1.0             3000   \n",
       "3     1.0      1.0          0            0.0             2583   \n",
       "4     1.0      0.0          1            0.0             6000   \n",
       "\n",
       "   CoapplicantIncome  LoanAmount  Loan_Amount_Term  Credit_History  \\\n",
       "0                0.0         NaN             360.0             1.0   \n",
       "1             1508.0       128.0             360.0             1.0   \n",
       "2                0.0        66.0             360.0             1.0   \n",
       "3             2358.0       120.0             360.0             1.0   \n",
       "4                0.0       141.0             360.0             1.0   \n",
       "\n",
       "   Loan_Status  Dependents_0  Dependents_1  Dependents_2  Dependents_3+  \\\n",
       "0            0             1             0             0              0   \n",
       "1            1             0             1             0              0   \n",
       "2            0             1             0             0              0   \n",
       "3            0             1             0             0              0   \n",
       "4            0             1             0             0              0   \n",
       "\n",
       "   Property_Area_Rural  Property_Area_Semiurban  Property_Area_Urban  \n",
       "0                    0                        0                    1  \n",
       "1                    1                        0                    0  \n",
       "2                    0                        0                    1  \n",
       "3                    0                        0                    1  \n",
       "4                    0                        0                    1  "
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "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>Credit_History</th>\n",
       "      <th>Self_Employed</th>\n",
       "      <th>LoanAmount</th>\n",
       "      <th>Loan_Amount_Term</th>\n",
       "      <th>Gender</th>\n",
       "      <th>Married</th>\n",
       "      <th>Dependents_1</th>\n",
       "      <th>Dependents_0</th>\n",
       "      <th>Loan_Status</th>\n",
       "      <th>Property_Area_Semiurban</th>\n",
       "      <th>Dependents_2</th>\n",
       "      <th>Dependents_3+</th>\n",
       "      <th>CoapplicantIncome</th>\n",
       "      <th>ApplicantIncome</th>\n",
       "      <th>Property_Area_Rural</th>\n",
       "      <th>Education</th>\n",
       "      <th>Property_Area_Urban</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Total</th>\n",
       "      <td>50.000000</td>\n",
       "      <td>32.000000</td>\n",
       "      <td>22.000000</td>\n",
       "      <td>14.00000</td>\n",
       "      <td>13.000000</td>\n",
       "      <td>3.000000</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",
       "      <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>Percent</th>\n",
       "      <td>8.143322</td>\n",
       "      <td>5.211726</td>\n",
       "      <td>3.583062</td>\n",
       "      <td>2.28013</td>\n",
       "      <td>2.117264</td>\n",
       "      <td>0.488599</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",
       "      <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",
       "</div>"
      ],
      "text/plain": [
       "         Credit_History  Self_Employed  LoanAmount  Loan_Amount_Term  \\\n",
       "Total         50.000000      32.000000   22.000000          14.00000   \n",
       "Percent        8.143322       5.211726    3.583062           2.28013   \n",
       "\n",
       "            Gender   Married  Dependents_1  Dependents_0  Loan_Status  \\\n",
       "Total    13.000000  3.000000           0.0           0.0          0.0   \n",
       "Percent   2.117264  0.488599           0.0           0.0          0.0   \n",
       "\n",
       "         Property_Area_Semiurban  Dependents_2  Dependents_3+  \\\n",
       "Total                        0.0           0.0            0.0   \n",
       "Percent                      0.0           0.0            0.0   \n",
       "\n",
       "         CoapplicantIncome  ApplicantIncome  Property_Area_Rural  Education  \\\n",
       "Total                  0.0              0.0                  0.0        0.0   \n",
       "Percent                0.0              0.0                  0.0        0.0   \n",
       "\n",
       "         Property_Area_Urban  \n",
       "Total                    0.0  \n",
       "Percent                  0.0  "
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#check for missing values \n",
    "total = data.isnull().sum().sort_values(ascending = False)\n",
    "percent = (data.isnull().sum()/data.isnull().count()*100).sort_values(ascending = False)\n",
    "pd.concat([total, percent], axis=1, keys=['Total', 'Percent']).transpose()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Normalize data for effective KNN Imputation\n",
    "#will manually normalize as NaN values present so preprocessing can't be used\n",
    "def normalize(data):\n",
    "    return (data - data.min()) * 1.0 / (data.max() - data.min())\n",
    "data = data.apply(normalize)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Apply KNN Imputation\n",
    "from sklearn.impute import KNNImputer\n",
    "imputer = KNNImputer(n_neighbors=15)\n",
    "df = imputer.fit_transform(data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "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>Gender</th>\n",
       "      <th>Married</th>\n",
       "      <th>Education</th>\n",
       "      <th>Self_Employed</th>\n",
       "      <th>ApplicantIncome</th>\n",
       "      <th>CoapplicantIncome</th>\n",
       "      <th>LoanAmount</th>\n",
       "      <th>Loan_Amount_Term</th>\n",
       "      <th>Credit_History</th>\n",
       "      <th>Loan_Status</th>\n",
       "      <th>Dependents_0</th>\n",
       "      <th>Dependents_1</th>\n",
       "      <th>Dependents_2</th>\n",
       "      <th>Dependents_3+</th>\n",
       "      <th>Property_Area_Rural</th>\n",
       "      <th>Property_Area_Semiurban</th>\n",
       "      <th>Property_Area_Urban</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.070489</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.159672</td>\n",
       "      <td>0.74359</td>\n",
       "      <td>1.0</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>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.054830</td>\n",
       "      <td>0.036192</td>\n",
       "      <td>0.172214</td>\n",
       "      <td>0.74359</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.035250</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.082489</td>\n",
       "      <td>0.74359</td>\n",
       "      <td>1.0</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>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.030093</td>\n",
       "      <td>0.056592</td>\n",
       "      <td>0.160637</td>\n",
       "      <td>0.74359</td>\n",
       "      <td>1.0</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>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.072356</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.191027</td>\n",
       "      <td>0.74359</td>\n",
       "      <td>1.0</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>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Gender  Married  Education  Self_Employed  ApplicantIncome  \\\n",
       "0     1.0      0.0        1.0            0.0         0.070489   \n",
       "1     1.0      1.0        1.0            0.0         0.054830   \n",
       "2     1.0      1.0        1.0            1.0         0.035250   \n",
       "3     1.0      1.0        0.0            0.0         0.030093   \n",
       "4     1.0      0.0        1.0            0.0         0.072356   \n",
       "\n",
       "   CoapplicantIncome  LoanAmount  Loan_Amount_Term  Credit_History  \\\n",
       "0           0.000000    0.159672           0.74359             1.0   \n",
       "1           0.036192    0.172214           0.74359             1.0   \n",
       "2           0.000000    0.082489           0.74359             1.0   \n",
       "3           0.056592    0.160637           0.74359             1.0   \n",
       "4           0.000000    0.191027           0.74359             1.0   \n",
       "\n",
       "   Loan_Status  Dependents_0  Dependents_1  Dependents_2  Dependents_3+  \\\n",
       "0          0.0           1.0           0.0           0.0            0.0   \n",
       "1          1.0           0.0           1.0           0.0            0.0   \n",
       "2          0.0           1.0           0.0           0.0            0.0   \n",
       "3          0.0           1.0           0.0           0.0            0.0   \n",
       "4          0.0           1.0           0.0           0.0            0.0   \n",
       "\n",
       "   Property_Area_Rural  Property_Area_Semiurban  Property_Area_Urban  \n",
       "0                  0.0                      0.0                  1.0  \n",
       "1                  1.0                      0.0                  0.0  \n",
       "2                  0.0                      0.0                  1.0  \n",
       "3                  0.0                      0.0                  1.0  \n",
       "4                  0.0                      0.0                  1.0  "
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#convert array back to dataframe\n",
    "data = pd.DataFrame(df)\n",
    "#bring data back in original form\n",
    "data.columns = ['Gender','Married','Education','Self_Employed','ApplicantIncome','CoapplicantIncome','LoanAmount',\n",
    "                'Loan_Amount_Term','Credit_History','Loan_Status','Dependents_0','Dependents_1','Dependents_2','Dependents_3+',\n",
    "                'Property_Area_Rural','Property_Area_Semiurban','Property_Area_Urban']\n",
    "data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "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>Property_Area_Urban</th>\n",
       "      <th>Loan_Amount_Term</th>\n",
       "      <th>Married</th>\n",
       "      <th>Education</th>\n",
       "      <th>Self_Employed</th>\n",
       "      <th>ApplicantIncome</th>\n",
       "      <th>CoapplicantIncome</th>\n",
       "      <th>LoanAmount</th>\n",
       "      <th>Credit_History</th>\n",
       "      <th>Property_Area_Semiurban</th>\n",
       "      <th>Loan_Status</th>\n",
       "      <th>Dependents_0</th>\n",
       "      <th>Dependents_1</th>\n",
       "      <th>Dependents_2</th>\n",
       "      <th>Dependents_3+</th>\n",
       "      <th>Property_Area_Rural</th>\n",
       "      <th>Gender</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Total</th>\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",
       "      <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",
       "      <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>Percent</th>\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",
       "      <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",
       "      <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",
       "</div>"
      ],
      "text/plain": [
       "         Property_Area_Urban  Loan_Amount_Term  Married  Education  \\\n",
       "Total                    0.0               0.0      0.0        0.0   \n",
