Credit Fraud - ML.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## TRANSACTION DATA (CREDIT FRAUD) - MACHINE LEARNING"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    " Sumber dataset : https://drive.google.com/file/d/1sr0k8_k7huFuHiR_C5P_r60VaBDzwoTb/view"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Dalam tahap ini, akan di lakukan proses **Classification Machine Learning** terhadap data *transaction.csv*. Pada **preprocessing** dalam tahap **EDA** sebelumnya sudah di lakukan pembersihan dari **Null Values**, sehingga dalam proses **Machine Learning** ini hanya terfokus untuk membangun model prediksi. Model yang di gunakan adalah ***LogisticRegression*** dan ***KNNeighbors***. Pembuatan model di lakukan dengan melalui metode *pipelining*, sehingga hasilnya juga akan cenderung lebih baik. Pembuatan model juga melewati beberapa tahap **feature engineering**, dimana data akan di lakukan uji **normalitas** dan **outliers** checking, **feature selection**, serta **cross-validation**. Dan karena data sangat *imbalanced*, maka di lakukan ***oversampling*** dengan metode ***SMOTE***.\n",
    "\n",
    "\n",
    "### FEATURE DESCRIPTION :\n",
    "\n",
    "- accountNumber             = The account number of The customer\n",
    "- customerId    =  The ID of The customer\n",
    "- creditLimit               =  The amount of money that can be charged to The debit card\n",
    "- availableMoney        = The amount of money in The debit card before adjusting for pending charges      \n",
    "- transactionDateTime       = The transaction timestamp when it happened\n",
    "- transactionAmount        = The amount of transaction \n",
    "- merchantName          = The merchant name of The particular transaction\n",
    "- acqCountry             = The country where The merchant is located\n",
    "- merchantCountryCode    = The country code for The specific merchant\n",
    "- posEntryMode        = a code that tells The processor how The transaction was captured\n",
    "- posConditionCode   = a code identifying transaction conditions at The point-of-sale or point-of-service\n",
    "- merchantCategoryCode    = The merchant category/types\n",
    "- currentExpDate       =  The expiry date of The credit card\n",
    "- accountOpenDate      = The date when The customer open The credit card\n",
    "- dateOfLastAddressChange    = The last date when The customer change The credit card address\n",
    "- cardCVV = The actual card verification value\n",
    "- enteredCVV       = The entered card verification value\n",
    "- cardLast4Digits     = The last 4 digits of The debit card\n",
    "- transactionType        = The types of transactions\n",
    "- isFraud      = The status of The fraud transaction\n",
    "- echoBuffer    = number of delayed response transactions\n",
    "- currentBalance =  The current balance of The debit card\n",
    "- merchantCity = The location for The specific merchant (City)\n",
    "- merchantState   = The location for The specific merchant (State)\n",
    "- merchantZip = The location for The specific merchant (Zip Code)           \n",
    "- cardPresent =     The physical presence of The debit card in The transaction\n",
    "- posOnPremises = The location of The point of sales          \n",
    "- recurringAuthInd  = wheTher The auThentication recurred or not\n",
    "- expirationDateKeyInMatch = The match between The expiration date in The system and what was inputted"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## IMPORT LIBRARIES"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "\n",
    "from sklearn.metrics import confusion_matrix, classification_report, roc_auc_score, roc_curve, auc, f1_score, precision_score, recall_score, matthews_corrcoef\n",
    "from sklearn.datasets import make_classification\n",
    "\n",
    "from sklearn.model_selection import train_test_split, RandomizedSearchCV, StratifiedKFold , cross_val_score\n",
    "from sklearn.preprocessing import RobustScaler, LabelEncoder\n",
    "from sklearn.pipeline import Pipeline\n",
    "\n",
    "from sklearn.feature_selection import RFE\n",
    "from sklearn.feature_selection import SelectKBest, chi2\n",
    "from sklearn.compose import ColumnTransformer\n",
    "\n",
    "from sklearn.ensemble import ExtraTreesClassifier\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "from sklearn.model_selection import RandomizedSearchCV\n",
    "\n",
    "import scipy.stats as ss\n",
    "import statsmodels.api as sm\n",
    "from sklearn.neighbors import KernelDensity\n",
    "from scipy.stats import mannwhitneyu, chi2_contingency\n",
    "from scipy.stats import anderson\n",
    "from imblearn.over_sampling import SMOTE\n",
    "\n",
    "from yellowbrick.model_selection import learning_curve, ValidationCurve\n",
    "\n",
    "import warnings\n",
    "warnings.filterwarnings(\"ignore\")\n",
    "warnings.filterwarnings(action='ignore', category=FutureWarning)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## OVERVIEW"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_csv('transaction_clean.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "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>acqCountry</th>\n",
       "      <th>merchantCountryCode</th>\n",
       "      <th>posEntryMode</th>\n",
       "      <th>posConditionCode</th>\n",
       "      <th>merchantCategoryCode</th>\n",
       "      <th>transactionType</th>\n",
       "      <th>creditLimit</th>\n",
       "      <th>availableMoney</th>\n",
       "      <th>transactionAmount</th>\n",
       "      <th>currentBalance</th>\n",
       "      <th>transactionDateTime</th>\n",
       "      <th>currentExpDate</th>\n",
       "      <th>accountOpenDate</th>\n",
       "      <th>dateOfLastAddressChange</th>\n",
       "      <th>isFraud</th>\n",
       "      <th>cardPresent</th>\n",
       "      <th>expirationDateKeyInMatch</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>US</td>\n",
       "      <td>US</td>\n",
       "      <td>5.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>rideshare</td>\n",
       "      <td>PURCHASE</td>\n",
       "      <td>5000</td>\n",
       "      <td>5000.00</td>\n",
       "      <td>111.33</td>\n",
       "      <td>0.00</td>\n",
       "      <td>2016-01-08 19:04:50</td>\n",
       "      <td>2020-01-04 00:00:00</td>\n",
       "      <td>2014-03-08 00:00:00</td>\n",
       "      <td>2014-03-08 00:00:00</td>\n",
       "      <td>Yes</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>US</td>\n",
       "      <td>US</td>\n",
       "      <td>9.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>rideshare</td>\n",
       "      <td>PURCHASE</td>\n",
       "      <td>5000</td>\n",
       "      <td>4888.67</td>\n",
       "      <td>24.75</td>\n",
       "      <td>111.33</td>\n",
       "      <td>2016-01-09 22:32:39</td>\n",
       "      <td>2023-01-06 00:00:00</td>\n",
       "      <td>2014-03-08 00:00:00</td>\n",
       "      <td>2014-03-08 00:00:00</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>US</td>\n",
       "      <td>US</td>\n",
       "      <td>5.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>rideshare</td>\n",
       "      <td>PURCHASE</td>\n",
       "      <td>5000</td>\n",
       "      <td>4863.92</td>\n",
       "      <td>187.40</td>\n",
       "      <td>136.08</td>\n",
       "      <td>2016-01-11 13:36:55</td>\n",
       "      <td>2027-01-12 00:00:00</td>\n",
       "      <td>2014-03-08 00:00:00</td>\n",
       "      <td>2014-03-08 00:00:00</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>US</td>\n",
       "      <td>US</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>rideshare</td>\n",
       "      <td>PURCHASE</td>\n",
       "      <td>5000</td>\n",
       "      <td>4676.52</td>\n",
       "      <td>227.34</td>\n",
       "      <td>323.48</td>\n",
       "      <td>2016-01-11 22:47:46</td>\n",
       "      <td>2029-01-09 00:00:00</td>\n",
       "      <td>2014-03-08 00:00:00</td>\n",
       "      <td>2014-03-08 00:00:00</td>\n",
       "      <td>Yes</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>US</td>\n",
       "      <td>US</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>rideshare</td>\n",
       "      <td>ADDRESS_VERIFICATION</td>\n",
       "      <td>5000</td>\n",
       "      <td>4449.18</td>\n",
       "      <td>0.00</td>\n",
       "      <td>550.82</td>\n",
       "      <td>2016-01-16 01:41:11</td>\n",
       "      <td>2024-01-10 00:00:00</td>\n",
       "      <td>2014-03-08 00:00:00</td>\n",
       "      <td>2014-03-08 00:00:00</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  acqCountry merchantCountryCode  posEntryMode  posConditionCode  \\\n",
       "0         US                  US           5.0               1.0   \n",
       "1         US                  US           9.0               1.0   \n",
       "2         US                  US           5.0               1.0   \n",
       "3         US                  US           2.0               1.0   \n",
       "4         US                  US           2.0               1.0   \n",
       "\n",
       "  merchantCategoryCode       transactionType  creditLimit  availableMoney  \\\n",
       "0            rideshare              PURCHASE         5000         5000.00   \n",
       "1            rideshare              PURCHASE         5000         4888.67   \n",
       "2            rideshare              PURCHASE         5000         4863.92   \n",
       "3            rideshare              PURCHASE         5000         4676.52   \n",
       "4            rideshare  ADDRESS_VERIFICATION         5000         4449.18   \n",
       "\n",
       "   transactionAmount  currentBalance  transactionDateTime  \\\n",
       "0             111.33            0.00  2016-01-08 19:04:50   \n",
       "1              24.75          111.33  2016-01-09 22:32:39   \n",
       "2             187.40          136.08  2016-01-11 13:36:55   \n",