       "Percent                  0.0               0.0      0.0        0.0   \n",
       "\n",
       "         Self_Employed  ApplicantIncome  CoapplicantIncome  LoanAmount  \\\n",
       "Total              0.0              0.0                0.0         0.0   \n",
       "Percent            0.0              0.0                0.0         0.0   \n",
       "\n",
       "         Credit_History  Property_Area_Semiurban  Loan_Status  Dependents_0  \\\n",
       "Total               0.0                      0.0          0.0           0.0   \n",
       "Percent             0.0                      0.0          0.0           0.0   \n",
       "\n",
       "         Dependents_1  Dependents_2  Dependents_3+  Property_Area_Rural  \\\n",
       "Total             0.0           0.0            0.0                  0.0   \n",
       "Percent           0.0           0.0            0.0                  0.0   \n",
       "\n",
       "         Gender  \n",
       "Total       0.0  \n",
       "Percent     0.0  "
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#check for missing values again\n",
    "total = data.isnull().sum().sort_values(ascending = False)\n",
    "percent = (data.isnull().sum()/data.isnull().count()*100).sort_values(ascending = False)\n",
    "pd.concat([total, percent], axis=1, keys=['Total', 'Percent']).transpose()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[0.48155919 1.36975064 0.52836225 ... 0.64147818 0.7820157  1.42814704]\n",
      " [0.48155919 0.73281659 0.52836225 ... 1.55889948 0.7820157  0.70020801]\n",
      " [0.48155919 0.73281659 0.52836225 ... 0.64147818 0.7820157  1.42814704]\n",
      " ...\n",
      " [0.48155919 0.73281659 0.52836225 ... 0.64147818 0.7820157  1.42814704]\n",
      " [0.48155919 0.73281659 0.52836225 ... 0.64147818 0.7820157  1.42814704]\n",
      " [2.10907171 1.36975064 0.52836225 ... 0.64147818 1.2787467  0.70020801]]\n",
      "(array([  7,   9,  14,  34,  61,  68,  68,  73,  74,  78,  79,  94, 109,\n",
      "       126, 126, 130, 133, 135, 155, 155, 155, 171, 171, 171, 172, 177,\n",
      "       177, 177, 183, 185, 202, 211, 213, 215, 242, 255, 257, 259, 262,\n",
      "       267, 278, 295, 308, 313, 321, 324, 333, 333, 338, 340, 343, 352,\n",
      "       359, 369, 376, 390, 391, 402, 409, 409, 417, 432, 442, 443, 461,\n",
      "       466, 472, 481, 487, 495, 497, 506, 515, 522, 523, 525, 531, 539,\n",
      "       546, 557, 561, 567, 574, 575, 581, 585, 592, 600, 600, 602, 604,\n",
      "       610], dtype=int64), array([13,  5,  7, 13, 13,  7, 13, 13, 13, 13, 13,  7, 13,  4, 13,  6,  7,\n",
      "       13,  4,  6, 13,  4,  6, 13, 13,  5,  6, 13,  4,  4, 13, 13, 13, 13,\n",
      "        7, 13, 13, 13,  7, 13,  6, 13,  6,  7, 13, 13,  4,  6, 13, 13, 13,\n",
      "       13, 13,  6, 13, 13, 13,  5,  4, 13,  5,  6, 13,  4, 13, 13, 13, 13,\n",
      "        6,  7,  7,  6, 13, 13,  6,  6, 13, 13,  7, 13,  6, 13, 13,  7,  5,\n",
      "        7, 13,  5, 13, 13,  6, 13], dtype=int64))\n"
     ]
    }
   ],
   "source": [
    "#check outliers in dataset using Z-Score\n",
    "from scipy import stats\n",
    "z = np.abs(stats.zscore(data))\n",
    "print(z)                            \n",
    "threshold = 3\n",
    "print(np.where(z > 3))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(533, 17)"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#remove outliers\n",
    "data = data[(z < 3).all(axis=1)]\n",
    "data.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Exploratory Data Analysis"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "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>Gender</th>\n",
       "      <th>Married</th>\n",
       "      <th>Education</th>\n",
       "      <th>Self_Employed</th>\n",
       "      <th>ApplicantIncome</th>\n",
       "      <th>CoapplicantIncome</th>\n",
       "      <th>LoanAmount</th>\n",
       "      <th>Loan_Amount_Term</th>\n",
       "      <th>Credit_History</th>\n",
       "      <th>Loan_Status</th>\n",
       "      <th>Dependents_0</th>\n",
       "      <th>Dependents_1</th>\n",
       "      <th>Dependents_2</th>\n",
       "      <th>Dependents_3+</th>\n",
       "      <th>Property_Area_Rural</th>\n",
       "      <th>Property_Area_Semiurban</th>\n",
       "      <th>Property_Area_Urban</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>533.000000</td>\n",
       "      <td>533.000000</td>\n",
       "      <td>533.000000</td>\n",
       "      <td>533.000000</td>\n",
       "      <td>533.000000</td>\n",
       "      <td>533.000000</td>\n",
       "      <td>533.000000</td>\n",
       "      <td>533.000000</td>\n",
       "      <td>533.000000</td>\n",
       "      <td>533.000000</td>\n",
       "      <td>533.000000</td>\n",
       "      <td>533.000000</td>\n",
       "      <td>533.000000</td>\n",
       "      <td>533.0</td>\n",
       "      <td>533.000000</td>\n",
       "      <td>533.000000</td>\n",
       "      <td>533.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>0.799500</td>\n",
       "      <td>0.630394</td>\n",
       "      <td>0.780488</td>\n",
       "      <td>0.132208</td>\n",
       "      <td>0.055329</td>\n",
       "      <td>0.034650</td>\n",
       "      <td>0.182060</td>\n",
       "      <td>0.719187</td>\n",
       "      <td>0.846404</td>\n",
       "      <td>0.307692</td>\n",
       "      <td>0.619137</td>\n",
       "      <td>0.174484</td>\n",
       "      <td>0.178236</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.290807</td>\n",
       "      <td>0.373358</td>\n",
       "      <td>0.335835</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>0.397906</td>\n",
       "      <td>0.482130</td>\n",
       "      <td>0.414305</td>\n",
       "      <td>0.331459</td>\n",
       "      <td>0.036816</td>\n",
       "      <td>0.039609</td>\n",
       "      <td>0.080011</td>\n",
       "      <td>0.107869</td>\n",
       "      <td>0.353320</td>\n",
       "      <td>0.461972</td>\n",
       "      <td>0.486055</td>\n",
       "      <td>0.379882</td>\n",
       "      <td>0.383071</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.454561</td>\n",
       "      <td>0.484150</td>\n",
       "      <td>0.472725</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.358974</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",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.032220</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.131693</td>\n",
       "      <td>0.743590</td>\n",
       "      <td>1.000000</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",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.044119</td>\n",
       "      <td>0.031248</td>\n",
       "      <td>0.169320</td>\n",
       "      <td>0.743590</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</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>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.066172</td>\n",
       "      <td>0.054792</td>\n",
       "      <td>0.218524</td>\n",
       "      <td>0.743590</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.253766</td>\n",
       "      <td>0.215518</td>\n",
       "      <td>0.536903</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           Gender     Married   Education  Self_Employed  ApplicantIncome  \\\n",
       "count  533.000000  533.000000  533.000000     533.000000       533.000000   \n",
       "mean     0.799500    0.630394    0.780488       0.132208         0.055329   \n",
       "std      0.397906    0.482130    0.414305       0.331459         0.036816   \n",
       "min      0.000000    0.000000    0.000000       0.000000         0.000000   \n",
       "25%      1.000000    0.000000    1.000000       0.000000         0.032220   \n",
       "50%      1.000000    1.000000    1.000000       0.000000         0.044119   \n",
       "75%      1.000000    1.000000    1.000000       0.000000         0.066172   \n",
       "max      1.000000    1.000000    1.000000       1.000000         0.253766   \n",
       "\n",
       "       CoapplicantIncome  LoanAmount  Loan_Amount_Term  Credit_History  \\\n",
       "count         533.000000  533.000000        533.000000      533.000000   \n",
       "mean            0.034650    0.182060          0.719187        0.846404   \n",
       "std             0.039609    0.080011          0.107869        0.353320   \n",
       "min             0.000000    0.000000          0.358974        0.000000   \n",
       "25%             0.000000    0.131693          0.743590        1.000000   \n",
       "50%             0.031248    0.169320          0.743590        1.000000   \n",
       "75%             0.054792    0.218524          0.743590        1.000000   \n",
       "max             0.215518    0.536903          1.000000        1.000000   \n",
       "\n",
       "       Loan_Status  Dependents_0  Dependents_1  Dependents_2  Dependents_3+  \\\n",
       "count   533.000000    533.000000    533.000000    533.000000          533.0   \n",
       "mean      0.307692      0.619137      0.174484      0.178236            0.0   \n",
       "std       0.461972      0.486055      0.379882      0.383071            0.0   \n",
       "min       0.000000      0.000000      0.000000      0.000000            0.0   \n",
       "25%       0.000000      0.000000      0.000000      0.000000            0.0   \n",
       "50%       0.000000      1.000000      0.000000      0.000000            0.0   \n",
       "75%       1.000000      1.000000      0.000000      0.000000            0.0   \n",
       "max       1.000000      1.000000      1.000000      1.000000            0.0   \n",
       "\n",
       "       Property_Area_Rural  Property_Area_Semiurban  Property_Area_Urban  \n",
       "count           533.000000               533.000000           533.000000  \n",
       "mean              0.290807                 0.373358             0.335835  \n",
       "std               0.454561                 0.484150             0.472725  \n",
       "min               0.000000                 0.000000             0.000000  \n",
       "25%               0.000000                 0.000000             0.000000  \n",
       "50%               0.000000                 0.000000             0.000000  \n",
       "75%               1.000000                 1.000000             1.000000  \n",
       "max               1.000000                 1.000000             1.000000  "
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "temp = data[\"Loan_Status\"].value_counts()\n",
    "df = pd.DataFrame({'Loan_Status': temp.index,'values': temp.values})\n",
    "plt.figure(figsize = (6,6))\n",
    "plt.title('Default Clients - target value - data unbalance\\n (Not Default = 0, Default = 1)')\n",
    "sns.set_color_codes(\"pastel\")\n",
    "sns.barplot(x = 'Loan_Status', y=\"values\", data=df)\n",
    "locs, labels = plt.xticks()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Data is imbalanced."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "numerical = ['ApplicantIncome','CoapplicantIncome','LoanAmount','Loan_Amount_Term']\n",
    "categorical = ['Credit_History','Gender','Married','Education','Self_Employed','Dependents_0','Dependents_1','Dependents_2',\n",
    "               'Dependents_3+','Property_Area_Rural','Property_Area_Semiurban','Property_Area_Urban']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 504x504 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#T Test for numerical columns\n",
    "p=[]\n",
    "from scipy.stats import ttest_ind\n",
    "\n",
    "for i in numerical:\n",
    "    df1=data.groupby('Loan_Status').get_group(0)\n",
    "    df2=data.groupby('Loan_Status').get_group(1)\n",
    "    t,pvalue=ttest_ind(df1[i],df2[i])\n",
    "    p.append(1-pvalue)\n",
    "plt.figure(figsize=(7,7))\n",
    "sns.barplot(x=p, y=numerical)\n",
    "plt.title('Best Numerical Features')\n",
    "plt.axvline(x=(1-0.05),color='r')\n",
    "plt.xlabel('1-p value')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "'CoappliantIncome' has the greatest importance. None of them show statistical significance though."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 504x504 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Chi Square test for Categorical Columns\n",
    "from scipy.stats import chi2_contingency\n",
    "l=[]\n",
    "for i in categorical:\n",
    "    pvalue  = chi2_contingency(pd.crosstab(data['Loan_Status'],data[i]))[1]\n",
    "    l.append(1-pvalue)\n",
    "plt.figure(figsize=(7,7))\n",
    "sns.barplot(x=l, y=categorical)\n",
    "plt.title('Best Categorical Features')\n",
    "plt.axvline(x=(1-0.05),color='r')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "'Credit_History', 'Property_Area_Rural' and 'Property_Area_Semiurban' show statistical significance. 'Dependents_3+' has very low significance, so drop it."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "data =  data.drop(columns = ['Dependents_3+'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "#drop target variable: 'Default'\n",
    "features =  data.drop(columns = ['Loan_Status'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "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>Gender</th>\n",
       "      <th>Married</th>\n",
       "      <th>Education</th>\n",
       "      <th>Self_Employed</th>\n",
       "      <th>ApplicantIncome</th>\n",
       "      <th>CoapplicantIncome</th>\n",
       "      <th>LoanAmount</th>\n",
       "      <th>Loan_Amount_Term</th>\n",
       "      <th>Credit_History</th>\n",
       "      <th>Dependents_0</th>\n",
       "      <th>Dependents_1</th>\n",
       "      <th>Dependents_2</th>\n",
       "      <th>Property_Area_Rural</th>\n",
       "      <th>Property_Area_Semiurban</th>\n",
       "      <th>Property_Area_Urban</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Gender</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.371364</td>\n",
       "      <td>0.050837</td>\n",
       "      <td>0.015270</td>\n",
       "      <td>0.012556</td>\n",
       "      <td>0.202383</td>\n",
       "      <td>0.171739</td>\n",
       "      <td>0.106348</td>\n",
       "      <td>0.008722</td>\n",
       "      <td>0.151308</td>\n",
       "      <td>0.044517</td>\n",
       "      <td>0.147746</td>\n",
       "      <td>0.082559</td>\n",
       "      <td>0.121970</td>\n",
       "      <td>0.045531</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Married</th>\n",
       "      <td>0.371364</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.021116</td>\n",
       "      <td>0.019860</td>\n",
       "      <td>0.007458</td>\n",
       "      <td>0.259718</td>\n",
       "      <td>0.205034</td>\n",
       "      <td>0.120622</td>\n",
       "      <td>0.025103</td>\n",
       "      <td>0.329109</td>\n",
       "      <td>0.137251</td>\n",
       "      <td>0.275940</td>\n",
       "      <td>0.014676</td>\n",
       "      <td>0.007126</td>\n",
       "      <td>0.006813</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Education</th>\n",
       "      <td>0.050837</td>\n",
       "      <td>0.021116</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.010973</td>\n",
       "      <td>0.151525</td>\n",
       "      <td>0.019276</td>\n",
       "      <td>0.135597</td>\n",
       "      <td>0.090932</td>\n",
       "      <td>0.075710</td>\n",
       "      <td>0.041435</td>\n",
       "      <td>0.006991</td>\n",
       "      <td>0.037264</td>\n",
       "      <td>0.099567</td>\n",
       "      <td>0.062626</td>\n",
       "      <td>0.031602</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Self_Employed</th>\n",
       "      <td>0.015270</td>\n",
       "      <td>0.019860</td>\n",
       "      <td>0.010973</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.251736</td>\n",
       "      <td>0.040629</td>\n",
       "      <td>0.123804</td>\n",
       "      <td>0.052935</td>\n",
       "      <td>0.030584</td>\n",
       "      <td>0.095227</td>\n",
       "      <td>0.117008</td>\n",
       "      <td>0.004544</td>\n",
       "      <td>0.019643</td>\n",
       "      <td>0.001280</td>\n",
       "      <td>0.017577</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ApplicantIncome</th>\n",
       "      <td>0.012556</td>\n",
       "      <td>0.007458</td>\n",
       "      <td>0.151525</td>\n",
       "      <td>0.251736</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.228698</td>\n",
       "      <td>0.441993</td>\n",
       "      <td>0.056216</td>\n",
       "      <td>0.015630</td>\n",
       "      <td>0.081156</td>\n",
       "      <td>0.054636</td>\n",
       "      <td>0.034840</td>\n",
       "      <td>0.001924</td>\n",
       "      <td>0.026774</td>\n",
       "      <td>0.029272</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CoapplicantIncome</th>\n",
       "      <td>0.202383</td>\n",
       "      <td>0.259718</td>\n",
       "      <td>0.019276</td>\n",
       "      <td>0.040629</td>\n",
       "      <td>0.228698</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.314743</td>\n",
       "      <td>0.089052</td>\n",
       "      <td>0.003216</td>\n",
       "      <td>0.010898</td>\n",
       "      <td>0.011022</td>\n",
       "      <td>0.013498</td>\n",
       "      <td>0.090991</td>\n",
       "      <td>0.017835</td>\n",
       "      <td>0.069229</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>LoanAmount</th>\n",
       "      <td>0.171739</td>\n",
       "      <td>0.205034</td>\n",
       "      <td>0.135597</td>\n",
       "      <td>0.123804</td>\n",
       "      <td>0.441993</td>\n",
       "      <td>0.314743</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.022247</td>\n",
       "      <td>0.012941</td>\n",
       "      <td>0.128236</td>\n",
       "      <td>0.076126</td>\n",
       "      <td>0.090088</td>\n",
       "      <td>0.061388</td>\n",
       "      <td>0.008507</td>\n",
       "      <td>0.050316</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Loan_Amount_Term</th>\n",
       "      <td>0.106348</td>\n",
       "      <td>0.120622</td>\n",
       "      <td>0.090932</td>\n",
       "      <td>0.052935</td>\n",
       "      <td>0.056216</td>\n",
       "      <td>0.089052</td>\n",
       "      <td>0.022247</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.022029</td>\n",
       "      <td>0.088680</td>\n",
       "      <td>0.117020</td>\n",
       "      <td>0.005350</td>\n",
       "      <td>0.022133</td>\n",
       "      <td>0.062503</td>\n",
       "      <td>0.085296</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Credit_History</th>\n",
       "      <td>0.008722</td>\n",
       "      <td>0.025103</td>\n",
       "      <td>0.075710</td>\n",
       "      <td>0.030584</td>\n",
       "      <td>0.015630</td>\n",
       "      <td>0.003216</td>\n",
       "      <td>0.012941</td>\n",
       "      <td>0.022029</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.042542</td>\n",
       "      <td>0.023092</td>\n",
       "      <td>0.006365</td>\n",
       "      <td>0.045168</td>\n",
       "      <td>0.061890</td>\n",
       "      <td>0.019953</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Dependents_0</th>\n",
       "      <td>0.151308</td>\n",
       "      <td>0.329109</td>\n",
       "      <td>0.041435</td>\n",
       "      <td>0.095227</td>\n",
       "      <td>0.081156</td>\n",
       "      <td>0.010898</td>\n",
       "      <td>0.128236</td>\n",
       "      <td>0.088680</td>\n",
       "      <td>0.042542</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.586171</td>\n",
       "      <td>0.593791</td>\n",
       "      <td>0.093872</td>\n",
       "      <td>0.001663</td>\n",
       "      <td>0.088561</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Dependents_1</th>\n",
       "      <td>0.044517</td>\n",
       "      <td>0.137251</td>\n",
       "      <td>0.006991</td>\n",
       "      <td>0.117008</td>\n",
       "      <td>0.054636</td>\n",
       "      <td>0.011022</td>\n",
       "      <td>0.076126</td>\n",
       "      <td>0.117020</td>\n",
       "      <td>0.023092</td>\n",
       "      <td>0.586171</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.214111</td>\n",
       "      <td>0.076689</td>\n",
       "      <td>0.017603</td>\n",
       "      <td>0.091770</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Dependents_2</th>\n",
       "      <td>0.147746</td>\n",
       "      <td>0.275940</td>\n",
       "      <td>0.037264</td>\n",
       "      <td>0.004544</td>\n",
       "      <td>0.034840</td>\n",
       "      <td>0.013498</td>\n",
       "      <td>0.090088</td>\n",
       "      <td>0.005350</td>\n",
       "      <td>0.006365</td>\n",
       "      <td>0.593791</td>\n",
       "      <td>0.214111</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.017559</td>\n",
       "      <td>0.015516</td>\n",
       "      <td>0.000993</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Property_Area_Rural</th>\n",
       "      <td>0.082559</td>\n",
       "      <td>0.014676</td>\n",
       "      <td>0.099567</td>\n",
       "      <td>0.019643</td>\n",
       "      <td>0.001924</td>\n",
       "      <td>0.090991</td>\n",
       "      <td>0.061388</td>\n",
       "      <td>0.022133</td>\n",
       "      <td>0.045168</td>\n",
       "      <td>0.093872</td>\n",
       "      <td>0.076689</td>\n",
       "      <td>0.017559</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.494280</td>\n",
       "      <td>0.455349</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Property_Area_Semiurban</th>\n",
       "      <td>0.121970</td>\n",
       "      <td>0.007126</td>\n",
       "      <td>0.062626</td>\n",
       "      <td>0.001280</td>\n",
       "      <td>0.026774</td>\n",
       "      <td>0.017835</td>\n",
       "      <td>0.008507</td>\n",
       "      <td>0.062503</td>\n",
       "      <td>0.061890</td>\n",
       "      <td>0.001663</td>\n",
       "      <td>0.017603</td>\n",
       "      <td>0.015516</td>\n",
       "      <td>0.494280</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.548881</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Property_Area_Urban</th>\n",
       "      <td>0.045531</td>\n",
       "      <td>0.006813</td>\n",
       "      <td>0.031602</td>\n",
       "      <td>0.017577</td>\n",
       "      <td>0.029272</td>\n",
       "      <td>0.069229</td>\n",
       "      <td>0.050316</td>\n",
       "      <td>0.085296</td>\n",
       "      <td>0.019953</td>\n",
       "      <td>0.088561</td>\n",
       "      <td>0.091770</td>\n",
       "      <td>0.000993</td>\n",
       "      <td>0.455349</td>\n",