       "3             227.34          323.48  2016-01-11 22:47:46   \n",
       "4               0.00          550.82  2016-01-16 01:41:11   \n",
       "\n",
       "        currentExpDate      accountOpenDate dateOfLastAddressChange isFraud  \\\n",
       "0  2020-01-04 00:00:00  2014-03-08 00:00:00     2014-03-08 00:00:00     Yes   \n",
       "1  2023-01-06 00:00:00  2014-03-08 00:00:00     2014-03-08 00:00:00      No   \n",
       "2  2027-01-12 00:00:00  2014-03-08 00:00:00     2014-03-08 00:00:00      No   \n",
       "3  2029-01-09 00:00:00  2014-03-08 00:00:00     2014-03-08 00:00:00     Yes   \n",
       "4  2024-01-10 00:00:00  2014-03-08 00:00:00     2014-03-08 00:00:00      No   \n",
       "\n",
       "  cardPresent expirationDateKeyInMatch  \n",
       "0          No                       No  \n",
       "1          No                       No  \n",
       "2          No                       No  \n",
       "3          No                       No  \n",
       "4          No                       No  "
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "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>dataFeatures</th>\n",
       "      <th>dataType</th>\n",
       "      <th>null</th>\n",
       "      <th>nullPct</th>\n",
       "      <th>Nunique</th>\n",
       "      <th>uniqueSample</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>acqCountry</td>\n",
       "      <td>object</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4</td>\n",
       "      <td>[MEX, US]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>merchantCountryCode</td>\n",
       "      <td>object</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4</td>\n",
       "      <td>[US, PR]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>posEntryMode</td>\n",
       "      <td>float64</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5</td>\n",
       "      <td>[9.0, 2.0]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>posConditionCode</td>\n",
       "      <td>float64</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3</td>\n",
       "      <td>[8.0, 99.0]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>merchantCategoryCode</td>\n",
       "      <td>object</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>19</td>\n",
       "      <td>[hotels, mobileapps]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>transactionType</td>\n",
       "      <td>object</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3</td>\n",
       "      <td>[ADDRESS_VERIFICATION, REVERSAL]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>creditLimit</td>\n",
       "      <td>int64</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>10</td>\n",
       "      <td>[7500, 20000]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>availableMoney</td>\n",
       "      <td>float64</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>446204</td>\n",
       "      <td>[39929.58, 4456.23]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>transactionAmount</td>\n",
       "      <td>float64</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>62571</td>\n",
       "      <td>[30.28, 326.3]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>currentBalance</td>\n",
       "      <td>float64</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>403487</td>\n",
       "      <td>[570.62, 228.48]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>transactionDateTime</td>\n",
       "      <td>object</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>627556</td>\n",
       "      <td>[2016-08-17 17:14:32, 2016-09-03 02:11:34]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>currentExpDate</td>\n",
       "      <td>object</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>165</td>\n",
       "      <td>[2025-01-12 00:00:00, 2027-01-02 00:00:00]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>accountOpenDate</td>\n",
       "      <td>object</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1826</td>\n",
       "      <td>[2009-03-03 00:00:00, 2006-10-26 00:00:00]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>dateOfLastAddressChange</td>\n",
       "      <td>object</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2186</td>\n",
       "      <td>[2016-02-21 00:00:00, 2015-12-07 00:00:00]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>isFraud</td>\n",
       "      <td>object</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2</td>\n",
       "      <td>[Yes, No]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>cardPresent</td>\n",
       "      <td>object</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2</td>\n",
       "      <td>[Yes, No]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>expirationDateKeyInMatch</td>\n",
       "      <td>object</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2</td>\n",
       "      <td>[Yes, No]</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                dataFeatures dataType  null  nullPct  Nunique  \\\n",
       "0                 acqCountry   object     0      0.0        4   \n",
       "1        merchantCountryCode   object     0      0.0        4   \n",
       "2               posEntryMode  float64     0      0.0        5   \n",
       "3           posConditionCode  float64     0      0.0        3   \n",
       "4       merchantCategoryCode   object     0      0.0       19   \n",
       "5            transactionType   object     0      0.0        3   \n",
       "6                creditLimit    int64     0      0.0       10   \n",
       "7             availableMoney  float64     0      0.0   446204   \n",
       "8          transactionAmount  float64     0      0.0    62571   \n",
       "9             currentBalance  float64     0      0.0   403487   \n",
       "10       transactionDateTime   object     0      0.0   627556   \n",
       "11            currentExpDate   object     0      0.0      165   \n",
       "12           accountOpenDate   object     0      0.0     1826   \n",
       "13   dateOfLastAddressChange   object     0      0.0     2186   \n",
       "14                   isFraud   object     0      0.0        2   \n",
       "15               cardPresent   object     0      0.0        2   \n",
       "16  expirationDateKeyInMatch   object     0      0.0        2   \n",
       "\n",
       "                                  uniqueSample  \n",
       "0                                    [MEX, US]  \n",
       "1                                     [US, PR]  \n",
       "2                                   [9.0, 2.0]  \n",
       "3                                  [8.0, 99.0]  \n",
       "4                         [hotels, mobileapps]  \n",
       "5             [ADDRESS_VERIFICATION, REVERSAL]  \n",
       "6                                [7500, 20000]  \n",
       "7                          [39929.58, 4456.23]  \n",
       "8                               [30.28, 326.3]  \n",
       "9                             [570.62, 228.48]  \n",
       "10  [2016-08-17 17:14:32, 2016-09-03 02:11:34]  \n",
       "11  [2025-01-12 00:00:00, 2027-01-02 00:00:00]  \n",
       "12  [2009-03-03 00:00:00, 2006-10-26 00:00:00]  \n",
       "13  [2016-02-21 00:00:00, 2015-12-07 00:00:00]  \n",
       "14                                   [Yes, No]  \n",
       "15                                   [Yes, No]  \n",
       "16                                   [Yes, No]  "
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.DataFrame({'dataFeatures' : df.columns, 'dataType' : df.dtypes.values, \n",
    "              'null' : [df[i].isna().sum() for i in df.columns],\n",
    "              'nullPct' : [((df[i].isna().sum()/len(df[i]))*100).round(1) for i in df.columns],\n",
    "              'Nunique' : [df[i].nunique() for i in df.columns],\n",
    "              'uniqueSample' : [list(pd.Series(df[i].unique()).sample(2)) for i in df.columns]}).reset_index(drop = True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> Data sudah bersih dari *null values*, namun sepertinya tipe datanya kembali seperti semula. Selanjutnya kita rapihkan dahulu setelah itu, kita masuk ke proses *Feature Engineering*."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Rubah data ke datetime\n",
    "\n",
    "error = []\n",
    "for i in df[['transactionDateTime', 'currentExpDate', 'accountOpenDate', 'dateOfLastAddressChange']]:\n",
    "    try : \n",
    "        df[i] = pd.to_datetime(df[i])\n",
    "    except :\n",
    "        error.append(i)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Rubah data ke object\n",
    "\n",
    "for i in df[['posEntryMode', 'posConditionCode']]:\n",
    "    df[i] = df[i].astype('object')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 633846 entries, 0 to 633845\n",
      "Data columns (total 17 columns):\n",
      " #   Column                    Non-Null Count   Dtype         \n",
      "---  ------                    --------------   -----         \n",
      " 0   acqCountry                633846 non-null  object        \n",
      " 1   merchantCountryCode       633846 non-null  object        \n",
      " 2   posEntryMode              633846 non-null  object        \n",
      " 3   posConditionCode          633846 non-null  object        \n",
      " 4   merchantCategoryCode      633846 non-null  object        \n",
      " 5   transactionType           633846 non-null  object        \n",
      " 6   creditLimit               633846 non-null  int64         \n",
      " 7   availableMoney            633846 non-null  float64       \n",
      " 8   transactionAmount         633846 non-null  float64       \n",
      " 9   currentBalance            633846 non-null  float64       \n",
      " 10  transactionDateTime       633846 non-null  datetime64[ns]\n",
      " 11  currentExpDate            633846 non-null  datetime64[ns]\n",
      " 12  accountOpenDate           633846 non-null  datetime64[ns]\n",
      " 13  dateOfLastAddressChange   633846 non-null  datetime64[ns]\n",
      " 14  isFraud                   633846 non-null  object        \n",
      " 15  cardPresent               633846 non-null  object        \n",
      " 16  expirationDateKeyInMatch  633846 non-null  object        \n",
      "dtypes: datetime64[ns](4), float64(3), int64(1), object(9)\n",
      "memory usage: 82.2+ MB\n"
     ]
    }
   ],
   "source": [
    "df.info()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> Data sudah rapih seperti sebelumnya. Kita bisa lanjutkan ke *Feature Engineering*."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## FEATURE ENGINEERING"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> Sebelum mulai, saya akan membagi menjadi 2 kategori agar mudah ketika feature engineering."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_cat = df.select_dtypes('object')\n",
    "df_con = df.select_dtypes(['int64', 'float64'])\n",
    "df_date = df.select_dtypes('datetime64[ns]')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>acqCountry</th>\n",
       "      <td>US</td>\n",
       "      <td>US</td>\n",
       "      <td>US</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>merchantCountryCode</th>\n",
       "      <td>US</td>\n",
       "      <td>US</td>\n",
       "      <td>US</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>posEntryMode</th>\n",
       "      <td>5</td>\n",
       "      <td>9</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>posConditionCode</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>merchantCategoryCode</th>\n",
       "      <td>rideshare</td>\n",
       "      <td>rideshare</td>\n",
       "      <td>rideshare</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>transactionType</th>\n",
       "      <td>PURCHASE</td>\n",
       "      <td>PURCHASE</td>\n",
       "      <td>PURCHASE</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>isFraud</th>\n",
       "      <td>Yes</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>cardPresent</th>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>expirationDateKeyInMatch</th>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                  0          1          2\n",
       "acqCountry                       US         US         US\n",
       "merchantCountryCode              US         US         US\n",
       "posEntryMode                      5          9          5\n",
       "posConditionCode                  1          1          1\n",
       "merchantCategoryCode      rideshare  rideshare  rideshare\n",
       "transactionType            PURCHASE   PURCHASE   PURCHASE\n",
       "isFraud                         Yes         No         No\n",
       "cardPresent                      No         No         No\n",
       "expirationDateKeyInMatch         No         No         No"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_cat.head(3).T"
   ]
  },
  {
   "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>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>creditLimit</th>\n",
       "      <td>5000.00</td>\n",
       "      <td>5000.00</td>\n",
       "      <td>5000.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>availableMoney</th>\n",
       "      <td>5000.00</td>\n",
       "      <td>4888.67</td>\n",
       "      <td>4863.92</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>transactionAmount</th>\n",
       "      <td>111.33</td>\n",
       "      <td>24.75</td>\n",
       "      <td>187.40</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>currentBalance</th>\n",
       "      <td>0.00</td>\n",
       "      <td>111.33</td>\n",
       "      <td>136.08</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                         0        1        2\n",
       "creditLimit        5000.00  5000.00  5000.00\n",
       "availableMoney     5000.00  4888.67  4863.92\n",
       "transactionAmount   111.33    24.75   187.40\n",
       "currentBalance        0.00   111.33   136.08"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_con.head(3).T"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### NORMALITY AND OUTLIER CHECK"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> Pertama kita lakukan uji normalitas dan cek outlier pada setiap data nominal."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "def normal(df):\n",
    "    \n",
    "    stat = anderson(df)[0]\n",
    "    cval = anderson(df)[1][2]\n",
    "    if stat < cval :\n",
    "        print('NORMALLY Distributed for ' + i + ' Variable')\n",
    "    else :\n",
    "        print('NOT NORMALLY Distributed for ' + i + ' Variable')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "NOT NORMALLY Distributed for creditLimit Variable\n",
      "NOT NORMALLY Distributed for availableMoney Variable\n",
      "NOT NORMALLY Distributed for transactionAmount Variable\n",
      "NOT NORMALLY Distributed for currentBalance Variable\n"
     ]
    }
   ],
   "source": [
    "for i in df_con.columns :\n",
    "    normal(df_con[i])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> Pengujian menggunakan **anderson method** menyatakan bahwa distribusi tidak normal. Di bawah ini kita lakukan visualisasi menggunakan **qqplot**."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Figure size 864x360 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x396 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 864x360 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x396 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 864x360 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x396 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 864x360 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x396 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "for i in df_con.columns :\n",
    "    \n",
    "    plt.figure(figsize = (12,5))\n",
    "    \n",
    "    sm.qqplot(df[i], line = '45')\n",
    "    plt.xlabel(i, fontsize = 12)\n",
    "    plt.title('QQ PLOT of ' + i, fontsize = 22)\n",
    "    \n",
    "    plt.tight_layout()\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> Sekarang kita cek outliersnya."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x720 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize = (20,10))\n",
    "sns.boxplot(data=df_con)\n",
    "plt.tight_layout()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> Terlihat sekali data memiliki banyak outliers. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "def outliers(df):\n",
    "    Q2 = np.quantile(df, 0.50)\n",
    "    Q1 = np.quantile(df, 0.25)\n",
    "    Q3 = np.quantile(df, 0.75)\n",
    "    \n",
    "    IQR   = (Q3 - Q1) * 1.5\n",
    "    lower =  Q1 - IQR\n",
    "    upper =  Q3 + IQR \n",
    "\n",
    "    outlier = ''\n",
    "    for i in range(len(df)):\n",
    "        if df[i] > upper :\n",
    "            outlier = outlier + str(df[i]) + ', '\n",
    "        if df[i] < lower :\n",
    "            outlier = outlier + str(df[i]) + ', '\n",
    "    return len(outlier)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "There are 263123 Outliers on creditLimit Features\n",
      "There are 437167 Outliers on availableMoney Features\n",
      "There are 262271 Outliers on transactionAmount Features\n",
      "There are 471246 Outliers on currentBalance Features\n"
     ]
    }
   ],
   "source": [
    "# Mencari outlier pada tiap features\n",
    "\n",
    "for i in df_con.columns:\n",
    "    outlier = outliers(df_con[i])\n",
    "    print(f'There are {outlier} Outliers on {i} Features')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> Terdapat ratusan ribu outlier pada tiap features. Dalam kasus semacam ini hal ini bisa saja terjadi, sehingga sepertinya tidak memungkinkan ketika di drop seluruh outliers ini. Akan lebih memungkinkan ketika kita menggunakan `RobustScaler()` nanti."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### LABELLING"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> Karena seluruh data dalam categorical features kali ini tergolong data nominal, maka kita gunakan `pd.get_dummies()` untuk proses labelling nya."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "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>dataFeatures</th>\n",
       "      <th>dataType</th>\n",
       "      <th>Nunique</th>\n",
       "      <th>uniqueSample</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>acqCountry</td>\n",
       "      <td>object</td>\n",
       "      <td>4</td>\n",
       "      <td>[MEX, PR]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>merchantCountryCode</td>\n",
       "      <td>object</td>\n",
       "      <td>4</td>\n",
       "      <td>[MEX, PR]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>posEntryMode</td>\n",
       "      <td>object</td>\n",
       "      <td>5</td>\n",
       "      <td>[5.0, 90.0]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>posConditionCode</td>\n",
       "      <td>object</td>\n",
       "      <td>3</td>\n",
       "      <td>[1.0, 8.0]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>merchantCategoryCode</td>\n",
       "      <td>object</td>\n",
       "      <td>19</td>\n",
       "      <td>[mobileapps, food]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>transactionType</td>\n",
       "      <td>object</td>\n",
       "      <td>3</td>\n",
       "      <td>[REVERSAL, PURCHASE]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>isFraud</td>\n",
       "      <td>object</td>\n",
       "      <td>2</td>\n",
       "      <td>[No, Yes]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>cardPresent</td>\n",
       "      <td>object</td>\n",
       "      <td>2</td>\n",
       "      <td>[No, Yes]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>expirationDateKeyInMatch</td>\n",
       "      <td>object</td>\n",
       "      <td>2</td>\n",
       "      <td>[Yes, No]</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               dataFeatures dataType  Nunique          uniqueSample\n",
       "0                acqCountry   object        4             [MEX, PR]\n",