       "      <td>0.548881</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                           Gender   Married  Education  Self_Employed  \\\n",
       "Gender                   1.000000  0.371364   0.050837       0.015270   \n",
       "Married                  0.371364  1.000000   0.021116       0.019860   \n",
       "Education                0.050837  0.021116   1.000000       0.010973   \n",
       "Self_Employed            0.015270  0.019860   0.010973       1.000000   \n",
       "ApplicantIncome          0.012556  0.007458   0.151525       0.251736   \n",
       "CoapplicantIncome        0.202383  0.259718   0.019276       0.040629   \n",
       "LoanAmount               0.171739  0.205034   0.135597       0.123804   \n",
       "Loan_Amount_Term         0.106348  0.120622   0.090932       0.052935   \n",
       "Credit_History           0.008722  0.025103   0.075710       0.030584   \n",
       "Dependents_0             0.151308  0.329109   0.041435       0.095227   \n",
       "Dependents_1             0.044517  0.137251   0.006991       0.117008   \n",
       "Dependents_2             0.147746  0.275940   0.037264       0.004544   \n",
       "Property_Area_Rural      0.082559  0.014676   0.099567       0.019643   \n",
       "Property_Area_Semiurban  0.121970  0.007126   0.062626       0.001280   \n",
       "Property_Area_Urban      0.045531  0.006813   0.031602       0.017577   \n",
       "\n",
       "                         ApplicantIncome  CoapplicantIncome  LoanAmount  \\\n",
       "Gender                          0.012556           0.202383    0.171739   \n",
       "Married                         0.007458           0.259718    0.205034   \n",
       "Education                       0.151525           0.019276    0.135597   \n",
       "Self_Employed                   0.251736           0.040629    0.123804   \n",
       "ApplicantIncome                 1.000000           0.228698    0.441993   \n",
       "CoapplicantIncome               0.228698           1.000000    0.314743   \n",
       "LoanAmount                      0.441993           0.314743    1.000000   \n",
       "Loan_Amount_Term                0.056216           0.089052    0.022247   \n",
       "Credit_History                  0.015630           0.003216    0.012941   \n",
       "Dependents_0                    0.081156           0.010898    0.128236   \n",
       "Dependents_1                    0.054636           0.011022    0.076126   \n",
       "Dependents_2                    0.034840           0.013498    0.090088   \n",
       "Property_Area_Rural             0.001924           0.090991    0.061388   \n",
       "Property_Area_Semiurban         0.026774           0.017835    0.008507   \n",
       "Property_Area_Urban             0.029272           0.069229    0.050316   \n",
       "\n",
       "                         Loan_Amount_Term  Credit_History  Dependents_0  \\\n",
       "Gender                           0.106348        0.008722      0.151308   \n",
       "Married                          0.120622        0.025103      0.329109   \n",
       "Education                        0.090932        0.075710      0.041435   \n",
       "Self_Employed                    0.052935        0.030584      0.095227   \n",
       "ApplicantIncome                  0.056216        0.015630      0.081156   \n",
       "CoapplicantIncome                0.089052        0.003216      0.010898   \n",
       "LoanAmount                       0.022247        0.012941      0.128236   \n",
       "Loan_Amount_Term                 1.000000        0.022029      0.088680   \n",
       "Credit_History                   0.022029        1.000000      0.042542   \n",
       "Dependents_0                     0.088680        0.042542      1.000000   \n",
       "Dependents_1                     0.117020        0.023092      0.586171   \n",
       "Dependents_2                     0.005350        0.006365      0.593791   \n",
       "Property_Area_Rural              0.022133        0.045168      0.093872   \n",
       "Property_Area_Semiurban          0.062503        0.061890      0.001663   \n",
       "Property_Area_Urban              0.085296        0.019953      0.088561   \n",
       "\n",
       "                         Dependents_1  Dependents_2  Property_Area_Rural  \\\n",
       "Gender                       0.044517      0.147746             0.082559   \n",
       "Married                      0.137251      0.275940             0.014676   \n",
       "Education                    0.006991      0.037264             0.099567   \n",
       "Self_Employed                0.117008      0.004544             0.019643   \n",
       "ApplicantIncome              0.054636      0.034840             0.001924   \n",
       "CoapplicantIncome            0.011022      0.013498             0.090991   \n",
       "LoanAmount                   0.076126      0.090088             0.061388   \n",
       "Loan_Amount_Term             0.117020      0.005350             0.022133   \n",
       "Credit_History               0.023092      0.006365             0.045168   \n",
       "Dependents_0                 0.586171      0.593791             0.093872   \n",
       "Dependents_1                 1.000000      0.214111             0.076689   \n",
       "Dependents_2                 0.214111      1.000000             0.017559   \n",
       "Property_Area_Rural          0.076689      0.017559             1.000000   \n",
       "Property_Area_Semiurban      0.017603      0.015516             0.494280   \n",
       "Property_Area_Urban          0.091770      0.000993             0.455349   \n",
       "\n",
       "                         Property_Area_Semiurban  Property_Area_Urban  \n",
       "Gender                                  0.121970             0.045531  \n",
       "Married                                 0.007126             0.006813  \n",
       "Education                               0.062626             0.031602  \n",
       "Self_Employed                           0.001280             0.017577  \n",
       "ApplicantIncome                         0.026774             0.029272  \n",
       "CoapplicantIncome                       0.017835             0.069229  \n",
       "LoanAmount                              0.008507             0.050316  \n",
       "Loan_Amount_Term                        0.062503             0.085296  \n",
       "Credit_History                          0.061890             0.019953  \n",
       "Dependents_0                            0.001663             0.088561  \n",
       "Dependents_1                            0.017603             0.091770  \n",
       "Dependents_2                            0.015516             0.000993  \n",
       "Property_Area_Rural                     0.494280             0.455349  \n",
       "Property_Area_Semiurban                 1.000000             0.548881  \n",
       "Property_Area_Urban                     0.548881             1.000000  "
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#correlation analysis\n",
    "corr_matrix = features.corr().abs()\n",
    "corr_matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(533, 16)"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Feature Selection \n",
    "\n",
    "#Select upper triangle of correlation matrix\n",
    "upper = corr_matrix.where(np.triu(np.ones(corr_matrix.shape), k=1).astype(np.bool))\n",
    "to_drop = [column for column in upper.columns if any(upper[column] > 0.95)]\n",
    "#Drop highly correlated attributes\n",
    "data = data.drop(data[to_drop], axis=1)\n",
    "data.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "No highly correlated attributes found."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "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>Loan_Status</th>\n",
       "      <th>0.0</th>\n",
       "      <th>1.0</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Married</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0.000000</th>\n",
       "      <td>124</td>\n",
       "      <td>72</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.400000</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.666667</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.933333</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1.000000</th>\n",
       "      <td>242</td>\n",
       "      <td>92</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "Loan_Status  0.0  1.0\n",
       "Married              \n",
       "0.000000     124   72\n",
       "0.400000       1    0\n",
       "0.666667       1    0\n",
       "0.933333       1    0\n",
       "1.000000     242   92"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.crosstab(data.Married, data.Loan_Status)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Those not married have a higher percentage on defaulting."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>Loan_Status</th>\n",
       "      <th>0.0</th>\n",
       "      <th>1.0</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Education</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0.0</th>\n",
       "      <td>73</td>\n",
       "      <td>44</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1.0</th>\n",
       "      <td>296</td>\n",
       "      <td>120</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "Loan_Status  0.0  1.0\n",
       "Education            \n",
       "0.0           73   44\n",
       "1.0          296  120"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.crosstab(data.Education, data.Loan_Status)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Graduates are more in number and thus are more in number for both default and not defaut. Not Graduates have a higher percentage of defaulting which makes sense. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "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>Loan_Status</th>\n",
       "      <th>0.0</th>\n",
       "      <th>1.0</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Dependents_0</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0.0</th>\n",
       "      <td>142</td>\n",
       "      <td>61</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1.0</th>\n",
       "      <td>227</td>\n",
       "      <td>103</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "Loan_Status   0.0  1.0\n",
       "Dependents_0          \n",
       "0.0           142   61\n",
       "1.0           227  103"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.crosstab(data.Dependents_0, data.Loan_Status)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Percentage of default for those with no dependents and those with dependents is almost the same."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1008x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "class_0 = data.loc[data['Loan_Status'] == 0][\"ApplicantIncome\"]\n",
    "class_1 = data.loc[data['Loan_Status'] == 1][\"ApplicantIncome\"]\n",
    "plt.figure(figsize = (14,6))\n",
    "plt.title('Default amount (Density Plot)')\n",
    "sns.set_color_codes(\"pastel\")\n",
    "sns.distplot(class_1,kde=True,bins=200, color=\"red\")\n",
    "sns.distplot(class_0,kde=True,bins=200, color=\"green\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Data is right skewed. Those with less Applicant Incomes have higher defaulting tendency. The highest default is for value 0.04."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1008x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "class_0 = data.loc[data['Loan_Status'] == 0][\"LoanAmount\"]\n",
    "class_1 = data.loc[data['Loan_Status'] == 1][\"LoanAmount\"]\n",
    "plt.figure(figsize = (14,6))\n",
    "plt.title('Default amount (Density Plot)')\n",
    "sns.set_color_codes(\"pastel\")\n",
    "sns.distplot(class_1,kde=True,bins=200, color=\"red\")\n",
    "sns.distplot(class_0,kde=True,bins=200, color=\"green\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Data has a normal distribution."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Pre-Processing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Split test and train data\n",
    "x = data.drop(columns = 'Loan_Status')\n",
    "y = data['Loan_Status']\n",
    "\n",
    "#will train on 70% of the data, test on the remaining 30%\n",
    "x_train, x_test, y_train, y_test = train_test_split(x, y, test_size=0.30, random_state=50)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.0    0.692308\n",
       "1.0    0.307692\n",
       "Name: Loan_Status, dtype: float64"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y.value_counts(normalize=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Majority class (not default) has 69% samples. Minority class (default) has 31% samples. Data is imbalanced. Thus, will apply SMOTE."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "from imblearn.over_sampling import SMOTE\n",
    "#so as not to lose any important information, apply SMOTE (Synthetic Minority Oversampling Technique) to fix class imbalance\n",
    "sm = SMOTE(random_state=10)\n",
    "x_smote, y_smote = sm.fit_sample(x_train, y_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Modeling"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 1. Logistic Regression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[99 14]\n",
      " [14 33]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "         0.0       0.88      0.88      0.88       113\n",
      "         1.0       0.70      0.70      0.70        47\n",
      "\n",
      "    accuracy                           0.82       160\n",
      "   macro avg       0.79      0.79      0.79       160\n",
      "weighted avg       0.82      0.82      0.82       160\n",
      "\n"
     ]
    }
   ],
   "source": [
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "param_grid = {'C': [0.001, 0.01, 1, 10]}\n",
    "\n",
    "lr = LogisticRegression(penalty='l2', class_weight='balanced')\n",
    "lr = GridSearchCV(estimator = lr, param_grid = param_grid, n_jobs = -1)\n",
    "lr.fit(x_train, y_train)\n",
    "y_pred = lr.predict(x_test)\n",
    "\n",
    "print(confusion_matrix(y_test, y_pred))\n",
    "print(classification_report(y_test, y_pred))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 0, 'False Positive Rate')"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYIAAAGDCAYAAAAmphcsAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjMsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+AADFEAAAgAElEQVR4nO3dd5hU5fn/8ffNsrCUpYN0QUAUkOZKMRohNtREjAUpFixgjCVGU8zXVGN+STSKPQqoqAhYEg0mqLG3sDQBpYggKF2qVBe23L8/zkHHZcssu7NnZ+fzuq65mFNmzn1mh3Ofp8zzmLsjIiKpq0bUAYiISLSUCEREUpwSgYhIilMiEBFJcUoEIiIpTolARCTFKRFI3MxslJn9N+o4qhIz221mR0Rw3A5m5mZWs7KPnQhmttjMBh3C6/SdrABKBEnKzD4zs6/CC9FGM5tkZvUTeUx3f8rdT0vkMWKZ2fFm9oaZ7TKzHWb2opl1q6zjFxHPW2Z2Zew6d6/v7isTdLwjzexZM9sSnv+HZnajmaUl4niHKkxIncvzHu7e3d3fKuU4ByW/yv5OVldKBMntB+5eH+gN9AF+FXE8h6Sou1ozGwj8F/gX0BroCCwE3k/EHXhVu7M2s07ALGANcIy7NwQuALKAzAo+VmTnXtU+95Tl7nok4QP4DDglZvl24D8xy7WBvwGrgS+Ah4A6MduHAguAncCnwJBwfUPgEWADsA64DUgLt40G3gufPwT8rVBM/wJuDJ+3Bv4BbAZWAdfH7Pd74Dlgcnj8K4s4v3eBB4tY/xLwRPh8ELAW+D9gS/iZjIrnM4h57S+BjcCTQGPg32HM28PnbcP9/wTkAznAbuD+cL0DncPnk4AHgP8Auwgu5J1i4jkNWAbsAB4E3i7q3MN9J8f+PYvY3iE89qXh+W0BbonZ3g+YCXwZ/i3vB2rFbHfgGmA5sCpcdw9B4tkJzANOjNk/LfycPw3PbR7QDngnfK894edyYbj/9wm+X18C/wN6Fvru/hL4ENgH1CTm+xzGPjeM4wvgrnD96vBYu8PHQGK+k+E+3YFXgW3ha/8v6v+ryfCIPAA9DvEP9+3/OG2Bj4B7YrbfDUwHmhDcQb4I/Dnc1i+8GJ1KUCpsAxwVbnsBeBioB7QAZgNXhdu+/k8HfDe8aFi43Bj4iiAB1AgvFL8FagFHACuB08N9fw/kAueE+9YpdG51CS66g4s478uADeHzQUAecBfBRf+k8ILUNY7P4MBr/xq+tg7QFDgvPH4m8CzwQsyx36LQhZuDE8G28POtCTwFTAu3NQsvbOeG234SfgbFJYKNwGUl/P07hMeeEMbei+CienS4/VhgQHisDsBS4IZCcb8afjYHkuNF4WdQE7gpjCEj3PZzgu9YV8DC4zUt/BmEy32BTUB/ggRyKcH3tXbMd3cBQSKpE7PuwPd5JnBx+Lw+MKDQOdeMOdZovvlOZhIkvZuAjHC5f9T/V5PhEXkAehziHy74j7Ob4O7MgdeBRuE2I7ggxt6NDuSbO7+HgXFFvOdh4cUktuQwAngzfB77n84I7tC+Gy6PAd4In/cHVhd6718Bj4XPfw+8U8K5tQ3P6agitg0BcsPngwgu5vVitj8D/CaOz2AQsP/Aha6YOHoD22OW36L0RDAxZtuZwMfh80uAmTHbjCCRFpcIcglLacVsP3BRbBuzbjYwvJj9bwCeLxT390r5jm0HeoXPlwFDi9mvcCL4O/DHQvssA06K+e5eXsT3+UAieAf4A9CsmHMuLhGMAOYn8v9ddX2ofi65nePur5nZScAUgrvOL4HmBHe188zswL5GcHcGwZ3YjCLe73AgHdgQ87oaBBesb3F3N7NpBP/53gFGElRnHHif1mb2ZcxL0giqew446D1jbAcKgFbAx4W2tSKoBvl6X3ffE7P8OUGppLTPAGCzu+d8vdGsLjCOINk0Dldnmlmau+eXEG+sjTHP9xLc0RLG9PU5h5/f2hLeZyvBuR7S8czsSIKSUhbB51CToJQW61t/AzO7CbgyjNWBBgTfKQi+M5/GEQ8Ef/9Lzey6mHW1wvct8tiFXAHcCnxsZquAP7j7v+M4bllilBhqLK4G3P1tgrvRv4WrthBU03R390bho6EHDcsQ/CfsVMRbrSEoETSLeV0Dd+9ezKGnAueb2eEEpYB/xLzPqpj3aOTume5+ZmzYJZzPHoLqgQuK2DyMoPRzQGMzqxez3B5YH8dnUFQMNxFUffR39wYE1V8QJJASY47DBoKSTvCGQXZqW/zuvEZQTXWo/k6QRLuE5/J/fHMeB3x9PmZ2IkG9/TCgsbs3Iqg+PPCa4r4zRVkD/KnQ37+uu08t6tiFuftydx9BUDX5V+C58G9c2udflhglhhJB9XE3cKqZ9Xb3AoK643Fm1gLAzNqY2enhvo8Al5nZyWZWI9x2lLtvIOipc6eZNQi3dQpLHAdx9/kEDasTgVfc/UAJYDaw08x+aWZ1zCzNzHqY2XFlOJ+bCe4qrzezTDNrbGa3EVTv/KHQvn8ws1rhxez7wLNxfAZFySRIHl+aWRPgd4W2f0HQ3nEo/gMcY2bnhD1lrgFalrD/74DjzewOM2sZxt/ZzCabWaM4jpdJ0Cax28yOAq6OY/88gr9nTTP7LUGJ4ICJwB/NrIsFeppZ03Bb4c9lAvAjM+sf7lvPzM4ys7h6O5nZRWbWPPwbHvhO5YexFVD83+DfQEszu8HMaoffm/7xHDPVKRFUE+6+GXiCoH4cgru7FUC2me0kuMPsGu47m6DRdRzBXd/bBMV5COqyawFLCKponqPkKoqpwCkEVVMHYskHfkBQx76K4O58IkGPpHjP5z3gdILG1Q0EVT59gBPcfXnMrhvDONcTNM7+yN0PVCcV+xkU426ChtctQDbwcqHt9xCUgLab2b3xnkt4PlsISji3E1T7dCPoGbOvmP0/JUh6HYDFZraDoMQ1l6BdqDQ/I6iu20VwYX66lP1fIeiR9QnBZ53Dt6tv7iJof/kvQYJ5hOCzgqDN53Ez+9LMhrn7XII2o/sJ/jYrCOry4zWE4Jx3E3zmw909x933EvTeej881oDYF7n7LoIOED8g+F4sBwaX4bgp60CPD5GkE/4SdbK7l1TFUiWZWQ2C7quj3P3NqOOR1KYSgUglMbPTzayRmdXmmzr77IjDElEiEKlEAwl6tWwhqL44x92/ijYkEVUNiYikPJUIRERSnBKBiEiKS7pfFjdr1sw7dOgQdRgiIkll3rx5W9y9eVHbki4RdOjQgblz50YdhohIUjGzz4vbpqohEZEUp0QgIpLilAhERFKcEoGISIpTIhARSXFKBCIiKU6JQEQkxSkRiIikOCUCEZEUl7BEYGaPmtkmM1tUzHYzs3vNbIWZfWhmfRMVi4iIFC+RJYJJBFPOFecMoEv4GEsw2baIiFSyhI015O7vmFmHEnYZCjzhwYQI2eHMTa3CCdRFJJmMHw9TppS+nxyyAqBG795w990V/t5RthG04duTY68N1x3EzMaa2Vwzm7t58+ZKCU5EymDKFFiwIOooqq2HWvXjmi5nk4sl5P2jHH20qDMqcro0dx8PjAfIysrSlGoiVVHv3vDWW1FHUe3c+/py7nr1E37QqzUM65WQY0SZCNYC7WKW2wLrI4pFRKTKWbphJ+Ne+4Rz+7Thjgt6kVaj+pUIpgPXmtk0oD+wQ+0DIiLfOLpVA565aiB92zdOWBKABCYCM5sKDAKamdla4HdAOoC7PwTMAM4EVgB7gcsSFYuISLJwd/7y0scM6NSUwV1bcFyHJgk/ZiJ7DY0oZbsD1yTq+CIiyaagwPnt9EVMzl4NwOCuLSrluPplsYgcmvHjYdCg4KEeQ+WWX+D86p8fMTl7NVeddAQ3n3FUpR1biUBEDk1sl9HevWHkyGjjSWL5Bc7Pn13I03PXcP33OnPzkKMwS1ybQGFJN3m9iFQh6jJaIWoYZNRK46ZTj+S6k7tU+vGVCEREIrI/r4Cte/bRqmEd/nROj0otBcRS1ZCISAT25eXz46fmcf7fZ7JnX15kSQCUCEREKl1Obj5jnpjHa0s38aNBnahXO9rKGVUNiVRVVX0gtwULgjYCKZO9+/O48vG5zFy5lb+edwwXHtc+6pBUIhCpsqr6QG7qKXRIbn95Gdkrt3LnBb2qRBIAlQhEqjb1yql2bjztSE46sjmDj6qcH4vFQyUCEZEE27E3lz+8uJic3HwaZKRXqSQASgQiIgm1bc9+RkzI5qns1SxatyPqcIqkqiERkQTZsnsfoybM4rOtexh/ybFkVcIAcodCiUBSU1XvkQPqlZPkNu3MYeTEWazdvpdHRx/Hdzo3izqkYqlqSFJTVe+RA+qVk+R25uSyP6+Axy/rV6WTAKhEIKlMPXIkAb7cu5+GddLp3CKT1286ifS0qn+/XfUjFBFJEp9v3cNZ977H/W+sAEiKJAAqEYiIVIhPN+9m1IRZ7MvLr3LdQ0ujRCAiUk7Lv9jFyImzcHemjh3AUS0bRB1SmSgRiIiUw979eYyaOAuAaWMH0LlFZsQRlZ0SgYhIOdStVZM/nN2dri0zOaJ5/ajDOSRKBCIih2DBmi/ZtDOH07q35IxjWkUdTrkoEYiIlNHcz7Yx+rE5HNagNoOPapE0vYOKk9zRi4hUspmfbuWSR2fTIrM2T105IOmTACgRiIjE7b3lW7hs0mzaNKrDtKsG0LJhRtQhVQhVDYmIxOnd5Zvp0LQek6/sT7P6taMOp8IoEUjVlOhB4TSgm5RBTm4+Gelp3HzGUVx3chfqRzzHcEVT1ZBUTYkeFE4DukmcZny0gZPvfJvVW/diZtUuCYBKBFKVaVA4idi/FqzjxmcW0rtdIxrXS486nIRRiUBEpAjPzVvLDU8vIOvwxjxxeT8yM6pvIlCJQESkkFeXfMHPn1vIdzo1Y8IlWdSplRZ1SAmlRCAiUsjxnZryo5M68ZOTu5CRXr2TACgRSHklqnePevVIBKYvXM/JR7WgXu2a/HLIUVGHU2nURiDlk6jePerVI5XswbdWcP3U+Tz63qqoQ6l0KhFI+al3jyQxd+fe11cw7rVPOLtXa64e1CnqkCqdEoGIpCx352//XcYDb37KeX3bcvv5PUmrYVGHVelUNSQiKWvbnv08O3ctI/q1444UTQKgEoGIpCB3B6Bp/dq8eN0JNK9fmxopmgRAJQIRSTEFBc4tLyzij/9eirtzWIOMlE4CoEQgIikkv8D55T8+ZMqs1dRO1+XvAFUNiUhKyMsv4OfPfcjz89fxk5O7cMMpXTBL7ZLAAUoEIpISfhEmgZ+f3pVrBneOOpwqRYlARFLC6T1acnSrBoz57hFRh1LlKBGISLWVk5vPB6u3c3ynZpzevWXU4VRZai0RkWopJzefMU/M5dJHZ7Nm296ow6nSVCIQkWpn7/48rpg0l+xVW/nruT1p16Ru1CFVaUoEIlKt7N6Xx+WPzWHu59u4a1gvftinbdQhVXlKBCJSrbwwfx3zVm/nnuF9+EGv1lGHkxSUCESkWhnVvz3HHt6Yo1s1iDqUpKHGYhFJetv27OfiR2ax/ItdmJmSQBkpEYhIUtu8ax/Dx89k9qptbNyZE3U4SSmhVUNmNgS4B0gDJrr7Xwptbw88DjQK97nZ3WckMiYph6KmpdSUkhKhL3bmMHJCNuu/zOGx0cdxfOdmUYeUlBJWIjCzNOAB4AygGzDCzLoV2u3XwDPu3gcYDjyYqHikAhQ1LaWmlJSIfLEzhwsfnsnGHTk8fnk/JYFySGSJoB+wwt1XApjZNGAosCRmHwcOVOY1BNYnMB6pCJqWUqqIBhnpHHlYJncO682xhzeOOpyklshE0AZYE7O8FuhfaJ/fA/81s+uAesApCYxHRKqB1Vv30qheOg0y0hl/SVbU4VQLiWwsLmp8Vy+0PAKY5O5tgTOBJ83soJjMbKyZzTWzuZs3b05AqCKSDFZs2s0FD/+PG59eUPrOErdEJoK1QLuY5bYcXPVzBfAMgLvPBDKAgyr63H28u2e5e1bz5s0TFK6IVGXLNu5i+Phs8gvg56cfFXU41UoiE8EcoIuZdTSzWgSNwdML7bMaOBnAzI4mSAS65ReRb1myficjJmRTw2Da2AF0bZkZdUjVSsLaCNw9z8yuBV4h6Br6qLsvNrNbgbnuPh24CZhgZj8lqDYa7QdmlZZoqauoVBHuzs+eXUjtmjWYMmYAHZvVizqkaseS7bqblZXlc+fOjTqM6m/QoKIv/CNHwtixkYQkqeuzLXtIq2EaRbQczGyeuxfZuq6xhqR46ioqEZrz2TZeWbSRW846mg4qBSSUhpgQkSpn5qdbufTR2bzx8SZ2fJUbdTjVnhKBiFQp7y7fzGWTZtOmUR2mXTWARnVrRR1StaeqIRGpMt78eBNXTZ5Hp+b1mXxFP5rWrx11SClBiUBEqpRj2jTkkUuzVBKoREoEIhK59V9+RetGdRh8VAsGdW2OWVEDE0iiqI1ARCL1wvx1nHTHm7y1bBOAkkAElAhEJDLPzl3DT59ZQNbhTTiuQ5Oow0lZqhoSkUhMmbWa/3v+I07s0ozxF2dRp1Za1CGlLCUCEal0C9d8yf89/xGDuzbn7xcdS0a6kkCUlAhSXVFjCoHGFZKE6tWuEfcM782QHi2pXVNJIGpqI0h1RU0/CZqCUhJi4rsrWbx+BwBDe7dREqgiVCIQjSkkCefu3PP6cu5+bTmXDDycW4c2jDokiaFEICIJ5e7c8coyHnzrU84/ti2/+0H3qEOSQpQIRCRh3J0//WcpE99bxYh+7fnTOT2oUUO/E6hq1EYgIgmTV+Cs2LybSwcezv/7oZJAVaUSQXVXXK+gA9Q7SBKgoMDZsz+PzIx0xl+cRXqa6RfDVZhKBNVdcb2CDlDvIKlg+QXOL/7xISMmZJOTm0+tmjWUBKo4lQhSgXoFSSXJyy/gpmcX8q8F67nhlC7Urql7zWSgRCAiFSI3v4Abpi3gPx9t4Oend+WawZ2jDknipEQgIhXitn8v4T8fbeDXZx3NlSceEXU4UgZKBCJSIcZ89wi6t27IsOPaRR2KlJEq8ETkkH21P5+J766koMBp27iukkCSUokgaqV17ywvdQ+VBNmzL48rHp/DrFXb6N2uEVmaTyBpqUQQtdK6d5aXuodKAuzKyeXSR2cze9U27r6wt5JAklOJoCpQ905JIju+CpLAonU7uG9EX87q2SrqkKSclAhEpEyWf7GLTzft5sFRfTmte8uow5EKoEQgInHZn1dArZo1yOrQhPd++T0a1k2POiSpIGojEJFSbdqVww/ue49n5qwBUBKoZpQIojB+PAwaFDwS2VAsUgE27shh+MPZrNm+l3ZN6kYdjiSAEkEUYnsKqVePVGHrvvyKC8fPZNOufTx+eT8GdmoadUiSAGojiIp6CkkVtysnlwsfnsmOr3J54op+9G3fOOqQJEGUCESkSJkZ6Yw+vgP9OjahZ9tGUYcjCaREICLfsmLTbvbuz6Nn20YaPC5FqI1ARL62bOMuho+fyU+fXkB+gUcdjlQSlQjK61DGCtL4P1IFLV6/g4smzqJWzRqMvySLNM0vnDJUIiivQxkrSD2FpIr5cO2XjJwwizrpaTw9diCdmtePOiSpRCoRVAT1AJIkN+l/n5GZUZOpYwbotwIpSIlAJIW5O2bGn889hh17c2nRICPqkCQCqhoSSVH/W7GFc//+P7bt2U/tmmlKAiksrkRgZrXMTDNRi1QTb3+ymcsmzWHPvjz1DpLSE4GZnQV8BLwaLvc2s+cTHZiIJMbrS79gzONzOaJ5faaOGUDzzNpRhyQRi6eN4FagP/AmgLsvSMnSQXHdRNUVVJLIm8s28aPJ8zi6VQOeuLwfjerWijokqQLiqRrKdfcvC61LvbJkcd1E1RVUkki3Vg0465hWTL6yv5KAfC2eEsFSMxsG1DCzjsBPgOzEhlVFqZuoJKnslVvJOrwxhzXI4O7hfaIOR6qYeEoE1wLHAgXAP4EcgmQgIkngmTlrGDEhm0feWxV1KFJFxVMiON3dfwn88sAKMzuXICmISBU2Oftzfv3CIk7s0oxLBnaIOhypouIpEfy6iHW3VHQgIlKxHnt/Fb9+YRHfO6oFEy7Jok6ttKhDkiqq2BKBmZ0ODAHamNldMZsaEFQTpYYDvYXUO0iSyBc7c7jjlWWc3v0w7hvRl1o19dtRKV5JVUObgEUEbQKLY9bvAm5OZFBVSmwSUO8gSRKHNcjgH1cfT+cW9UlPUxKQkhWbCNx9PjDfzJ5y95xKjKnqUW8hSQLuzrjXltOsfi0uGdiBo1s1iDokSRLx3Cq0MbNpZvahmX1y4BHPm5vZEDNbZmYrzKzIUoSZDTOzJWa22MzKOLC/iECQBP768jLufX05i9ftxD31fuojhy6eXkOTgNuAvwFnAJcRRxuBmaUBDwCnAmuBOWY23d2