       "1       merchantCountryCode   object        4             [MEX, PR]\n",
       "2              posEntryMode   object        5           [5.0, 90.0]\n",
       "3          posConditionCode   object        3            [1.0, 8.0]\n",
       "4      merchantCategoryCode   object       19    [mobileapps, food]\n",
       "5           transactionType   object        3  [REVERSAL, PURCHASE]\n",
       "6                   isFraud   object        2             [No, Yes]\n",
       "7               cardPresent   object        2             [No, Yes]\n",
       "8  expirationDateKeyInMatch   object        2             [Yes, No]"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.DataFrame({'dataFeatures' : df_cat.columns, 'dataType' : df_cat.dtypes.values, \n",
    "              'Nunique' : [df_cat[i].nunique() for i in df_cat.columns],\n",
    "              'uniqueSample' : [list(pd.Series(df_cat[i].unique()).sample(2)) for i in df_cat.columns]}).reset_index(drop = True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "encoder = LabelEncoder()\n",
    "df_cat['isFraud'] = encoder.fit_transform(df_cat[['isFraud']])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> **isFraud** features menggunakan `LabelEncoder()` karena merupakan target."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "for i in df_cat.drop(columns = ['isFraud']):\n",
    "    df_cat = pd.get_dummies(df_cat, drop_first=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "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>isFraud</th>\n",
       "      <th>acqCountry_MEX</th>\n",
       "      <th>acqCountry_PR</th>\n",
       "      <th>acqCountry_US</th>\n",
       "      <th>merchantCountryCode_MEX</th>\n",
       "      <th>merchantCountryCode_PR</th>\n",
       "      <th>merchantCountryCode_US</th>\n",
       "      <th>posEntryMode_5.0</th>\n",
       "      <th>posEntryMode_9.0</th>\n",
       "      <th>posEntryMode_80.0</th>\n",
       "      <th>...</th>\n",
       "      <th>merchantCategoryCode_online_gifts</th>\n",
       "      <th>merchantCategoryCode_online_retail</th>\n",
       "      <th>merchantCategoryCode_online_subscriptions</th>\n",
       "      <th>merchantCategoryCode_personal care</th>\n",
       "      <th>merchantCategoryCode_rideshare</th>\n",
       "      <th>merchantCategoryCode_subscriptions</th>\n",
       "      <th>transactionType_PURCHASE</th>\n",
       "      <th>transactionType_REVERSAL</th>\n",
       "      <th>cardPresent_Yes</th>\n",
       "      <th>expirationDateKeyInMatch_Yes</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>3 rows × 35 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   isFraud  acqCountry_MEX  acqCountry_PR  acqCountry_US  \\\n",
       "0        1               0              0              1   \n",
       "1        0               0              0              1   \n",
       "2        0               0              0              1   \n",
       "\n",
       "   merchantCountryCode_MEX  merchantCountryCode_PR  merchantCountryCode_US  \\\n",
       "0                        0                       0                       1   \n",
       "1                        0                       0                       1   \n",
       "2                        0                       0                       1   \n",
       "\n",
       "   posEntryMode_5.0  posEntryMode_9.0  posEntryMode_80.0  ...  \\\n",
       "0                 1                 0                  0  ...   \n",
       "1                 0                 1                  0  ...   \n",
       "2                 1                 0                  0  ...   \n",
       "\n",
       "   merchantCategoryCode_online_gifts  merchantCategoryCode_online_retail  \\\n",
       "0                                  0                                   0   \n",
       "1                                  0                                   0   \n",
       "2                                  0                                   0   \n",
       "\n",
       "   merchantCategoryCode_online_subscriptions  \\\n",
       "0                                          0   \n",
       "1                                          0   \n",
       "2                                          0   \n",
       "\n",
       "   merchantCategoryCode_personal care  merchantCategoryCode_rideshare  \\\n",
       "0                                   0                               1   \n",
       "1                                   0                               1   \n",
       "2                                   0                               1   \n",
       "\n",
       "   merchantCategoryCode_subscriptions  transactionType_PURCHASE  \\\n",
       "0                                   0                         1   \n",
       "1                                   0                         1   \n",
       "2                                   0                         1   \n",
       "\n",
       "   transactionType_REVERSAL  cardPresent_Yes  expirationDateKeyInMatch_Yes  \n",
       "0                         0                0                             0  \n",
       "1                         0                0                             0  \n",
       "2                         0                0                             0  \n",
       "\n",
       "[3 rows x 35 columns]"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_cat.head(3)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> Selanjutnya kita akan persiapkan data untuk di scaling. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "# for i in df_con.columns :\n",
    "#     scaler = RobustScaler()\n",
    "#     df_con[i] = scaler.fit_transform(df[[i]])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Unscaled data\n",
    "\n",
    "df_new = pd.concat([df_con, df_cat], axis = 1)"
   ]
  },
  {
   "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></th>\n",
       "      <th>creditLimit</th>\n",
       "      <th>availableMoney</th>\n",
       "      <th>transactionAmount</th>\n",
       "      <th>currentBalance</th>\n",
       "      <th>isFraud</th>\n",
       "      <th>acqCountry_MEX</th>\n",
       "      <th>acqCountry_PR</th>\n",
       "      <th>acqCountry_US</th>\n",
       "      <th>merchantCountryCode_MEX</th>\n",
       "      <th>merchantCountryCode_PR</th>\n",
       "      <th>...</th>\n",
       "      <th>merchantCategoryCode_online_gifts</th>\n",
       "      <th>merchantCategoryCode_online_retail</th>\n",
       "      <th>merchantCategoryCode_online_subscriptions</th>\n",
       "      <th>merchantCategoryCode_personal care</th>\n",
       "      <th>merchantCategoryCode_rideshare</th>\n",
       "      <th>merchantCategoryCode_subscriptions</th>\n",
       "      <th>transactionType_PURCHASE</th>\n",
       "      <th>transactionType_REVERSAL</th>\n",
       "      <th>cardPresent_Yes</th>\n",
       "      <th>expirationDateKeyInMatch_Yes</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>5000</td>\n",
       "      <td>5000.00</td>\n",
       "      <td>111.33</td>\n",
       "      <td>0.00</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>5000</td>\n",
       "      <td>4888.67</td>\n",
       "      <td>24.75</td>\n",
       "      <td>111.33</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>5000</td>\n",
       "      <td>4863.92</td>\n",
       "      <td>187.40</td>\n",
       "      <td>136.08</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>3 rows × 39 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   creditLimit  availableMoney  transactionAmount  currentBalance  isFraud  \\\n",
       "0         5000         5000.00             111.33            0.00        1   \n",
       "1         5000         4888.67              24.75          111.33        0   \n",
       "2         5000         4863.92             187.40          136.08        0   \n",
       "\n",
       "   acqCountry_MEX  acqCountry_PR  acqCountry_US  merchantCountryCode_MEX  \\\n",
       "0               0              0              1                        0   \n",
       "1               0              0              1                        0   \n",
       "2               0              0              1                        0   \n",
       "\n",
       "   merchantCountryCode_PR  ...  merchantCategoryCode_online_gifts  \\\n",
       "0                       0  ...                                  0   \n",
       "1                       0  ...                                  0   \n",
       "2                       0  ...                                  0   \n",
       "\n",
       "   merchantCategoryCode_online_retail  \\\n",
       "0                                   0   \n",
       "1                                   0   \n",
       "2                                   0   \n",
       "\n",
       "   merchantCategoryCode_online_subscriptions  \\\n",
       "0                                          0   \n",
       "1                                          0   \n",
       "2                                          0   \n",
       "\n",
       "   merchantCategoryCode_personal care  merchantCategoryCode_rideshare  \\\n",
       "0                                   0                               1   \n",
       "1                                   0                               1   \n",
       "2                                   0                               1   \n",
       "\n",
       "   merchantCategoryCode_subscriptions  transactionType_PURCHASE  \\\n",
       "0                                   0                         1   \n",
       "1                                   0                         1   \n",
       "2                                   0                         1   \n",
       "\n",
       "   transactionType_REVERSAL  cardPresent_Yes  expirationDateKeyInMatch_Yes  \n",
       "0                         0                0                             0  \n",
       "1                         0                0                             0  \n",