XxOzTBfgV8B13325mLcp8BiIpzt3547+X8uj7qxjVvz1/HNoDM00qI/GLp0RQ191fAXD3T93918DgOF7XD1jh7ivdfT8wDRhaaJ8xwAPuvj18/03xhy4i7s7vpi/m0fdXMfr4Dtx2Tg9qaGYxKaN4SgT7LLi9+NTMfgSsA+K5c28DrIlZXkswZlGsIwHM7H0gDfi9u79c+I3MbCwwFqB9+/ZxHLoMSptqUr2FpAozM9o3qctV3z2Cm884SiUBOSTxJIKfAvWB64E/AQ2By+N4XVHfyMIVlzWBLsAgoC3wrpn1KDy2kbuPB8YDZGVlVWzlZ2ldQ9VbSKqg/AJn1ZbddG6RyZUnHhF1OJLkSk0E7j4rfLoLuBjAzNrG8d5rgXYxy22B9UXsk+3uucAqM1tGkBjmxPH+FUe9giSJ5OUXcOMzC3nj4028duNJtGyoCWWkfEpsIzCz48zsHDNrFi53N7MniG/QuTlAFzPraGa1gOHA9EL7vEDY3hAe40hgZRnPQSRl5OYXcN3U+UxfuJ5rBndWEpAKUWwiMLM/A08Bo4CXzewWgjkJFhLW7ZfE3fMIBqx7BVgKPOPui83sVjM7O9ztFWCrmS0J3/vn7r61PCckUl3ty8vn6skf8NKijfz6rKO5elCnqEOSaqKkqqGhQC93/8rMmhBU6/Ry92Xxvrm7zwBmFFr325jnDtwYPkSkBI//7zNeW/oFtw7trgHkpEKVlAhy3P0rAHffZmYflyUJiEjFuuw7HenasgEnHdk86lCkmikpERxhZgeGmjagQ8wy7n5uQiMTEfbsy+PWF5dw0+lH0iIzQ0lAEqKkRHBeoeX7ExmIiHzbrpxcLntsDvPXfMmp3Q7jlG5qGJbEKGnQudcrMxAR+caOvblc8thsFq/bwf0j+nBKt8OiDkmqsXh+UCYilWj7nv1c/Ogslm3cxd8vOpZTlQQkwZQIRKqY/HDk0PGXZDG4q8ZhlMSLOxGYWW1335fIYERS2dbd+2hQJ51m9Wsz/ZoTNHicVJpSRx81s35m9hGwPFzuZWb3JTwykRSycUcOFzw0k1/98yMAJQGpVPEMQ30v8H1gK4C7LyS+YahFJA5rt+9l2MMz2bRrH8OPa1f6C0QqWDxVQzXc/fNCw9vmJygekZSyeuteRkzIZmdOLk9e0Y8+7RtHHZKkoHgSwRoz6wd4OOvYdUBcU1WKSPHyC5wrn5jDnv15TB0zgB5tGkYdkqSoeBLB1QTVQ+2BL4DXwnUiUg5pNYw/n9uTurXSNNG8RCqeRJDn7sMTHolIivh4405mr9rGJQM7cOzhqgqS6MXTWDzHzGaY2aVmlpnwiESqsUXrdjBifDYPvvkpO3Nyow5HBIgjEbh7J+A24FjgIzN7wcxUQhApowVrvmTkhGzq1qrJ01cNoEFGetQhiQDxlQhw9/+5+/VAX2AnwYQ1IhKneZ9v46KJs2hYN52nrxrA4U3rRR2SyNfi+UFZfTMbZWYvArOBzcDxCY9MpBr55IvdtMiszTNXDaRt47pRhyPyLfE0Fi8CXgRud/d3ExyPSLWye18e9WvXZES/9vywTxsy0tOiDknkIPFUDR3h7tcpCYiUzVvLNnHiX99g3ufbAZQEpMoqtkRgZne6+03AP8zMC2/XDGUixXttyRf8+KkP6HJYfY5opvYAqdpKqhp6OvxXM5OJlMFLH23guqnz6d66AU9c3p+GddU7SKq2kmYomx0+Pdrdv5UMzOxaQDOYiRQy7/PtXDt1Pr3bNeKxy45TF1FJCvG0EVxexLorKjoQkeqgd7tG3HjqkTx+eT8lAUkaJbURXAgMBzqa2T9jNmUCXyY6MJFk8q8F6+jXsQmtGtbhmsGdow5HpExKaiOYTTAHQVvggZj1u4D5iQxKJJk8mf05v3lhERcNaM9t5xwTdTgiZVZSG8EqYBXBaKMiUoRH31vFrf9ewilHt+A33+8WdTgih6SkqqG33f0kM9sOxHYfNcDdvUnCoxOpwh5++1P+/NLHDOnekntH9KFWzbhGbBGpckqqGjowHWWzyghEJJnk5Obz/Px1/KBXa+4a1ov0NCUBSV4lVQ0VhE/bAevdfb+ZnQD0BCYTDD4nklLcnfwCJyM9jafHDqRe7TRqKglIkovnG/wCwTSVnYAngKOBKQmNSqQKcnf+8vLHXPXkPHLzC2hYN11JQKqFeL7FBe6eC5wL3O3u1wFtEhuWSNXi7tz67yU8/PZKWjXKIM0s6pBEKkxcU1Wa2QXAxcA54Tr9UkZSRkGB89vpi5icvZrLvtOB336/G6ZEINVIvL8sHkwwDPVKM+sITE1sWCJVx/+bsZTJ2au56qQjlASkWiq1RODui8zseqCzmR0FrHD3PyU+NJGqYWjvNjSqm841gzsrCUi1VGoiMLMTgSeBdQS/IWhpZhe7+/uJDk4kKrn5Bby+9AuG9GjFMW0bckzbhlGHJJIw8VQNjQPOdPfvuPvxwFnAPYkNSyQ6+/MKuG7KfH40+QPmr94edTgiCRdPY3Etd19yYMHdl5pZrQTGJBKZfXn5XPPUB7y2dBO//X43+rRvHHVIIgkXTyL4wMweJqgeAhiFBp2TaignN5+xT87jnU8288dzenDxgMOjDkmkUsSTCH4EXA/8gqCN4B3gvkQGJRKFmSu38v6KLfz1vGO48Lj2UYcjUmlKTARmdgiHsDYAAB0rSURBVAzQCXje3W+vnJBEKpe7Y2YM7tqC1288iQ6aY1hSTLGNxWb2fwTDS4wCXjWzomYqE0lqO3NyGTVxFu8t3wKgJCApqaReQ6OAnu5+AXAccHXlhCRSOXbszeXiibOYvWobu/flRh2OSGRKqhra5+57ANx9s5lpdC2pNrbt2c9FE2exYtNuHrroWE7pdljUIYlEpqREcETMXMUGdIqdu9jdz01oZCIJsuOrXEaMz+azrXsYf8mxDOraIuqQRCJVUiI4r9Dy/YkMRKSyZNauSf8jmvDbH3TjO50175JISRPTvF6ZgYgk2oYdX5GX77RrUpdbh/aIOhyRKkP1/pIS1mzby7CHZ3LVk/MoKPDSXyCSQuL5QZlIUvt86x5GTpjFrpxc7h/Rlxo1NIKoSKy4E4GZ1Xb3fYkMRqSifbp5NyMnZLM/r4ApYwbQo41GERUprNSqITPrZ2YfAcvD5V5mpiEmJCn8ecZS8gucqWOVBESKE08bwb3A94GtAO6+kGDGslKZ2RAzW2ZmK8zs5hL2O9/M3Myy4nlfkXjdeUFvnr5qIEe1bBB1KCJVVjyJoIa7f15oXX5pLzKzNOAB4AygGzDCzLoVsV8mwaB2s+KIRaRUi9bt4Lqp88nJzadh3XQ6Na8fdUgiVVo8iWCNmfUD3MzSzOwG4JM4XtePYFrLle6+H5gGDC1ivz8CtwM58QYtUpz5q7czYkI2H3y+nW179kcdjkhSiCcRXA3cCLQHvgAGEN+4Q22ANTHLa8N1XzOzPkA7d/93SW9kZmPNbK6Zzd28eXMch5ZUNPezbVz8yGwa163F01cNoHWjOlGHJJIU4pm8fhMw/BDeu6g+el934A7HLhoHjI4jhvHAeICsrCx1ApeDzFq5lcsmzaFlgwymjBlAy4YZUYckkjTimbx+AjEX8APcfWwpL10LtItZbgusj1nOBHoAb5kZQEtgupmd7e5zS4tLJFbDuun0aNOQ+0f2oUWmkoBIWcTzO4LXYp5nAD/k21U+xZkDdDGzjsA6glLFyAMb3X0H8PVAL2b2FvAzJQEpi0837+aIZvU4qmUDnh47gPCmQkTKoNQ2And/OubxOHAuQS+g0l6XB1wLvAIsBZ5x98VmdquZnV3ewEVeXfIFZ9z9LlNmrwZQEhA5RIcyxERHIK5Zvd19BjCj0LrfFrPvoEOIRVLUjI82cP3U+XRv05Dv92wddTgiSS2eNoLtfNNGUAPYBhT74zCRRPvXgnXc+MxCerdrxKTLjiMzIz3qkESSWmmT1xvQi6COH6DA3dVrRyKz7suv+NmzC8k6vDGPjj6OerU1bqJIeZX4v8jd3cyed/djKysgkZK0aVSHCZdk0b9jU+rUSos6HJFqIZ4flM02s74Jj0SkBE9mf87rS78AYFDXFkoCIhWo2ERgZgdKCycQJINlZvaBmc03sw8qJzwRmPjuSn7zwiL+OX9d6TuLSJmVVDU0G+gLnFNJsSTW+PEwZcrB6xcsgN69Kz8eicuDb63g9peXceYxLbn7Qv2dRBKhpERgAO7+aSXFklhTphR90e/dG0aOLPo1Eql7XlvOuNc+4exerblrWC9qpmlmVZFEKCkRNDezG4vb6O53JSCexOrdG956K+ooJA7uzpbd+zivb1tuP78naZpeUiRhSkoEaUB9ih48TiQh3J2te/bTrH5t/nB2dwDNMSySYCUlgg3ufmulRSIpz935w4tLeGXxRl687gSa1a8ddUgiKaGkSlfdhkmlKShwbnlhEZP+9xlnHtOKpvVqRR2SSMooqURwcqVFISktv8C5+R8f8uy8tVw9qBO/OL2rBpATqUTFJgJ331aZgUjqeujtT3l23lp+cnIXbjili5KASCXTQC0SuUsGHk7zzNoMy2pX+s4iUuHUMVsisT+vgHteW85X+/PJzEhXEhCJkBKBVLqc3HyunjyPca99wtufbIo6HJGUp6ohqVQ5ufmMeWIu7y7fwp9+2IMhPVpFHZJIylMikEqzd38eV0yaS/aqrdx+fk9VB4lUEUoEUmk27dzHyi27uWtYL37Yp23U4YhISIlAEu6r/flkpNegQ7N6vPmzQdStpa+dSFWixmJJqC/37mfYwzP523+XASgJiFRBSgSSMNv27GfEhFks27iLvu0bRx2OiBRDt2eSEJt37WPUxGw+37qXiZdm8d0jm0cdkogUQ4lAKlxefgEXPzKLNdu+4rHRx3F852ZRhyQiJVAikApXM60G15/chWb1a9OvY5OowxGRUigRSIVZs20vyzbu4pRuh3HmMfqhmEiyUCKQCvHZlj2MnJDN/vwC3u40mHq19dUSSRb63yrltmLTbkZNzGZ/XgGTr+yvJCCSZPQ/Vspl2cZdjJo4C3CmjR1I15aZUYckImWkRCDl8tKiDdQwmDJmIJ1b1I86HBE5BEoEckjyC5y0GsZPTu7CqP7BxDIikpz0y2Ipsw9Wb+fUcW/z6ebdmJmSgEiSUyKQMpnz2TYunjiL/AInIz0t6nBEpAKoakjiNvPTrVw+aQ6tGmUw5coBtGyYEXVIIlIBlAgkLh+s3s5lk2bTrnFdnhrTnxaZSgIi1YUSgcTlqJaZnNu3LTedeiRN66tNQKQ6URuBlOh/K7awe18edWvV5P/98BglAZFqSIlAivWfDzdwyaOz+dsry6IORUQSSIlAivTC/HVcN/UD+rRvxE2nHRl1OCKSQGojkIM8O3cNv/jHhwzo2JSJl2Zp7CCRak7/w+Vb9u7P465XP+GEzs0Yf3EWdWrptwIi1Z0SgXzN3albqybPXDWQ5pm19YMxkRShNgIBYMI7K/nDi0twd9o1qaskIJJClAiEB95cwZ9mLGXzrn3kF3jU4YhIJVPVUApzd+55fTl3v7acob1bc+cFvaiZpnsDkVSjRJDCxr22nHtfX875x7blr+f1JK2GRR2SiERAt38prGebhlw84HBuVxIQSWkqEaSYggLno3U76NWuEad0O4xTuh0WdUgiEjGVCFJIQYFzywuL+OGD77N4/Y6owxGRKkIlghSRX+D88h8f8ty8tVwzuBPdWjWIOiQRqSISWiIwsyFmtszMVpjZzUVsv9HMlpjZh2b2upkdnsh4UlVefgE3PrOA5+at5aenHMnPTuuKmdoERCSQsERgZmnAA8AZQDdghJl1K7TbfCDL3XsCzwG3JyqeVPby4o38a8F6fn56V35yShclARH5lkRWDfUDVrj7SgAzmwYMBZYc2MHd34zZPxu4KIHxpKyzjmlFi6sy6NexSdShiEgVlMiqoTbAmpjlteG64lwBvJTAeFJKTm4+Nz69gGUbd2FmSgIiUqxEJoKi6h+KHL/AzC4CsoA7itk+1szmmtnczZs3V2CI1dNX+/MZ88Rc/jl/HR+u/TLqcESkiktkIlgLtItZbgusL7yTmZ0C3AKc7e77inojdx/v7lnuntW8efOEBFtd7NmXx2WTZvPeii3cfn5PLshqV/qLRCSlJTIRzAG6mFlHM6sFDAemx+5gZn2AhwmSwKYExpISdu/L49JHZzN71TbGDevNMCUBEYlDwhKBu+cB1wKvAEuBZ9x9sZndamZnh7vdAdQHnjWzBWY2vZi3kzjUrGHUz6jJfSP6ck6fkppjRES+kdAflLn7DGBGoXW/jXl+SiKPnyq+3LsfM6NhnXQeG32cuoeKSJloiIkkt3X3PoaPz2bME3NxdyUBESkzJYIktmlXDsPHZ7Nqyx6u+15nJQEROSQaayhJbdyRw8gJ2WzYkcNjlx3H8Z2aRR2SiCQpJYIkddOzC9i0ax9PXNGP4zrox2IicuiUCJLUn3/Yk21799O7XaOoQxGRJKc2giSyassebn/5YwoKnPZN6yoJiEiFUCJIEis27eLCh2cybc4aNuzMiTocEalGVDWUBJZt3MWoidmAMW3sANo0qhN1SCJSjahEUMUtXr+D4eNnklbDePqqARx5WGbUIYlINaMSQRW3dfd+GtWtxWOjj6NDs3pRhyMi1ZASQRW1fc9+GterxXePbM5/f/pd0tNUeBORxNDVpQqatXIr3739TWZ8tAFASUBEEkpXmCrmfyu2MPqxObRoUJtjD28cdTgikgKUCKqQtz/ZzGWT5tC+SV2mjR3IYQ0yog5JRFKA2giqiM+37mHM43Pp3KI+k6/sT5N6taIOSURShBJBFXF403r8YWh3zujRkkZ1lQREpPKoaihiMz7awKJ1OwAY0a+9koCIVDolggg9P38t1075gHtfXx51KCKSwpQIIvLMnDXc+MxCBhzRlLuH9446HBFJYWojiMBTsz7nlucXcWKXZky4JIuM9LSoQxKRFKYSQSVzd95YuonvHdVCSUBEqgSVCCpRTm4+GelpPDCqLzXMqFVTeVhEoqcrUSW5/43lnPvg/9iZk0tGepqSgIhUGboaJZi7c9ern/C3/35C15aZ1FVVkIhUMaoaSiB3568vL+Ohtz/lgmPb8pfzepJWw6IOS0TkW5QIEuiht1fy0NufMqp/e/44tAc1lAREpApSIkigob1bk5dfwLXf64yZkoCIVE1qI6hgBQXOM3PXkF/gtG5Uh+tO7qIkICJVmkoEFSi/wPnFcx/yjw/W0iAjnSE9WkYdkohIqZQIKkhefgE3PrOQ6QvXc+OpRyoJiEjSUCKoAPvzCvjJtPm8tGgjvxxyFFcP6hR1SCIicVMiqADLN+3irWWb+fVZR3PliUdEHY6ISJkoEZRDQYFTo4bRvXVD3vzZIFo21NSSIpJ81GvoEH21P59LH5vNlFmrAZQERCRpKREcgj378hj92GzeX7GF2hozSESSnKqGymhnTi6XPTaHBWu+ZNyFvRnau03UIYmIlIsSQRnszyvg4kdms3jdDu4f0YczjmkVdUgiIuWmRFAGtWrW4KxjWnLt4M6c2u2wqMMREakQSgRx2LJ7Hxu+zOGYtg0Z+139RkBEqhe1dJZi084cho/P5orH55CTmx91OCIiFU4lghJs3JHDyAnZbNyZw6Ojj9P8wiJSLSkRFGPt9r2MnDCLbXv288Tl/cjq0CTqkEREEkKJoBgT3lnJ9r37efKKfvRp3zjqcEREEkaJoBi3nNWNiwd2oHOL+lGHIiKSUGosjrFi0y4umjiLrbv3UatmDSUBEUkJKhGEPt64k1ETZlGjhrF9by5N69eOOiQRkUqhRAAsWreDix+ZRe2aaUwZ058jmqskIFIRcnNzWbt2LTk5OVGHkjIyMjJo27Yt6enpcb8m5RPBonU7GDkhm8yMdKaM6c/hTetFHZJItbF27VoyMzPp0KGD5u6uBO7O1q1bWbt2LR07doz7dSnfRtAiszZ92jfm6asGKAmIVLCcnByaNm2qJFBJzIymTZuWuQSWsong4407ycsvoEWDDB6/vB9tG9eNOiSRaklJoHIdyuedkong/RVbOOeB97nz1U+iDkVEKsHzzz+PmfHxxx9/ve6tt97i+9///rf2Gz16NM899xwQtG/cfPPNdOnShR49etCvXz9eeumlcsfy5z//mc6dO9O1a1deeeWVIvd5/fXX6du3L7179+aEE05gxYoVAKxevZrBgwfTp08fevbsyYwZM8odDyQ4EZjZEDNbZmYrzOzmIrbXNrOnw+2zzKxDIuMBePuTzVw+aQ6HN6nH5d+Jvw5NRJLX1KlTOeGEE5g2bVrcr/nNb37Dhg0bWLRoEYsWLeLFF19k165d5YpjyZIlTJs2jcWLF/Pyyy/z4x//mPz8g8cwu/rqq3nqqadYsGABI0eO5LbbbgPgtttuY9iwYcyfP59p06bx4x//uFzxHJCwRGBmacADwBlAN2CEmXUrtNsVwHZ37wyMA/6aqHgAXmvUiTGPz6VT8/pMHTuA5pnqIipS3e3evZv333+fRx55JO5EsHfvXiZMmMB9991H7drBdeKwww5j2LBh5YrlX//6F8OHD6d27dp07NiRzp07M3v27IP2MzN27twJwI4dO2jdunWJ68srkb2G+gEr3H0lgJlNA4YCS2L2GQr8Pnz+HHC/mZm7e0UHsyutFj/rdAZHt8rkicv707Bu/F2rRKQC3HADLFhQse/ZuzfcfXeJu7zwwgsMGTKEI488kiZNmvDBBx/Qt2/fEl+zYsUK2rdvT4MGDUoN4ac//SlvvvnmQeuHDx/OzTd/uyJk3bp1DBgw4Ovltm3bsm7duoNeO3HiRM4880zq1KlDgwYNyM7OBuD3v/89p512Gvfddx979uzhtddeKzW+eCQyEbQB1sQsrwX6F7ePu+eZ2Q6gKbAldiczGwuMBWjfvv0hBZN5TDceL/iQjlcOpUGGkoBIqpg6dSo33HADEFycp06dSt++fYttVC1rY+u4cePi3reoe9yijjdu3DhmzJhB//79ueOOO7jxxhuZOHEiU6dOZfTo0dx0003MnDmTiy++mEWLFlGjRvkqdxKZCIr6NAt/CvHsg7uPB8YDZGVlHVpp4e676XVILxSRClHKnXsibN26lTfeeINFixZhZuTn52Nm3H777TRt2pTt27d/a/9t27bRrFkzOnfuzOrVq9m1axeZmZklHqMsJYK2bduyZs0398dr1649qHpn8+bNLFy4kP79g/vmCy+8kCFDhgDwyCOP8PLLLwMwcOBAcnJy2LJlCy1atIjzEylaIhuL1wLtYpbbAuuL28fMagINgW0JjElEUshzzz3HJZdcwueff85nn33GmjVr6NixI++99x5dunRh/fr1LF26FIDPP/+chQsX0rt3b+rWrcsVV1zB9ddfz/79+wHYsGEDkydPPugY48aNY8GCBQc9CicBgLPPPptp06axb98+Vq1axfLly+nXr9+39mncuDE7duzgk0+CXo2vvvoqRx99NBDUiLz++usALF26lJycHJo3b17uzymRJYI5QBcz6wisA4YDIwvtMx24FJgJnA+8kYj2ARFJTVOnTj3ognzeeecxZcoUTjzxRCZPnsxll11GTk4O6enpTJw4kYYNGwJBD51f//rXdOvWjYyMDOrVq8ett95arni6d+/OsGHD6NatGzVr1uSBBx4gLS2Y8OrMM89k4sSJtG7dmgkTJnDeeedRo0YNGjduzKOPPgrAnXfeyZgxYxg3bhxmxqRJkyrkdxqWyOuumZ0J3A2kAY+6+5/M7FZgrrtPN7MM4EmgD0FJYPiBxuXiZGVl+dy5cxMWs4hUnKVLl359NyuVp6jP3czmuXtWUfsndKwhd58BzCi07rcxz3OACxIZg4iIlCwlf1ksIiLfUCIQEUlxSgQiklDq/1G5DuXzViIQkYTJyMhg69atSgaV5MB8BBkZGWV6XcpPTCMiidO2bVvWrl3L5s2bow4lZRyYoawslAhEJGHS09PLNFOWRENVQyIiKU6JQEQkxSkRiIikuIQOMZEIZrYZ+PwQX96MQkNcpwCdc2rQOaeG8pzz4e5e5Ah1SZcIysPM5hY31kZ1pXNODTrn1JCoc1bVkIhIilMiEBFJcamWCMZHHUAEdM6pQeecGhJyzinVRiAiIgdLtRKBiIgUUi0TgZkNMbNlZrbCzA6aONTMapvZ0+H2WWbWofKjrFhxnPONZrbEzD40s9fN7PAo4qxIpZ1zzH7nm5mbWdL3MInnnM1sWPi3XmxmUyo7xooWx3e7vZm9aWbzw+/3mVHEWVHM7FEz22Rmi4rZbmZ2b/h5fGhmfct9UHevVg+CaTE/BY4AagELgW6F9vkx8FD4fDjwdNRxV8I5Dwbqhs+vToVzDvfLBN4BsoGsqOOuhL9zF2A+0DhcbhF13JVwzuOBq8Pn3YDPoo67nOf8XaAvsKiY7WcCLwEGDABmlfeY1bFE0A9Y4e4r3X0/MA0YWmifocDj4fPngJOtImaAjk6p5+zub7r73nAxGyjb8IRVTzx/Z4A/ArcDOZUZXILEc85jgAfcfTuAu2+q5BgrWjzn7ECD8HlDYH0lxlfh3P0dgjncizMUeMID2UAjM2tVnmNWx0TQBlgTs7w2XFfkPu6eB+wAmlZKdIkRzznHuoLgjiKZlXrOZtYHaOfu/67MwBIonr/zkcCRZva+mWWb2ZBKiy4x4jnn3wMXmdlagjnSr6uc0CJT1v/vpaqOw1AXdWdfuGtUPPskk7jPx8wuArKAkxIaUeKVeM5mVgMYB4yurIAqQTx/55oE1UODCEp975pZD3f/MsGxJUo85zwCmOTud5rZQODJ8JwLEh9eJCr8+lUdSwRrgXYxy205uKj49T5mVpOgOFlSUayqi+ecMbNTgFuAs919XyXFliilnXMm0AN4y8w+I6hLnZ7kDcbxfrf/5e657r4KWEaQGJJVPOd8BfAMgLvPBDIIxuSpruL6/14W1TERzAG6mFlHM6tF0Bg8vdA+04FLw+fnA2942AqTpEo957Ca5GGCJJDs9cZQyjm7+w53b+buHdy9A0G7yNnuPjeacCtEPN/tFwg6BmBmzQiqilZWapQVK55zXg2cDGBmRxMkguo8Jdp04JKw99AAYIe7byjPG1a7qiF3zzOza4FXCHocPOrui83sVmCuu08HHiEoPq4gKAkMjy7i8ovznO8A6gPPhu3iq9397MiCLqc4z7laifOcXwFOM7MlQD7wc3ffGl3U5RPnOd8ETDCznxJUkYxO5hs7M5tKULXXLGz3+B2QDuDuDxG0g5wJrAD2ApeV+5hJ/HmJiEgFqI5VQyIiUgZKBCIiKU6JQEQkxSkRiIikOCUCEZEUp0QgVY6Z5ZvZgphHhxL27VDcKI1lPOZb4QiXC8PhGboewnv8yMwuCZ+PNrPWMdsmmlm3Co5zjpn1juM1N5hZ3fIeW6ovJQKpir5y994xj88q6bij3L0XwYCEd5T1xe7+kLs/ES6OBlrHbLvS3ZdUSJTfxPkg8cV5A6BEIMVSIpCkEN75v2tmH4SP44vYp7uZzQ5LER+aWZdw/UUx6x82s7RSDvcO0Dl87cnhOPcfhePE1w7X/8W+md/hb+G635vZz8zsfILxnJ4Kj1knvJPPMrOrzez2mJhHm9l9hxjnTGIGGzOzv5vZXAvmIfhDuO56goT0ppm9Ga47zcxmhp/js2ZWv5TjSDWnRCBVUZ2YaqHnw3WbgFPdvS9wIXBvEa/7EXCPu/cmuBCvDYccuBD4Trg+HxhVyvF/AHxkZhnAJOBCdz+G4Jf4V5tZE+CHQHd37wncFvtid38OmEtw597b3b+K2fwccG7M8oXA04cY5xCCISUOuMXds4CewElm1tPd7yUYh2awuw8Oh534NXBK+FnOBW4s5ThSzVW7ISakWvgqvBjGSgfuD+vE8wnG0ClsJnCLmbUF/unuy83sZOBYYE44tEYdgqRSlKfM7CvgM4KhjLsCq9z9k3D748A1wP0E8xtMNLP/AHEPc+3um81sZThGzPLwGO+H71uWOOsRDLkQOzvVMDMbS/D/uhXBJC0fFnrtgHD9++FxahF8bpLClAgkWfwU+ALoRVCSPWiiGXefYmazgLOAV8zsSoIhex9391/FcYxRsYPSmVmRc1SE49/0IxjobDhwLfC9MpzL08Aw4GPgeXd3C67KccdJMFPXX4AHgHPNrCPwM+A4d99uZpMIBl8rzIBX3X1EGeKVak5VQ5IsGgIbwjHmLya4G/4WMzsCWBlWh0wnqCJ5HTjfzFqE+zSx+Odr/hjoYGadw+WLgbfDOvWG7j6DoCG2qJ47uwiGwi7KP4FzCMbRfzpcV6Y43T2XoIpnQFit1ADYA+wws8OAM4qJJRv4zoFzMrO6ZlZU6UpSiBKBJIsHgUvNLJugWmhPEftcCCwyswXAUQTT+S0huGD+18w+BF4lqDYplbvnEIzs+KyZfQQUAA8RXFT/Hb7f2wSllcImAQ8daCwu9L7bgSXA4e4+O1xX5jjDtoc7gZ+5+0KCuYoXA48SVDcdMB54yczedPfNBD2apobHySb4rCSFafRREZEUpxKBiEiKUyIQEUlxSgQiIilOiUBEJMUpEYiIpDglAhGRFKdEICKS4pQIRERS3P8He5brLjcftJsAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.metrics import roc_curve, auc\n",
    "y_prob = lr.predict_proba(x_test)[:,1]\n",
    "false_positive_rate, true_positive_rate, thresholds = roc_curve(y_test, y_prob)\n",
    "roc_auc = auc(false_positive_rate, true_positive_rate)\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "plt.figure(figsize=(6,6))\n",
    "plt.title('Receiver Operating Characteristic')\n",
    "plt.plot(false_positive_rate,true_positive_rate, color='red',label = 'AUC = %0.2f' % roc_auc)\n",
    "plt.legend(loc = 'lower right')\n",
    "plt.plot([0, 1], [0, 1],linestyle='--')\n",
    "plt.axis('tight')\n",
    "plt.ylabel('True Positive Rate')\n",
    "plt.xlabel('False Positive Rate')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[98 15]\n",
      " [15 32]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "         0.0       0.87      0.87      0.87       113\n",
      "         1.0       0.68      0.68      0.68        47\n",
      "\n",
      "    accuracy                           0.81       160\n",
      "   macro avg       0.77      0.77      0.77       160\n",
      "weighted avg       0.81      0.81      0.81       160\n",
      "\n"
     ]
    }
   ],
   "source": [
    "lr.fit(x_smote, y_smote)\n",
    "y_pred = lr.predict(x_test)\n",
    "\n",
    "print(confusion_matrix(y_test, y_pred))\n",
    "print(classification_report(y_test, y_pred))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Best result on original data = 0.82, AUC = 0.88"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 2. Decision Tree"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training score:  0.8257372654155496\n",
      "Testing score:  0.85\n"
     ]
    }
   ],
   "source": [
    "from sklearn.tree import DecisionTreeClassifier\n",
    "\n",
    "# Decision tree with depth = 2\n",
    "dt2 = tree.DecisionTreeClassifier(random_state=42, max_depth=2)\n",
    "dt2.fit(x_train, y_train)\n",
    "dt2_score_train = dt2.score(x_train, y_train)\n",
    "print(\"Training score: \",dt2_score_train)\n",
    "dt2_score_test = dt2.score(x_test, y_test)\n",
    "print(\"Testing score: \",dt2_score_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training score:  0.8284182305630027\n",
      "Testing score:  0.84375\n"
     ]
    }
   ],
   "source": [
    "# Decision tree with depth = 3\n",
    "dt3 = tree.DecisionTreeClassifier(random_state=42, max_depth=3)\n",
    "dt3.fit(x_train, y_train)\n",
    "dt3_score_train = dt3.score(x_train, y_train)\n",
    "print(\"Training score: \",dt3_score_train)\n",
    "dt3_score_test = dt3.score(x_test, y_test)\n",
    "print(\"Testing score: \",dt3_score_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training score:  0.8445040214477212\n",
      "Testing score:  0.84375\n"
     ]
    }
   ],
   "source": [
    "# Decision tree with depth = 4\n",
    "dt4 = tree.DecisionTreeClassifier(random_state=42, max_depth=4)\n",
    "dt4.fit(x_train, y_train)\n",
    "dt4_score_train = dt4.score(x_train, y_train)\n",
    "print(\"Training score: \",dt4_score_train)\n",
    "dt4_score_test = dt4.score(x_test, y_test)\n",
    "print(\"Testing score: \",dt4_score_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training score:  1.0\n",
      "Testing score:  0.7625\n"
     ]
    }
   ],
   "source": [
    "# Decision tree: To the full depth\n",
    "dt1 = tree.DecisionTreeClassifier()\n",
    "dt1.fit(x_train, y_train)\n",
    "dt1_score_train = dt1.score(x_train, y_train)\n",
    "print(\"Training score: \", dt1_score_train)\n",
    "dt1_score_test = dt1.score(x_test, y_test)\n",
    "print(\"Testing score: \", dt1_score_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "depth      Training score       Testing score       \n",
      "-----      --------------       -------------       \n",
      "2        0.8257372654155496                 0.85\n",
      "3        0.8284182305630027              0.84375\n",
      "4        0.8445040214477212              0.84375\n",
      "max                     1.0               0.7625\n"
     ]
    }
   ],
   "source": [
    "#Compare Training and Testing scores for various tree depths used\n",
    "print('{:10} {:20} {:20}'.format('depth', 'Training score','Testing score'))\n",
    "print('{:10} {:20} {:20}'.format('-----', '--------------','-------------'))\n",
    "print('{:1} {:>25} {:>20}'.format(2, dt2_score_train, dt2_score_test))\n",
    "print('{:1} {:>25} {:>20}'.format(3, dt3_score_train, dt3_score_test))\n",
    "print('{:1} {:>25} {:>20}'.format(4, dt4_score_train, dt4_score_test))\n",
    "print('{:1} {:>23} {:>20}'.format(\"max\", dt1_score_train, dt1_score_test))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Highest testing accuracy reached at depth 2. At this depth, training accuracy is good too so will go with this."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[109   4]\n",
      " [ 20  27]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "         0.0       0.84      0.96      0.90       113\n",
      "         1.0       0.87      0.57      0.69        47\n",
      "\n",
      "    accuracy                           0.85       160\n",
      "   macro avg       0.86      0.77      0.80       160\n",
      "weighted avg       0.85      0.85      0.84       160\n",
      "\n"
     ]
    }
   ],
   "source": [
    "y_pred = dt2.predict(x_test)\n",
    "\n",
    "print(confusion_matrix(y_test, y_pred))\n",
    "print(classification_report(y_test, y_pred))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 0, 'False Positive Rate')"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "y_prob = dt2.predict_proba(x_test)[:,1]\n",
    "false_positive_rate, true_positive_rate, thresholds = roc_curve(y_test, y_prob)\n",
    "roc_auc = auc(false_positive_rate, true_positive_rate)\n",
    "\n",
    "plt.figure(figsize=(6,6))\n",
    "plt.title('Receiver Operating Characteristic')\n",
    "plt.plot(false_positive_rate,true_positive_rate, color='red',label = 'AUC = %0.2f' % roc_auc)\n",
    "plt.legend(loc = 'lower right')\n",
    "plt.plot([0, 1], [0, 1],linestyle='--')\n",
    "plt.axis('tight')\n",
    "plt.ylabel('True Positive Rate')\n",
    "plt.xlabel('False Positive Rate')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training score:  0.755859375\n",
      "Testing score:  0.85\n",
      "[[109   4]\n",
      " [ 20  27]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "         0.0       0.84      0.96      0.90       113\n",
      "         1.0       0.87      0.57      0.69        47\n",
      "\n",
      "    accuracy                           0.85       160\n",
      "   macro avg       0.86      0.77      0.80       160\n",
      "weighted avg       0.85      0.85      0.84       160\n",
      "\n"
     ]
    }
   ],
   "source": [
    "dt2.fit(x_smote, y_smote)\n",
    "dt2_score_train = dt2.score(x_smote, y_smote)\n",
    "print(\"Training score: \",dt2_score_train)\n",
    "dt2_score_test = dt2.score(x_test, y_test)\n",
    "print(\"Testing score: \",dt2_score_test)\n",
    "\n",
    "y_pred = dt2.predict(x_test)\n",
    "print(confusion_matrix(y_test, y_pred))\n",
    "print(classification_report(y_test, y_pred))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Best result on original data = 0.85, AUC = 0.82"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 3. Random Forest Classifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.model_selection import RandomizedSearchCV\n",
    "rf = RandomForestClassifier(n_estimators=100)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[108   5]\n",
      " [ 14  33]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "         0.0       0.89      0.96      0.92       113\n",
      "         1.0       0.87      0.70      0.78        47\n",
      "\n",
      "    accuracy                           0.88       160\n",
      "   macro avg       0.88      0.83      0.85       160\n",
      "weighted avg       0.88      0.88      0.88       160\n",
      "\n"
     ]
    }
   ],
   "source": [
    "rf.fit(x_train, y_train)\n",
    "y_pred = rf.predict(x_test)\n",
    "\n",
    "print(confusion_matrix(y_test, y_pred))\n",
    "print(classification_report(y_test, y_pred))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 0, 'False Positive Rate')"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "y_prob = rf.predict_proba(x_test)[:,1]\n",
    "false_positive_rate, true_positive_rate, thresholds = roc_curve(y_test, y_prob)\n",
    "roc_auc = auc(false_positive_rate, true_positive_rate)\n",
    "\n",
    "plt.figure(figsize=(6,6))\n",
    "plt.title('Receiver Operating Characteristic')\n",
    "plt.plot(false_positive_rate,true_positive_rate, color='red',label = 'AUC = %0.2f' % roc_auc)\n",
    "plt.legend(loc = 'lower right')\n",
    "plt.plot([0, 1], [0, 1],linestyle='--')\n",
    "plt.axis('tight')\n",
    "plt.ylabel('True Positive Rate')\n",
    "plt.xlabel('False Positive Rate')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[102  11]\n",
      " [ 14  33]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "         0.0       0.88      0.90      0.89       113\n",
      "         1.0       0.75      0.70      0.73        47\n",
      "\n",
      "    accuracy                           0.84       160\n",
      "   macro avg       0.81      0.80      0.81       160\n",
      "weighted avg       0.84      0.84      0.84       160\n",
      "\n"
     ]
    }
   ],
   "source": [
    "rf.fit(x_smote, y_smote)\n",
    "y_pred = rf.predict(x_test)\n",
    "print(confusion_matrix(y_test, y_pred))\n",
    "print(classification_report(y_test, y_pred))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Best result on original data = 0.86, AUC = 0.86"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 4. Support Vector Machine (SVM)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn import svm\n",
    "from sklearn import metrics"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.70625\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "         0.0       0.71      1.00      0.83       113\n",