       "2                         0                0                             0  \n",
       "\n",
       "[3 rows x 39 columns]"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_new.head(3)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### FEATURE IMPORTANCES"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> Feature importances atau feature / variable selection, merupakan pemilihan attribute atau subset variable yang relevan untuk digunakan pada model construction. Beberapa metode dapat dilakukan untuk melakukan feature importances ini. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### RECURSIVE FEATURE ELIMINATION"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> Recursive feature elimination (RFE) adalah suatu metode feature selection dengan cara metode model fitting lalu mengeliminasi feature yang paling lemah. Feature tersebut di urutkan berdasarkan coefficient modelnya. RFE mencoba mengeliminasi dependencies dan collinearity yang ada pada model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = df_new.drop(['isFraud'], axis = 1)\n",
    "y = df_new['isFraud']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(X, y, train_size = .8, random_state = 42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "creditLimit                 True\n",
       "availableMoney              True\n",
       "transactionAmount           True\n",
       "currentBalance              True\n",
       "acqCountry_US               True\n",
       "merchantCountryCode_US      True\n",
       "posEntryMode_5.0            True\n",
       "posEntryMode_9.0            True\n",
       "transactionType_PURCHASE    True\n",
       "cardPresent_Yes             True\n",
       "dtype: bool"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Select 10 best feature using RFE\n",
    "\n",
    "model = LogisticRegression(solver = 'lbfgs', penalty = 'none', max_iter = 10)\n",
    "rfe = RFE(model, n_features_to_select= 10)\n",
    "rfe = rfe.fit(X_train, y_train)\n",
    "\n",
    "# Summarize the selection of the attributes\n",
    "\n",
    "rfe_feat = pd.Series(rfe.support_, index=X.columns)\n",
    "rfe_feat[rfe_feat == True]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "rfe_col = rfe_feat[rfe_feat == True].index\n",
    "rfe_tgt = pd.DataFrame(df_new['isFraud'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_rfe = pd.concat([df_new[rfe_col], rfe_tgt], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "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>creditLimit</th>\n",
       "      <th>availableMoney</th>\n",
       "      <th>transactionAmount</th>\n",
       "      <th>currentBalance</th>\n",
       "      <th>acqCountry_US</th>\n",
       "      <th>merchantCountryCode_US</th>\n",
       "      <th>posEntryMode_5.0</th>\n",
       "      <th>posEntryMode_9.0</th>\n",
       "      <th>transactionType_PURCHASE</th>\n",
       "      <th>cardPresent_Yes</th>\n",
       "      <th>isFraud</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>5000</td>\n",
       "      <td>5000.00</td>\n",
       "      <td>111.33</td>\n",
       "      <td>0.00</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>5000</td>\n",
       "      <td>4888.67</td>\n",
       "      <td>24.75</td>\n",
       "      <td>111.33</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>5000</td>\n",
       "      <td>4863.92</td>\n",
       "      <td>187.40</td>\n",
       "      <td>136.08</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   creditLimit  availableMoney  transactionAmount  currentBalance  \\\n",
       "0         5000         5000.00             111.33            0.00   \n",
       "1         5000         4888.67              24.75          111.33   \n",
       "2         5000         4863.92             187.40          136.08   \n",
       "\n",
       "   acqCountry_US  merchantCountryCode_US  posEntryMode_5.0  posEntryMode_9.0  \\\n",
       "0              1                       1                 1                 0   \n",
       "1              1                       1                 0                 1   \n",
       "2              1                       1                 1                 0   \n",
       "\n",
       "   transactionType_PURCHASE  cardPresent_Yes  isFraud  \n",
       "0                         1                0        1  \n",
       "1                         1                0        0  \n",
       "2                         1                0        0  "
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_rfe.head(3)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> Kita sudah mendapatkan 10 feature yang dipilih dari metode RFE."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### TREE BASED CLASSIFIER"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> Extremely Randomized Trees Classifier(Extra Trees Classifier) adalah salah satu jenis feature importances yang menggunakan teknik Ensemble Learning. Konsepnya mirip dengan Random Forest yang meng-agregasi hasil dari beberapa de-correlated Decision Tree untuk mengeluarkan outputnya."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[3.11144075e-02 3.00087932e-01 3.38051174e-01 2.98987730e-01\n",
      " 2.46435089e-04 1.91641527e-04 3.92262523e-04 2.76443747e-04\n",
      " 2.00666289e-04 3.66606578e-04 3.27315200e-03 3.09203116e-03\n",
      " 1.29641696e-03 1.99964003e-03 3.18314053e-03 1.45165033e-03\n",
      " 6.04047037e-04 6.99920943e-05 9.89438417e-04 1.24161936e-03\n",
      " 8.07676910e-04 1.49125287e-04 5.53000856e-04 3.69897279e-04\n",
      " 7.34412267e-05 1.07787573e-03 7.06955370e-04 3.77653517e-04\n",
      " 1.13154923e-03 1.19393418e-03 3.03856267e-04 3.98308275e-04\n",
      " 1.00212432e-03 5.20280521e-04 9.52413529e-04 8.06085172e-04\n",
      " 2.08160042e-03 3.77794201e-04]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x396 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# feature importances using Tree Based Classifier\n",
    "\n",
    "model = ExtraTreesClassifier()\n",
    "model.fit(X,y)\n",
    "print(model.feature_importances_)\n",
    "\n",
    "# plot graph of feature importances for better visualization\n",
    "\n",
    "feat_importances = pd.Series(model.feature_importances_, index=X.columns)\n",
    "feat_importances.nlargest(10).plot(kind = 'barh')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "feat_col = feat_importances.nlargest(10).index.tolist()\n",
    "feat_tgt = pd.DataFrame(df_new['isFraud'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_etc = pd.concat([df_new[feat_col], feat_tgt], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>transactionAmount</th>\n",
       "      <th>availableMoney</th>\n",
       "      <th>currentBalance</th>\n",
       "      <th>creditLimit</th>\n",
       "      <th>posEntryMode_5.0</th>\n",
       "      <th>posConditionCode_8.0</th>\n",
       "      <th>posEntryMode_9.0</th>\n",
       "      <th>cardPresent_Yes</th>\n",
       "      <th>posEntryMode_90.0</th>\n",
       "      <th>posConditionCode_99.0</th>\n",
       "      <th>isFraud</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>111.33</td>\n",
       "      <td>5000.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>5000</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>24.75</td>\n",
       "      <td>4888.67</td>\n",
       "      <td>111.33</td>\n",
       "      <td>5000</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>187.40</td>\n",
       "      <td>4863.92</td>\n",
       "      <td>136.08</td>\n",
       "      <td>5000</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   transactionAmount  availableMoney  currentBalance  creditLimit  \\\n",
       "0             111.33         5000.00            0.00         5000   \n",
       "1              24.75         4888.67          111.33         5000   \n",
       "2             187.40         4863.92          136.08         5000   \n",
       "\n",
       "   posEntryMode_5.0  posConditionCode_8.0  posEntryMode_9.0  cardPresent_Yes  \\\n",
       "0                 1                     0                 0                0   \n",
       "1                 0                     0                 1                0   \n",
       "2                 1                     0                 0                0   \n",
       "\n",
       "   posEntryMode_90.0  posConditionCode_99.0  isFraud  \n",
       "0                  0                      0        1  \n",
       "1                  0                      0        0  \n",
       "2                  0                      0        0  "
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_etc.head(3)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> Kita juga mendapatkan 10 feature yang di pilih menggunakan Extra Tree Classifier. Sebetulnya keduanya sama saja, maka kita pilih salah satu saja."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## MODEL PREPARATION"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> Di bawah ini merupakan persiapan untuk proses modelling."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_ready = df_etc.copy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_ready.to_csv('transaction_HP.csv', index = False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = df_ready.drop(['isFraud'], axis=1)\n",
    "y = df_ready['isFraud']\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> Modelling dilakukan menggunakan pipeline dan scaling menggunakan Robust Scaler sekaligus bagi model yang menggunakan default parameter maupun model yang menggunakan hyperparameter."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Default model build + pipelining\n",
    "\n",
    "# Logistic Regression \n",
    "\n",
    "LRG = Pipeline([('scale', RobustScaler()),\n",
    "                ('clf', LogisticRegression())])\n",
    "\n",
    "# K-Nearest Neighbors \n",
    "\n",
    "KNN = Pipeline([('scale', RobustScaler()),\n",
    "                ('clf', KNeighborsClassifier())])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> Tuning hyperparameter di lakukan di notebook yang terpisah. Hasil terbaik di terapkan di sini."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Hyperparameter tuned model build + pipelining\n",
    "\n",
    "# Logistic Regression \n",
    "\n",
    "hyper_LRG = Pipeline([('scale', RobustScaler()),\n",
    "                      ('clf', LogisticRegression(solver = 'lbfgs', penalty = 'l2', max_iter = 200))])\n",
    "\n",
    "# K-Nearest Neighbors \n",
    "\n",
    "hyper_KNN = Pipeline([('scale', RobustScaler()),\n",
    "                      ('clf', KNeighborsClassifier(p = 2, n_neighbors = 25, leaf_size = 20))])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### MODEL FITTING"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> Selanjutnya kita akan lakukan fitting model kepada data train, baik pada model default maupun model hyperparameter tuned."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Default model fitting\n",
    "\n",
    "# Logistic Regression \n",
    "\n",
    "LRG = LRG.fit(X_train, y_train)\n",
    "\n",
    "# K-Nearest Neighbors \n",
    "\n",
    "KNN = KNN.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Hyperparameter tuned model fitting\n",
    "\n",
    "# Logistic Regression \n",
    "\n",
    "hyper_LRG = hyper_LRG.fit(X_train, y_train)\n",
    "\n",
    "# K-Nearest Neighbors \n",
    "\n",
    "hyper_KNN = hyper_KNN.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### CROSS VALIDATION"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> Sebelum melanjutkan proses modelling, sebaiknya kita lakukan *cross validation* terlebih dahulu untuk mengetahui apakah model benar - benar sudah fit atau terdapat overfitting maupun underfitting."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Score pada train data\n",
    "\n",
    "LRG_score_tr = LRG.score(X_train, y_train)\n",
    "KNN_score_tr = KNN.score(X_train, y_train)\n",
    "hyper_LRG_score_tr = hyper_LRG.score(X_train, y_train)\n",
    "hyper_KNN_score_tr = hyper_KNN.score(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Score pada test data\n",
    "\n",
    "LRG_score_ts = LRG.score(X_test, y_test)\n",
    "KNN_score_ts = KNN.score(X_test, y_test)\n",
    "hyper_LRG_score_ts = hyper_LRG.score(X_test, y_test)\n",
    "hyper_KNN_score_ts = hyper_KNN.score(X_test, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_score = pd.DataFrame({'Model' : ['LogisticRegression', 'KNNeighbors'], \n",
    "                            'Normal': [LRG_score_tr, KNN_score_tr], \n",
    "                            'Hyper' : [hyper_LRG_score_tr, hyper_KNN_score_tr]})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "test_score = pd.DataFrame({'Model' : ['LogisticRegression', 'KNNeighbors'], \n",
    "                           'Normal': [LRG_score_ts, KNN_score_ts], \n",
    "                           'Hyper' : [hyper_LRG_score_ts, hyper_KNN_score_ts]})"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> `pd.concat()` untuk menampilkan score berdasarkan *accuracy*."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "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></th>\n",
       "      <th>Model</th>\n",
       "      <th>Normal</th>\n",
       "      <th>Hyper</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">Train Score</th>\n",
       "      <th>0</th>\n",
       "      <td>LogisticRegression</td>\n",
       "      <td>0.982878</td>\n",
       "      <td>0.982878</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>KNNeighbors</td>\n",
       "      <td>0.982965</td>\n",
       "      <td>0.982902</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">Test Score</th>\n",
       "      <th>0</th>\n",
       "      <td>LogisticRegression</td>\n",
       "      <td>0.982441</td>\n",
       "      <td>0.982441</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>KNNeighbors</td>\n",
       "      <td>0.982228</td>\n",
       "      <td>0.982472</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                            Model    Normal     Hyper\n",
       "Train Score 0  LogisticRegression  0.982878  0.982878\n",
       "            1         KNNeighbors  0.982965  0.982902\n",
       "Test Score  0  LogisticRegression  0.982441  0.982441\n",
       "            1         KNNeighbors  0.982228  0.982472"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.concat([train_score, test_score], keys = ['Train Score', 'Test Score'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> Berdasarkan score *accuracy* yang di tampilkan di atas, sepertinya tidak terlihat adanya overfitting. Namun untuk lebih memastikan lagi, maka kita pastikan dengan menampilkan `learning_curve()`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x396 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x396 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Learning curve untuk default parameter\n",
    "\n",
    "curve = [LRG, KNN]\n",
    "\n",
    "for i in curve:\n",
    "    learning_curve(i, X, y, scoring='accuracy')\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> Berdasarkan plot *learning curves* di atas, sepertinya model memang stabil dan tidak terdapat overfitting pada kedua model ini."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x396 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x396 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Learning curve untuk tuned parameter\n",
    "\n",
    "curve = [hyper_LRG, hyper_KNN]\n",
    "\n",
    "for i in curve:\n",
    "    learning_curve(i, X, y, scoring='accuracy')\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> Hasil yang sama baiknya juga di tunjukkan pada model dengan tuning hyperparameter. Sehingga dapat di simpulkan bahwa tidak terdapat overfitting pada kedua model pada kedua parameter. Detail *kfold* nya bisa di lihat di bawah ini."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "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>LRG kfold</th>\n",
       "      <th>KNN kfold</th>\n",
       "      <th>Hyper LRG kfold</th>\n",
       "      <th>Hyper KNN kfold</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.982863</td>\n",
       "      <td>0.982587</td>\n",
       "      <td>0.982863</td>\n",
       "      <td>0.982902</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.982863</td>\n",
       "      <td>0.982587</td>\n",
       "      <td>0.982863</td>\n",
       "      <td>0.982902</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.982882</td>\n",
       "      <td>0.982705</td>\n",
       "      <td>0.982882</td>\n",
       "      <td>0.982902</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.982902</td>\n",
       "      <td>0.982527</td>\n",
       "      <td>0.982902</td>\n",
       "      <td>0.982902</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.982882</td>\n",
       "      <td>0.982665</td>\n",
       "      <td>0.982882</td>\n",
       "      <td>0.982902</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.982863</td>\n",
       "      <td>0.982587</td>\n",
       "      <td>0.982863</td>\n",
       "      <td>0.982902</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0.982902</td>\n",
       "      <td>0.982626</td>\n",
       "      <td>0.982902</td>\n",
       "      <td>0.982902</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0.982862</td>\n",
       "      <td>0.982665</td>\n",
       "      <td>0.982862</td>\n",
       "      <td>0.982902</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>0.982902</td>\n",
       "      <td>0.982685</td>\n",
       "      <td>0.982902</td>\n",
       "      <td>0.982902</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>0.982862</td>\n",
       "      <td>0.982626</td>\n",
       "      <td>0.982862</td>\n",
       "      <td>0.982902</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   LRG kfold  KNN kfold  Hyper LRG kfold  Hyper KNN kfold\n",
       "0   0.982863   0.982587         0.982863         0.982902\n",
       "1   0.982863   0.982587         0.982863         0.982902\n",
       "2   0.982882   0.982705         0.982882         0.982902\n",
       "3   0.982902   0.982527         0.982902         0.982902\n",
       "4   0.982882   0.982665         0.982882         0.982902\n",
       "5   0.982863   0.982587         0.982863         0.982902\n",
       "6   0.982902   0.982626         0.982902         0.982902\n",
       "7   0.982862   0.982665         0.982862         0.982902\n",
       "8   0.982902   0.982685         0.982902         0.982902\n",
       "9   0.982862   0.982626         0.982862         0.982902"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cv = StratifiedKFold(n_splits = 10, shuffle= True, random_state=0)\n",
    "LRG_kfold = cross_val_score(LRG, X_train, y_train, cv = cv, scoring = 'accuracy')\n",
    "KNN_kfold = cross_val_score(KNN, X_train, y_train, cv = cv, scoring = 'accuracy')\n",
    "hyper_LRG_kfold = cross_val_score(hyper_LRG, X_train, y_train, cv = cv, scoring = 'accuracy')\n",
    "hyper_KNN_kfold = cross_val_score(hyper_KNN, X_train, y_train, cv = cv, scoring = 'accuracy')\n",
    "\n",
    "pd.DataFrame({'LRG kfold' : LRG_kfold.tolist(),\n",
    "              'KNN kfold' : KNN_kfold.tolist(),\n",
    "              'Hyper LRG kfold' : hyper_LRG_kfold.tolist(),\n",
    "              'Hyper KNN kfold' : hyper_KNN_kfold.tolist()})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Model</th>\n",
       "      <th>Default</th>\n",
       "      <th>Hyper</th>\n",
       "      <th>Mean Kfold</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Random Forest</td>\n",
       "      <td>0.982878</td>\n",
       "      <td>0.982878</td>\n",
       "      <td>0.982878</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Support Vector</td>\n",
       "      <td>0.982965</td>\n",
       "      <td>0.982902</td>\n",
       "      <td>0.982902</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            Model   Default     Hyper  Mean Kfold\n",
       "0   Random Forest  0.982878  0.982878    0.982878\n",