      "         1.0       0.00      0.00      0.00        47\n",
      "\n",
      "    accuracy                           0.71       160\n",
      "   macro avg       0.35      0.50      0.41       160\n",
      "weighted avg       0.50      0.71      0.58       160\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\Lt col Haider\\anaconda3\\lib\\site-packages\\sklearn\\metrics\\_classification.py:1272: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n"
     ]
    }
   ],
   "source": [
    "#Radial Basis Function Kernel with low gamma 0.01\n",
    "\n",
    "clf = svm.SVC(kernel='rbf', random_state=0, gamma=.01, C=1) \n",
    "\n",
    "#Train the model using the training sets\n",
    "clf.fit(x_train, y_train)\n",
    "\n",
    "#Predict the response for test dataset\n",
    "y_pred = clf.predict(x_test)\n",
    "\n",
    "print(\"Accuracy:\",metrics.accuracy_score(y_test, y_pred))\n",
    "confusion_matrix(y_test, y_pred)\n",
    "print(classification_report(y_test, y_pred))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.8625\n",
      "[[111   2]\n",
      " [ 20  27]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "         0.0       0.85      0.98      0.91       113\n",
      "         1.0       0.93      0.57      0.71        47\n",
      "\n",
      "    accuracy                           0.86       160\n",
      "   macro avg       0.89      0.78      0.81       160\n",
      "weighted avg       0.87      0.86      0.85       160\n",
      "\n"
     ]
    }
   ],
   "source": [
    "#Radial Basis Fnction Kernel with penalty parameter c = 1000.\n",
    "#classifier starts to become very intolerant to misclassified data points and thus the decision boundary becomes less biased and has more variance (i.e. more dependent on the individual data points)\n",
    "clf = svm.SVC(kernel='rbf', random_state=0, gamma=.01, C=1000) \n",
    "\n",
    "#Train the model using the training sets\n",
    "clf.fit(x_train, y_train)\n",
    "\n",
    "#Predict the response for test dataset\n",
    "y_pred = clf.predict(x_test)\n",
    "\n",
    "#takes too much training time.\n",
    "from sklearn import metrics\n",
    "\n",
    "# model accuracy:\n",
    "print(\"Accuracy:\",metrics.accuracy_score(y_test, y_pred))\n",
    "print(confusion_matrix(y_test, y_pred))\n",
    "print(classification_report(y_test, y_pred))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 0, 'False Positive Rate')"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "false_positive_rate, true_positive_rate, thresholds = roc_curve(y_test, y_pred)\n",
    "roc_auc = auc(false_positive_rate, true_positive_rate)\n",
    "\n",
    "plt.figure(figsize=(6,6))\n",
    "plt.title('Receiver Operating Characteristic')\n",
    "plt.plot(false_positive_rate,true_positive_rate, color='red',label = 'AUC = %0.2f' % roc_auc)\n",
    "plt.legend(loc = 'lower right')\n",
    "plt.plot([0, 1], [0, 1],linestyle='--')\n",
    "plt.axis('tight')\n",
    "plt.ylabel('True Positive Rate')\n",
    "plt.xlabel('False Positive Rate')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.83125\n",
      "[[103  10]\n",
      " [ 17  30]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "         0.0       0.86      0.91      0.88       113\n",
      "         1.0       0.75      0.64      0.69        47\n",
      "\n",
      "    accuracy                           0.83       160\n",
      "   macro avg       0.80      0.77      0.79       160\n",
      "weighted avg       0.83      0.83      0.83       160\n",
      "\n"
     ]
    }
   ],
   "source": [
    "#Radial Basis Fnction Kernel with penalty parameter c = 1000.\n",
    "#classifier starts to become very intolerant to misclassified data points and thus the decision boundary becomes less biased and has more variance (i.e. more dependent on the individual data points)\n",
    "clf = svm.SVC(kernel='rbf', random_state=0, gamma=.01, C=1000) \n",
    "\n",
    "#Train the model using the training sets\n",
    "clf.fit(x_smote, y_smote)\n",
    "\n",
    "#Predict the response for test dataset\n",
    "y_pred = clf.predict(x_test)\n",
    "\n",
    "#takes too much training time.\n",
    "from sklearn import metrics\n",
    "\n",
    "# model accuracy:\n",
    "print(\"Accuracy:\",metrics.accuracy_score(y_test, y_pred))\n",
    "print(confusion_matrix(y_test, y_pred))\n",
    "print(classification_report(y_test, y_pred))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Best result on original data = 0.86, AUC = 0.78"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 5. Naive Bayes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.8625\n",
      "[[111   2]\n",
      " [ 20  27]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "         0.0       0.85      0.98      0.91       113\n",
      "         1.0       0.93      0.57      0.71        47\n",
      "\n",
      "    accuracy                           0.86       160\n",
      "   macro avg       0.89      0.78      0.81       160\n",
      "weighted avg       0.87      0.86      0.85       160\n",
      "\n"
     ]
    }
   ],
   "source": [
    "from sklearn.naive_bayes import GaussianNB\n",
    "naive = GaussianNB()\n",
    "naive.fit(x_train, y_train)\n",
    "    \n",
    "# Make predictions and evalute\n",
    "y_pred = naive.predict(x_test)\n",
    "\n",
    "print(\"Accuracy:\",metrics.accuracy_score(y_test, y_pred))\n",
    "print(confusion_matrix(y_test, y_pred))    \n",
    "print(classification_report(y_test, y_pred))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 0, 'False Positive Rate')"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "y_prob = naive.predict_proba(x_test)[:,1]\n",
    "false_positive_rate, true_positive_rate, thresholds = roc_curve(y_test, y_prob)\n",
    "roc_auc = auc(false_positive_rate, true_positive_rate)\n",
    "\n",
    "plt.figure(figsize=(6,6))\n",
    "plt.title('Receiver Operating Characteristic')\n",
    "plt.plot(false_positive_rate,true_positive_rate, color='red',label = 'AUC = %0.2f' % roc_auc)\n",
    "plt.legend(loc = 'lower right')\n",
    "plt.plot([0, 1], [0, 1],linestyle='--')\n",
    "plt.axis('tight')\n",
    "plt.ylabel('True Positive Rate')\n",
    "plt.xlabel('False Positive Rate')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.8375\n",
      "[[106   7]\n",
      " [ 19  28]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "         0.0       0.85      0.94      0.89       113\n",
      "         1.0       0.80      0.60      0.68        47\n",
      "\n",
      "    accuracy                           0.84       160\n",
      "   macro avg       0.82      0.77      0.79       160\n",
      "weighted avg       0.83      0.84      0.83       160\n",
      "\n"
     ]
    }
   ],
   "source": [
    "naive.fit(x_smote, y_smote)\n",
    "    \n",
    "# Make predictions and evalute\n",
    "y_pred = naive.predict(x_test)\n",
    "print(\"Accuracy:\",metrics.accuracy_score(y_test, y_pred))\n",
    "print(confusion_matrix(y_test, y_pred))    \n",
    "print(classification_report(y_test, y_pred))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Best result on original data = 0.86, AUC = 0.88"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 6. XGBoost Classifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [],
   "source": [
    "from xgboost import XGBClassifier\n",
    "\n",
    "model = XGBClassifier(silent=False, \n",
    "                      scale_pos_weight=1,\n",
    "                      learning_rate=0.01,  \n",
    "                      colsample_bytree = 0.8,\n",
    "                      subsample = 0.8,\n",
    "                      objective='binary:logistic', \n",
    "                      n_estimators=100, \n",
    "                      reg_alpha = 0.3,\n",
    "                      max_depth=4, \n",
    "                      gamma=5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.85625\n",
      "[[109   4]\n",
      " [ 19  28]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "         0.0       0.85      0.96      0.90       113\n",
      "         1.0       0.88      0.60      0.71        47\n",
      "\n",
      "    accuracy                           0.86       160\n",
      "   macro avg       0.86      0.78      0.81       160\n",
      "weighted avg       0.86      0.86      0.85       160\n",
      "\n"
     ]
    }
   ],
   "source": [
    "model.fit(x_train, y_train)\n",
    "    \n",
    "# Make predictions and evalute\n",
    "y_pred = model.predict(x_test)\n",
    "\n",
    "print(\"Accuracy:\",metrics.accuracy_score(y_test, y_pred))\n",
    "print(confusion_matrix(y_test, y_pred)) \n",
    "print(classification_report(y_test, y_pred))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 0, 'False Positive Rate')"
      ]
     },
     "execution_count": 57,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "y_prob = model.predict_proba(x_test)[:,1]\n",
    "false_positive_rate, true_positive_rate, thresholds = roc_curve(y_test, y_prob)\n",
    "roc_auc = auc(false_positive_rate, true_positive_rate)\n",
    "\n",
    "plt.figure(figsize=(6,6))\n",
    "plt.title('Receiver Operating Characteristic')\n",
    "plt.plot(false_positive_rate,true_positive_rate, color='red',label = 'AUC = %0.2f' % roc_auc)\n",
    "plt.legend(loc = 'lower right')\n",
    "plt.plot([0, 1], [0, 1],linestyle='--')\n",
    "plt.axis('tight')\n",
    "plt.ylabel('True Positive Rate')\n",
    "plt.xlabel('False Positive Rate')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.86875\n",
      "[[109   4]\n",
      " [ 17  30]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "         0.0       0.87      0.96      0.91       113\n",
      "         1.0       0.88      0.64      0.74        47\n",
      "\n",
      "    accuracy                           0.87       160\n",
      "   macro avg       0.87      0.80      0.83       160\n",
      "weighted avg       0.87      0.87      0.86       160\n",
      "\n"
     ]
    }
   ],
   "source": [
    "model.fit(x_smote, y_smote)\n",
    "    \n",
    "# Make predictions and evalute\n",
    "y_pred = model.predict(x_test)\n",
    "print(\"Accuracy:\",metrics.accuracy_score(y_test, y_pred))\n",
    "print(confusion_matrix(y_test, y_pred)) \n",
    "print(classification_report(y_test, y_pred))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 0, 'False Positive Rate')"
      ]
     },
     "execution_count": 59,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "y_prob = model.predict_proba(x_test)[:,1]\n",
    "false_positive_rate, true_positive_rate, thresholds = roc_curve(y_test, y_prob)\n",
    "roc_auc = auc(false_positive_rate, true_positive_rate)\n",
    "\n",
    "plt.figure(figsize=(6,6))\n",
    "plt.title('Receiver Operating Characteristic')\n",
    "plt.plot(false_positive_rate,true_positive_rate, color='red',label = 'AUC = %0.2f' % roc_auc)\n",
    "plt.legend(loc = 'lower right')\n",
    "plt.plot([0, 1], [0, 1],linestyle='--')\n",
    "plt.axis('tight')\n",
    "plt.ylabel('True Positive Rate')\n",
    "plt.xlabel('False Positive Rate')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Best result on original data = 0.87, AUC = 0.89"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "Best Accuracy = 0.87, XGB | Best AUC = 0.89 XGB | Best Model = XGBoost Classifier"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}