       "1  Support Vector  0.982965  0.982902    0.982902"
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.DataFrame({'Model' : ['Random Forest', 'Support Vector'], \n",
    "              'Default' : [LRG_score_tr, KNN_score_tr],\n",
    "              'Hyper': [hyper_LRG_score_tr, hyper_KNN_score_tr], \n",
    "              'Mean Kfold' : [hyper_LRG_kfold.mean(), hyper_KNN_kfold.mean()]})"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## MODEL PREDICTION & PERFORMANCE EVALUATION"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> Selanjutnya kita lakukan prediksi dan evaluasi model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Logistic Regression\n",
    "\n",
    "yp_LRG = hyper_LRG.predict(X_test)\n",
    "\n",
    "# KNNeighbors\n",
    "\n",
    "yp_KNN = hyper_KNN.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Label untuk confusion matrix\n",
    "\n",
    "labels = ['Fraud No', 'Fraud Yes']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### LOGISTIC REGRESSION"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "    Fraud No       0.98      1.00      0.99    124548\n",
      "   Fraud Yes       0.17      0.00      0.00      2222\n",
      "\n",
      "    accuracy                           0.98    126770\n",
      "   macro avg       0.57      0.50      0.50    126770\n",
      "weighted avg       0.97      0.98      0.97    126770\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x360 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# LogisticRegression evaluation\n",
    "\n",
    "print(classification_report(y_test, yp_LRG, target_names = labels))\n",
    "f, ax = plt.subplots(figsize=(8,5))\n",
    "sns.heatmap(confusion_matrix(y_test, yp_LRG), annot=True, fmt=\".0f\", ax=ax, cmap = 'Blues')\n",
    "\n",
    "ax.xaxis.set_ticklabels(labels)\n",
    "ax.yaxis.set_ticklabels(labels)\n",
    "\n",
    "plt.title('LogisticRegression Matrix')\n",
    "plt.xlabel('PREDICTED')\n",
    "plt.ylabel('ACTUAL')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> Melihat evaluasi model **Logistic Regression**, sepertinya masih harus di lakukan balancing data lebih lanjut."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### K-NEAREST NEIGHBORS"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "    Fraud No       0.98      1.00      0.99    124548\n",
      "   Fraud Yes       0.00      0.00      0.00      2222\n",
      "\n",
      "    accuracy                           0.98    126770\n",
      "   macro avg       0.49      0.50      0.50    126770\n",
      "weighted avg       0.97      0.98      0.97    126770\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAd4AAAFKCAYAAACzaT6WAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+j8jraAAAgAElEQVR4nO3de1xUdf7H8ddwVRlow7LygoqKZa4XJM1C3drKS5pmaohRpnkrIU0N75e8p6ilsWVZmimIyZY9rG23y2IkoZFo2pqJW95vicYgMAjz+6Nfk6yiQnNmZOb93Mc8HnDmnPP9nlnyPZ9zvud7TDabzYaIiIg4hZerOyAiIuJJFLwiIiJOpOAVERFxIgWviIiIEyl4RUREnEjBKyIi4kQKXrkmHDp0iNatW5dZ9uGHH9KuXTsyMjI4dOgQTZs2Zf369WXWWbFiBePHjwdg6dKltG/fnpMnT5ZZp3v37mRmZl62/W+//Za4uLgr9rNp06acPn36ouWpqakMGzbsits7wvjx42natClfffVVmeWHDh3i1ltv5YUXXrjiPiZPnsyuXbsu+d6kSZPYsmWLQ/oqIhdT8Mo1KTk5mXnz5rFy5Urat28PgJeXF/Pnz2f//v3lbmexWIiPj6eit6f/+c9/5uWXX/5DfXam2rVr8/7775dZ9t5771GzZs2r2n7Lli3lfkazZ8/mrrvu+sN9FJFLU/DKNWf58uWsXLmStWvXctttt9mXV6tWjSeffJKxY8ditVovue1DDz3EiRMnePPNNy/5fk5ODoMGDaJ379707NmTd999F4DMzEy6d+8OwOnTpxk2bBhdu3alf//+xMXFsXTpUvs+li5dSu/evbn33ntZs2aNffnJkycZPHgwPXr0YPjw4fbK+9ixYwwfPpwePXrQvXt33njjDeDXCrVTp04MGjSIzp07c/ToUaZNm0aPHj3o3bs3cXFx5OfnX/I4unXrxmeffUZhYaF92UcffUTXrl3tv2dnZzNgwAD69u3LX/7yFyZOnAjA4sWLOXHiBGPHjmXHjh3ExMQwcuRIunXrxurVq4mJieEf//gHX331Fe3ateP48eOUlpYSExPDK6+8Us7/ayJytRS8ck158cUXSUhIICYmhrp16170/ogRI6hRowaLFy++5Pb+/v4kJCSQmJjI7t27y7x3/vx54uLiGDNmDKmpqbzzzju8+eabZGdnl1lv1qxZNG7cmI8++oiXXnqJb775psz79erVIzU1lWXLljFv3jyKi4sB+O9//8vUqVP54IMPCAsLY/bs2QCMHTuWdu3a8cEHH5CUlMTGjRvZtGkT8GsoP/3003z88cccPnyYrVu3snHjRlJTU6lXrx7ff//9JY8zODiY1q1b89lnnwHw9ddf06hRI6677jr7Om+//TZxcXGsX7+eTZs28dlnn7Fr1y5Gjx5NrVq1WLhwIS1btgQgKCiIDz/8kJiYGPv2d955J1FRUUyePJnExET8/PwYMWLEJfsjIldPwSvXjHPnzrF3716WL19OQkIC33333UXreHl5sWDBAlJTU/nyyy8vuZ+mTZsyatQoxowZw7lz5+zLf/zxRw4cOMDEiRPp2bMnjz32GIWFhRe1k5aWxqOPPgpArVq16NKlS5n3f6uMb7vtNqxWKxaLBYC77rqL+vXrA9CnTx+2bNnCuXPn+OabbxgwYAAAgYGB9O7dm82bNwPg4+NDq1atAAgLC8Pb25u+ffuyZMkSOnfuTHh4eLmfV8+ePdm4cSPw62nmhx9+uMz78+bNIy8vj1dffZUZM2ZQVFRU5vO4UERExCWXx8bGcvbsWZKSkliwYAFeXvonQ+SP0n9Fcs2oVq0af/vb3+jUqRPDhg1j5MiRnDlz5qL1brnlFmbMmEF8fDy5ubmX3FdMTAz169e3V50AJSUlBAYG8v7779tfKSkpPPLII2W29fHxKXP983/DxsfHBwCTyQRgX9fb29u+TmlpKT4+PpSWll50LbW0tJTz588D4OfnZ99fUFAQ77//PvHx8Xh7ezNq1Kgyp7L/11//+ld27NjB0aNH2bZtGx06dCjz/mOPPUZaWhqhoaE888wz1KpVq9zrujVq1Ljk8ry8PE6ePInJZOKnn34qty8icvUUvHLN8PLywtfXF4ChQ4fSuHFjxowZQ2lp6UXrdunShY4dO7Jq1apy9zd37lzS0tLsgdGwYUOqVatmH5R09OhRunfvftHo3k6dOtmv/ebm5vLJJ5/YQ/ZyMjMzOXLkCPDr4LCOHTtiNptp2bKlPUDz8vJ47733Ljl46fPPP2fgwIG0bt2a2NhYevXqVe7IY/g1tO+//36ef/557r33XnuAA/zyyy98++23jB07lgceeIBjx45x4MAB+2fp7e1tD//LmTRpEg899BBz585l7Nix5OXlXXEbEbk8Ba9ck0wmE/PnzycnJ4clS5Zccp3JkydTu3btcvcRHBxc5hqsn58fiYmJvPvuu/To0YNBgwbx7LPP0qZNmzLbTZgwgf3799OjRw/i4uKoXbs21apVu2Kfw8LCmDhxIt27d+fIkSP225wWLlxIRkYGPXr0oE+fPjzwwAP07t37ou07duxI48aN6d69O71792b79u0888wzl22zZ8+ebN269aLTzEFBQQwdOpSHH36Y7t27s3z5csLDw+1fQu6//37GjRtHenp6uftes2YNR48eZeTIkXTo0IHIyEimTJlyxc9BRC7PpMcCipS1Zs0amjVrRuvWrbFarURHRxMbG0unTp1c3TURcQM+V15FxLM0btyYmTNnUlpaSnFxMV26dFHoiojDqOIVERFxIl3jFRERcSIFr4iIiBMpeEVERJzomhxcVb31SFd3QeQPy922zNVdEHGIagYmRWX+vS/YXrX/21LFKyIi4kTXZMUrIiIewuR59Z+CV0REXOcqpmN1NwpeERFxHVW8IiIiTqSKV0RExIk8sOL1vCMWEZFrh8lU8ddV2LFjBzExMQD85z//ITo6mpiYGAYPHsypU6cASElJoXfv3vTr14/PP/8cgMLCQmJjY4mOjmbIkCGcPn0agOzsbPr27UtUVBTLlv1+O9OyZcvo06cPUVFR7Ny586r6popXRERcx4CK9/XXX2fjxo1Ur14dgNmzZzNlyhRuu+02kpOTef3113nqqadYvXo1GzZsoKioiOjoaO6++26SkpIICwsjNjaWTZs2kZiYyOTJk5k2bRpLly6lXr16DB06lN27dwOwdetW1q9fz9GjR4mNjWXDhg1X7J8qXhERcR0DKt6QkBCWLl1q/33RokXcdtttAJSUlODv78/OnTtp3bo1fn5+BAYGEhISwp49e8jKyqJDhw7Ar8/IzsjIwGKxYLVaCQkJwWQyERkZSUZGBllZWURGRmIymahduzYlJSX2CvlyFLwiIuI6Jq+Kv66gc+fO+Pj8fkK3Vq1aAHzzzTe88847DBw4EIvFQmBgoH2dgIAALBZLmeUBAQHk5eVhsVgwm81l1r3c8ivRqWYREXEdJ41q/vDDD/nb3/7G8uXLCQ4Oxmw2k5+fb38/Pz+fwMDAMsvz8/MJCgq65LpBQUH4+vpech9XoopXRERcx4CK93+9//77vPPOO6xevZp69eoB0KJFC7KysigqKiIvL4+cnBzCwsIIDw8nLS0NgM2bN9OmTRvMZjO+vr4cOHAAm81Geno6ERERhIeHk56eTmlpKUeOHKG0tJTg4OAr9kcVr4iIuI7BFW9JSQmzZ8/mlltuITY2FoA77riDuLg4YmJiiI6OxmazMXr0aPz9/enfvz/x8fH0798fX19fEhISAJgxYwZjx46lpKSEyMhIWrZsCUBERASPPvoopaWlTJ069ar6ZLLZbDZjDrfy9HQicQd6OpG4C0OfTtRxeoW3Kdhc8W2uJap4RUTEdTxwAg0Fr4iIuI6XpowUERFxHg+seD3viEVERFxIFa+IiLiOnk4kIiLiRB54qlnBKyIirqOKV0RExIlU8YqIiDiRKl4REREnUsUrIiLiRKp4RUREnEgVr4iIiBOp4hUREXEiVbwiIiJOpOAVERFxIp1qFhERcSJVvCIiIk6kildERMSJVPGKiIg4kQdWvJ73VUNERMSFVPGKiIjLmDyw4lXwioiIyyh4RUREnMnzclfBKyIirqOKV0RExIkUvCIiIk6k4BUREXEiBa+IiIgzeV7uKnhFRMR1VPGKiIg4kYJXRETEiRS8IiIiTqTgFRERcSbPy10Fr4iIuI4nVrx6LKCIiLiMyWSq8Otq7Nixg5iYGAB++ukn+vfvT3R0NNOmTaO0tBSAlJQUevfuTb9+/fj8888BKCwsJDY2lujoaIYMGcLp06cByM7Opm/fvkRFRbFs2TJ7O8uWLaNPnz5ERUWxc+fOq+qbgldERFzGiOB9/fXXmTx5MkVFRQDMnTuXUaNGsXbtWmw2G59++iknT55k9erVJCcns2LFChYtWoTVaiUpKYmwsDDWrl1Lr169SExMBGDatGkkJCSQlJTEjh072L17N7t372br1q2sX7+eRYsWMWPGjKs6ZgWviIi4lZCQEJYuXWr/fffu3bRt2xaAjh07smXLFnbu3Enr1q3x8/MjMDCQkJAQ9uzZQ1ZWFh06dLCvm5GRgcViwWq1EhISgslkIjIykoyMDLKysoiMjMRkMlG7dm1KSkrsFfLlKHhFRMR1TJV4XUHnzp3x8fl9CJPNZrNXygEBAeTl5WGxWAgMDLSvExAQgMViKbP8wnXNZnOZdS+3/Eo0uEpERFzGGYOrvLx+rzHz8/MJCgrCbDaTn59fZnlgYGCZ5ZdbNygoCF9f30vu44r9ccRBiYiIVIZRg6su1KxZMzIzMwHYvHkzERERtGjRgqysLIqKisjLyyMnJ4ewsDDCw8NJS0uzr9umTRvMZjO+vr4cOHAAm81Geno6ERERhIeHk56eTmlpKUeOHKG0tJTg4OAr9kcVr4iIuIwzKt74+HimTJnCokWLCA0NpXPnznh7exMTE0N0dDQ2m43Ro0fj7+9P//79iY+Pp3///vj6+pKQkADAjBkzGDt2LCUlJURGRtKyZUsAIiIiePTRRyktLWXq1KlX1R+TzWazGXa0lVS99UhXd0HkD8vdtuzKK4lUAdUMLNFqD0ut8DZHXuttQE+cRxWviIi4jufNn6HgFRER1/HEmasUvCIi4jIKXhERESdS8IqIiDiT5+Wugreqev2FGHb/cIQlqz+lmr8vS8b3I6J5fUwmE9u+/ZFR81IoLCq2r1+/dk22rH2eHk+/wjffHQAgaeFT/LlJHSwFv85nunnbXp5P+H2Eoa+PN5++OZrUf21nyepPAbirVSgvjuuDj7cXRUXFjJ6/3r4/EWfanPZvXl6SgNVqJSysKdNnzikzi5BUDZ5Y8WoCjSqmacOb+Oi1WB6+r5V9Wfzgzvj4eHFHv7nc0W8O1av5Mm7QA/b3/f18eGv24/j5lv2e1a5FQ+5/agl3Rs3jzqh5ZUIXYMG4R2hQt2aZZW/OfoJJS97jzqh5JKz8F6+/EGPAUYpc3unTp5k6eQIJS5aycdPH1Klbj5cWLXR1t6QSnDGBxrVGwVvFDO/XkZV/zyD1X9vty9K/2ce81z/GZrNRWmpjx55DhNzy++wpSyb0Y/XGTH4+Y7Evq1+7JuYa/rwypT/bUiby2vTHuD6ohv39/g/ewXXm6vzji91l2vf28rKvZw6oVqaqFnGWjC3pNG/+Z+rXbwBAv6j+fLjpA67BaQnkChS8DlRSUsKaNWuYMWMGq1atwmq1GtWURxk9fz3r/vF1mWWffrWHfQdOABByy/WMHHCPPZgHPtweXx9v3vr7ljLb1Ao283nm98TNTqZd1FzyC4p4bfoAAG5vXJtnou/hmZlJF7U/bPoa3pgZw75/zOSlCY/y3Pz1RhymyGUdO3qMm26+2f77TTfdjMViKTNvrlQNCl4HmjJlCgcOHODuu+/m8OHDTJ482aim5P+1vq0en6wYzd+S0/joi120urUuT/XpQOzs5IvW3bbrJx4d8zqHjp+htNTGrFc/pEtkc4LM1XhjZgxPTXmbc4VlvyzVCg4kcWp/HnjqJRp3mcKgSatYu2AwNar5OesQRQCw2Uov+Q/whZPhSxVhwNOJrnWGDa766aefWLNmDQD33XcfUVFRRjUlQN/ObVgyoR/Pzfu9Ih7QvR1BAdX4fOUYAG658Tremv0EE5e8x5lfzvGnoBpsSvsWAJMJSm2lPHBXM64PqsHKOQMBqHfz9fz1zlsJNFdj1w+HOXD0tH0w1Qf/3smCcY9wa+jNGmAlTnXzLbfw7c4d9t9PnDhOUNB11KhR4zJbybXIHSrYijIseIuKiigoKKB69eoUFhZSUlJiVFMer1vH5ix8vk+ZEcsA4xZuYNzCDfbf92yawZOTVvHNdwd44O5mLIrvy5btOeT+co7Rj9/H3z/J5t1/fsO7//zGvs3yGY/x3b6jLFn9KY1DatGsUW0ah9Ri34ET3NG8PtWr+fHDTyecerwi7e+KJGHBfH766Ufq12/A+nXJ/OXev7q6WyJXxbDgffzxx+nZsydNmjRh3759xMbGGtWUx5s7+mFMJkicGm1flpG9n9HzUsrd5p9ffkdi0r/57K3n8PIysXvfEZ5+Ye1l29l34ARxs5NJWvgUNpuNgkIr/ce8Tl5+ocOOReRq1KxZkxdmzWXsqDiKzxdTt14Is+fMd3W3pBI8seI19OlEZ86c4eDBg9StW5frr7/+qrfT04nEHejpROIujHw6UeOxH1V4m30LuxrQE+dx+Me5bFn5/9iMHKlAFRGR33lixevwIYA33HBDmVdAQABr164lLS3N0U2JiEgVZzJV/FXVObzivXD08tdff82UKVMYMGAAw4cPd3RTIiJSxXlixWvImfvi4mIWLVpERkYGCQkJNGvWzIhmRESkivPA3HX8qebvvvuOPn364O3tzfr16xW6IiJSLi8vU4VfVZ3DK95+/foREBDAtm3biIn5dQJ9m82GyWQiOfniGZRERMRzeWLF6/Dg/fjjjx29SxERcVO6xusAderUcfQuRUTETXlg7ho3c5WIiMiVqOIVERFxIgWvA0yYMKHc9+bOnevo5kREpArzwNx1/O1E3bp1o1u3bpw9e5bQ0FD69OlD06ZNsVqtV95YREQ8ytU8+P5/X1Wdw4O3Q4cOdOjQgcLCQoYMGUKbNm0YOHAgp0+fdnRTIiJSxXnilJEOD97fnDt3joyMDCwWC1988QXFxcVGNSUiIlWUKl4Hmj17NklJSfTr14+UlBQWL15sVFMiIlJFeWLFa9io5kaNGvHyyy8btXsREZEqybDgjYyMtP985swZ6tWrx0cfVfyBxyIi4r7c4dRxRRkWvOnp6fafDx8+zLJly4xqSkREqigPzF3nTKBRp04d9u/f74ymRESkClHF60DPPfec/QM9ceIENWvWNKopERGpojwwd40L3qioKPvP/v7+NG/e3KimRESkivLEitew24nCwsI4ceIER44cYf/+/bzxxhtGNSUiIlWUbidyoLi4OBo0aMDevXvx9/enevXqRjUlIiJVlBEVb3FxMePHj+fw4cN4eXkxc+ZMfHx8GD9+PCaTiSZNmjBt2jS8vLxISUkhOTkZHx8fRowYwT333ENhYSHjxo3j559/JiAggPnz5xMcHEx2djazZ8/G29ubyMhIRo4cWan+GVbxArzwwgs0bNiQt956i7NnzxrZlIiIVEFGVLxpaWmcP3+e5ORknnnmGZYsWcLcuXMZNWoUa9euxWaz8emnn3Ly5ElWr15NcnIyK1asYNGiRVitVpKSkggLC2Pt2rX06tWLxMREAKZNm0ZCQgJJSUns2LGD3bt3V+qYDQ3eoqIiCgoKMJlMnDt3zsimRESkCjJiysiGDRtSUlJCaWkpFosFHx8fdu/eTdu2bQHo2LEjW7ZsYefOnbRu3Ro/Pz8CAwMJCQlhz549ZGVl0aFDB/u6v01/bLVaCQkJwWQyERkZSUZGRqWO2bBTzQMGDGDVqlXcfffddOrUiTZt2hjVlIiIVFFGnGquUaMGhw8fpmvXruTm5vLqq6+ybds2e1sBAQHk5eVhsVgIDAy0bxcQEIDFYimz/MJ1zWZzmXUPHjxYqf4ZFrxFRUUMHToUgK5du5bpsIiICBgzWGrlypVERkYyZswYjh49yhNPPFHmQT35+fkEBQVhNpvJz88vszwwMLDM8sutGxQUVKn+GXaqOSUlxf6zQldERC7FiFPNQUFB9or1uuuu4/z58zRr1ozMzEwANm/eTEREBC1atCArK4uioiLy8vLIyckhLCyM8PBw0tLS7Ou2adMGs9mMr68vBw4cwGazkZ6eTkREROWO2Waz2Sq15RX069cPq9VKw4YN8fL6Nd8TEhKuatvqrSs3UkzkWpK7TdOkinuoZuAch/e8tKXC23z+7F2XfT8/P5+JEydy8uRJiouLefzxx2nevDlTpkyhuLiY0NBQZs2ahbe3NykpKaxbtw6bzcawYcPo3LkzBQUFxMfHc/LkSXx9fUlISODGG28kOzubOXPmUFJSQmRkJKNHj67UMRsWvFu3br1o2W8Xtq9EwSvuQMEr7sLI4L335YoPUPosrr0BPXEeh3+ciYmJPP3007Rt25YTJ05Qq1YtRzchIiJuwh0mxKgoh1/j/eqrr+w/jx071tG7FxERN+JlMlX4VdU5PHgvPHNt0FlsERGRKsvhp5ovHHHmiZNfi4jI1fPEmHB48O7evZuoqChsNhv79u2z/2wymUhOTnZ0cyIiUoV5YoHm8ODduHGjo3cpIiJuysvzctfxwVunTh1H71JERNyUKl4REREn8sDcVfCKiIjrmPC85FXwioiIy+gar4iIiBPpGq+IiIgTeWDuKnhFRMR13GEKyIpS8IqIiMt4YO5Wbq7m8PBwR/dDREQ80NU8+P5/X1VdpSpePfxAREQcwQ1ytMIqFbzu8I1DRERcT9d4L3DkyJFLLrfZbKp4RUTEITwvdi8TvI899hgmk+mSIXv99dcb2ikREfEMnngGtdzg/eyzz8rd6L///a8hnREREXF3V32N9/z58/zzn/8kOTmZb7/9lu3btxvZLxER8QCaMvISDh48yLp160hNTeWXX35h+PDhLFmyxBl9ExERN+eJp5rLvY/3X//6F4MHD6Zv376cPXuWBQsWUKtWLUaOHElwcLAz+ygiIm7KZKr4q6ort+KNjY2la9eurFu3jvr16wOe+c1ERESM44m5Um7wbty4kdTUVKKjo6lTpw4PPvggJSUlzuybiIi4OU+8xlvuqeawsDDGjx9PWloaQ4cOJTMzk1OnTjF06FDS0tKc2UcREXFTnjhl5BXnavbx8eG+++4jMTGRzZs3c+edd5KQkOCMvomIiJszVeJV1V31zFVeXl4EBQUxaNAgBg0aZHjHRETE/WnKyAtcauaq3NxcwsLCWLx4MbVr13ZKB0VExH15YO5WfOaq9957j+nTp7N8+XLDOiUiIp7BHa7ZVlSFn8fbq1cvjh49akRfRETEw+g+XhERESfSNd4rsFgsrF+/3j6hhoiIyB/hgblbfvDeeuutZc69m0wmgoKCaN++PdOnTze0U6cylxq6fxERuTZ44jXecoN3z549zuyHiIh4oAoPNHID5R5zbGysM/shIiIeyBNnriq34j148KAz+yEiIuIwr732Gp999hnFxcX079+ftm3bMn78eEwmE02aNGHatGl4eXmRkpJCcnIyPj4+jBgxgnvuuYfCwkLGjRvHzz//TEBAAPPnzyc4OJjs7Gxmz56Nt7c3kZGRjBw5slJ9Kzd4z507x9dff11mAo0L3XHHHZVqUERE5DdGPCQhMzOT7du3k5SUREFBAW+++SZz585l1KhRtGvXjqlTp/Lpp5/SqlUrVq9ezYYNGygqKiI6Opq7776bpKQkwsLCiI2NZdOmTSQmJjJ58mSmTZvG0qVLqVevHkOHDmX37t3cfvvtFe5fucF78uRJXn755UsGr8lk4u23365wYyIiIhcyInjT09MJCwvjmWeewWKx8Pzzz5OSkkLbtm0B6NixI19++SVeXl60bt0aPz8//Pz8CAkJYc+ePWRlZfHUU0/Z101MTMRisWC1WgkJCQEgMjKSjIwMxwZv/fr1Fa4iImIoI67Z5ubmcuTIEV599VUOHTrEiBEjsNls9rYCAgLIy8vDYrEQGBho3y4gIACLxVJm+YXrms3mMutW9pKsJtAQERGXMaLi/dOf/kRoaCh+fn6Ehobi7+/PsWPH7O/n5+cTFBSE2WwmPz+/zPLAwMAyyy+3blBQUKX6V+6o5rFjx1ZqhyIiIlfLiCkj27RpwxdffIHNZuP48eMUFBTQvn17MjMzAdi8eTMRERG0aNGCrKwsioqKyMvLIycnh7CwMMLDw+3Pnd+8eTNt2rTBbDbj6+vLgQMHsNlspKenExERUbljtpU3egrYsGEDTZo0oUWLFgAkJCRQv359+vTpU6nGrla+tdwuiVQZ3kZ8lRdxgWoGnhsd/+HeCm8zr1vYFdd58cUXyczMxGazMXr0aOrWrcuUKVMoLi4mNDSUWbNm4e3tTUpKCuvWrcNmszFs2DA6d+5MQUEB8fHxnDx5El9fXxISErjxxhvJzs5mzpw5lJSUEBkZyejRoytzyOUH7+rVq9m4cSPz588nNDQUgG3btjFv3jweeeQRoqOjK9Xg1VDwijtQ8Iq7MDJ4J1YieOdcRfBey8o91fzuu+/y1ltv2UMXfr2F6PXXXyc5OdkpnRMREfempxNdwMvLq8wIrt8EBwfj5eWJk3yJiIijeeLTicpNUG9vb37++eeLlp86dYqSkhJDOyUiIp7BEyvecoP3scceY8iQIXz11Vf88ssvFBUV8fXXXzNixAiioqKc2UcREXFTXqaKv6q6ck819+rVi6KiIiZMmMDRo0cBCAkJYfDgwfbZP0RERP4ITzzVfNnbiX6Tm5uLzWbjq6++IikpiV27drF9+3bDOqVRzeIONKpZ3IWRo5pnfrKvwttMua+xAT1xnit+nAcPHiQlJYXU1FTOnj3L8OHDeemll5zRNxERcXOe+P203Gu8//rXvxg8eDB9+/blzJkzvPjii9SqVYuRI0cSHBzszD6KiK868ygAABeESURBVIibMlXif1VduRVvbGwsXbt2Zd26ddSvXx8wZjJrERHxXJ5Y8ZYbvBs3biQ1NZXo6Gjq1KnDgw8+qNuIRERE/qByTzWHhYUxfvx40tLSGDp0KJmZmZw6dYqhQ4faJ48WERH5IzzxdqKrGtX8m9OnT/Pee+/x3nvvsXHjRsM6pVHN4g40qlnchZGjmhf8e3+Ftxn3l9Arr3QNq1DwOouCV9yBglfchZHBm5BW8eAd06lqB6+BH6eIiMjleeKYXQWviIi4jCfOXKXgFRERl/HEKzIKXhERcRkPLHgVvCIi4jpebjATVUUpeEVExGVU8YqIiDiRrvGKiIg4kUY1i4iIOJEH5q6CV0REXEcVr4iIiBN5YO4qeEVExHXKfUSeG1PwioiIy5g8sORV8IqIiMt4Xux6ZpUvIiLiMqp4RUTEZTSqWURExIk8L3YVvCIi4kIeWPAqeEVExHU0qllERMSJPHGEr4JXRERcRhWviIiIE3le7HpmlS8iItcIk8lU4dfV+Pnnn+nUqRM5OTn89NNP9O/fn+joaKZNm0ZpaSkAKSkp9O7dm379+vH5558DUFhYSGxsLNHR0QwZMoTTp08DkJ2dTd++fYmKimLZsmV/6JgVvCIi4jJelXhdSXFxMVOnTqVatWoAzJ07l1GjRrF27VpsNhuffvopJ0+eZPXq1SQnJ7NixQoWLVqE1WolKSmJsLAw1q5dS69evUhMTARg2rRpJCQkkJSUxI4dO9i9e/cfOmYRERGXMKLinT9/PlFRUdSqVQuA3bt307ZtWwA6duzIli1b2LlzJ61bt8bPz4/AwEBCQkLYs2cPWVlZdOjQwb5uRkYGFosFq9VKSEgIJpOJyMhIMjIyKn3MCl4REXEZUyVel5OamkpwcLA9PAFsNps9sAMCAsjLy8NisRAYGGhfJyAgAIvFUmb5heuazeYy6+bl5VX6mDW4SkREXMbRg5o3bNiAyWQiIyOD//znP8THx9uv0wLk5+cTFBSE2WwmPz+/zPLAwMAyyy+3blBQUKX7qIpXRERcxgtThV+Xs2bNGt555x1Wr17Nbbfdxvz58+nYsSOZmZkAbN68mYiICFq0aEFWVhZFRUXk5eWRk5NDWFgY4eHhpKWl2ddt06YNZrMZX19fDhw4gM1mIz09nYiIiEofsypeERFxGWfcxhsfH8+UKVNYtGgRoaGhdO7cGW9vb2JiYoiOjsZmszF69Gj8/f3p378/8fHx9O/fH19fXxISEgCYMWMGY8eOpaSkhMjISFq2bFnp/phsNpvNUQfnKPnWa65LIhXm7eWJdyiKO6pmYIm2adeJCm/zYPNaBvTEeVTxioiIy3jgxFW6xisiIuJMqnhFRMRlrjRYyh0peEVExGU88VSzgldERFxGwSsiIuJEJp1qFhERcR5PvOtOwSsiIi6jildERMSJdI1XRETEiVTxSpW26YONvL1yBSaTiWrVqvH8hEmENmrCvNkvsHvXt9hsNpr/uQXjJ/36gOjdu75l4fw5FBQUUFpSwhODhvBgj4cAWL3qTd7/eyre3t5cf30wk6bNoF69EBcfocjvNqf9m5eXJGC1WgkLa8r0mXPKPLpNqgZPvMaruZrdxI//3c/QQU+wJmUDN95Yi/TNacyZOZ0He/Tk2LGjzJg1F5vNxuQJ4wgJacDwZ2J58IF7mfbCbNq1v4vjx44R/Whv3np7LUePHOHFebNZtWYdZrOZlOS1fPzRh6xY9Y6rD7NK0VzNxjl9+jS9ez7IqneSqF+/AYsTFnAuP59JU6e7umtuyci5mr/Ym1vhbTqEXW9AT5xHU0a6CT8/P6bMmMmNN/46eXiz25tz6tQpwiMieGrocLy8vPD29qbprc04evQwVquVoSOeoV37uwC46eabuf76YI4fP0bNG25g4uRp9uqh2e3NOXr0iMuOTeR/ZWxJp3nzP1O/fgMA+kX158NNH3AN1hFyBSZTxV9VnWHfY7Zt20ZBQQE2m42ZM2fy7LPP0qNHD6Oa83i169Sldp26ANhsNhIWzKPTPffQ/q5I+zpHjhxm7TurmDztBfz9/enVu4/9vQ3r13EuP58/t2hFtWrV7MutVisvL0ng/gc6O+9gRK7g2NFj3HTzzfbfb7rpZiwWC/n5+TrdXMW4QY5WmGEV74IFC2jQoAFvv/02SUlJJCcnG9WUXKDg3Dnix4zi4MEDTJ0+y778u927eOqJx3i0/wA6drqnzDZvvbGc1xKXsWTZ38qEbu7p0zw9dBA1atRg5LOjnXYMIldis5ViukTp4+Wlk3hVjZfJVOFXVWfYX6m/vz81a9bEx8eHG2+8EavValRT8v+OHj3CwJj+eHl7s3zFKgKDggD4+KNNPD10MLGjnmPwkOH29a1WKxOef45/fLSJle8kEdb0Vvt7e7//nsf69+HW224nYckyfH39nH48IuW5+ZZbOHni9+e4njhxnKCg66hRo4YLeyWVYarEq6ozLHjNZjNPPvkkXbt2Zc2aNdxyyy1GNSVAfr6FoU8+zr333c+8BYvslWvavz/jxXmzSVz+Bl0fLHuqf/L4ceRb8lm5Osl+mhrg+LFjDH/qCYYMf4ax8RPw9vZ26rGIXEn7uyLZuXMHP/30IwDr1yXzl3v/6tpOSeV4YPIaNqrZarVy4MABGjduzN69e2nQoAF+fldXNWlUc8W9+cZrJC59icZNwsosLygoIC/vF2rVusm+rGWrcLp178GTMf2p36AB/v6/n16OGz2Gzz75F5s+eJ/6DRral/v5+fH22hTjD8SNaFSzsb7YnMbLixMoPl9M3XohzJ4zn+v+9CdXd8stGTmqOTPnbIW3adfoOgN64jyGBe/x48dZsGABubm5dO7cmaZNm9KyZcur2lbBK+5AwSvuQsHrWIadap4yZQqPPPIIVquViIgIZs+ebVRTIiJSRXni7USGBW9RURHt27fHZDIRGhqKv7+/UU2JiEgV5YGXeI27j9fPz48vvviC0tJSsrOzr/r6roiIeBB3SNIKcnjF+/333wMwc+ZMUlNTyc3N5c0332T69OmObkpERKo4UyX+V9U5fHBVly5diIqKYuDAgZXehwZXiTvQ4CpxF0YOrsr68ZcKb9OmQZABPXEeh1e8qamp/Pe//2Xw4MGcPHnS0bsXERE34onXeA27nWjbtm1MmDChzC1ECQkJV7WtKl5xB6p4xV0YWfF+81PFK97w+lW74jXk48zJySEhIYG2bdvSq1cvI5oQERE34A7XbCvK4cG7fPlykpOTmTp1Kn/5y18cvXsREXEj7nBfbkU5PHh37drFhg0buP76qv2gYhERMZ4H5q5x13j/CF3jFXega7ziLoy8xrvjYF6Ft2lZL9CAnjiPgR+niIjI5ekar4iIiBPpGq8DTJgwodz35s6d6+jmRESkCvPA3HX8BBrdunWjW7dunD17ltDQUPr06UPTpk2xWq2ObkpERKo6D5xBw+HB26FDBzp06EBhYSFDhgyhTZs2DBw4kNOnTzu6KRERqeI8ca5mwx4LeO7cOTIyMrBYLHzxxRcUFxcb1ZSIiEiVYdjtRDk5Obz00kvs27ePRo0aMXXqVG688car2la3E4k70O1E4i6MvJ3ouyP5Fd6mWe2Ay75fXFzMxIkTOXz4MFarlREjRtC4cWPGjx+PyWSiSZMmTJs2DS8vL1JSUkhOTsbHx4cRI0Zwzz33UFhYyLhx4/j5558JCAhg/vz5BAcHk52dzezZs/H29iYyMpKRI0dW6ph1H6+IQRS84i6MDN7/VCJ4b7tC8G7YsIE9e/YwadIkcnNzefjhh7n11lt58sknadeuHVOnTqVDhw60atWKQYMGsWHDBoqKioiOjmbDhg2sWbMGi8VCbGwsmzZtYvv27UyePJmePXuydOlS6tWrx9ChQxk1ahS33357hftv2McZGRlp//nMmTPUq1ePjz76yKjmRESkKjLg+2mXLl3o3Lmz/Xdvb292795N27ZtAejYsSNffvklXl5etG7dGj8/P/z8/AgJCWHPnj1kZWXx1FNP2ddNTEzEYrFgtVoJCQkBfs24jIyMSgWvYdd409PT7a+PP/6YVq1aGdWUiIhUUUYMrgoICMBsNmOxWIiLi2PUqFHYbDZM/3/TcEBAAHl5eVgsFgIDA8tsZ7FYyiy/cF2z2Vxm3by8is+6BQYG74Xq1KnD/v37ndGUiIhUISZTxV9X4+jRozz++OP07NmTHj164OX1e9zl5+cTFBSE2WwmPz+/zPLAwMAyyy+3blBQ5R5PaNip5ueee87+7eLEiRPUrFnTqKZERKSKMmIkxKlTpxg0aBBTp06lffv2ADRr1ozMzEzatWvH5s2bufPOO2nRogVLliyhqKgIq9VKTk4OYWFhhIeHk5aWRosWLdi8eTNt2rTBbDbj6+vLgQMHqFevHunp6dfe4KqtW7faf/b396d58+Z4e3tf1bYaXCXuQIOrxF0YObhq7/FzFd4m7KYal31/1qxZfPTRR4SGhtqXTZo0iVmzZlFcXExoaCizZs3C29ublJQU1q1bh81mY9iwYXTu3JmCggLi4+M5efIkvr6+JCQkcOONN5Kdnc2cOXMoKSkhMjKS0aNHV7jvYGDwnjlzhvT0dM6fP4/NZuPEiRMMGzbsqrZV8Io7UPCKuzAyeH84XlDhbZrcVN2AnjiPYR9nXFwcDRo0YO/evfj7+1O9etX+oERExPE88SEJhg6ueuGFF2jYsCFvvfUWZ8+eNbIpERGpgjxwqmZjHwtYVFREQUEBJpOJc+cqfh5fRETcnDskaQUZVvEOGDCAVatWcffdd9OpU6cyF7lFRETAMx+SYFjFW1RUxNChQwHo2rVrmRuPRUREQNd4HSolJcX+s0JXREQuRdd4HchqtdKrVy8aNmxonzEkISHBqOZERKQqcockrSDDgnfs2LFG7VpERNyEO1yzrSiHT6CRmJjI008/Dfw6VWStWrUqvA9NoCHuQBNoiLswcgKNn34uqvA29Wv6G9AT53H4Nd6vvvrK/rOqXhERuRyjHpJwLXP495gLC2iDZqMUERE34QY5WmEOD17TBV9HTO7w1URERAzjiTHh8Gu8bdq0oUmTJthsNvbt22f/2WQykZycfFX70DVecQe6xivuwshrvIdyrRXepu71fgb0xHkc/nFu3LjR0bsUERE35YkVr8ODt06dOo7epYiIuCkPzF1jH5IgIiJyOap4RUREnMgTJ9BQ8IqIiOt4Xu4qeEVExHU8MHcVvCIi4jq6xisiIuJEusYrIiLiTJ6XuwpeERFxHQ/MXQWviIi4jq7xioiIOJGu8YqIiDiRJ1a8Xq7ugIiIiCdR8IqIiDiRTjWLiIjLeOKpZgWviIi4jAZXiYiIOJEqXhERESfywNxV8IqIiAt5YPIqeEVExGV0jVdERMSJdI1XRETEiTwwdxW8IiLiQgYkb2lpKdOnT+f777/Hz8+PWbNmUb9+fcc3VEmauUpERFzGVIn/Xcknn3yC1Wpl3bp1jBkzhnnz5jnhSK6eKl4REXEZI67xZmVl0aFDBwBatWrFrl27HN/IH3BNBm+Anyee9RcR8TzVDEghi8WC2Wy2/+7t7c358+fx8bk2Ik+nmkVExK2YzWby8/Ptv5eWll4zoQsKXhERcTPh4eFs3rwZgOzsbMLCwlzco7JMNpvN5upOiIiIOMpvo5r37t2LzWZjzpw5NGrUyNXdslPwioiIOJFONYuIiDiRgldERMSJFLxVwKFDhwgPDycmJsb+WrZsmcP2v3nzZsaPH39Rm7fffnuZ+9+SkpJYunSpw9oVz+GKv+G4uDiWL19u/z0/P5/OnTuzZ88eh7UrUhnXzvhquazGjRuzevVqp7ZpNpuZMGECGzZswM/Pz6lti/tx9t/w9OnTeeSRR7j33ntp3Lgx8+fP59FHH+XWW291Wh9ELkXBW4VlZmaycOFCfH196devH9WqVWPNmjX291966SV++OEHkpOTWbx4MQB33303X375JTk5OUycOJHq1atTvXp1rrvuuov2X79+fSIiIli8eDHx8fFl3tu4cSOrVq3Cz8+PBg0a8MILL+Dr62vsAYvbMfJvODg4mClTpjB58mSee+45Dh48yIwZMzh69ChTpkyhqKgIf39/Zs6cSXBwMM8++ywWi4XCwkLGjRtHu3btnPpZiOfQqeYqYt++fWVO0x0/fhyAoqIi1q5dS69evfjxxx9Zvnw5q1evpmHDhqSnp5e7v5deeom4uDhWrlxJ69aty11v1KhRfPnll3z99df2Zbm5uSxdupRVq1aRlJREYGAg69atc9zBiltyxd/wvffeS8OGDRk/fjzz5s3DZDIxf/58YmJiWL16NYMHD2bhwoUcOHCAU6dO8eqrr5KQkEBhYaEhn4EIqOKtMi51mu7HH3+kYcOG9t9r1qxJfHw8AQEB7N+/n1atWl20n9/uHvvhhx9o0aIF8OvN5vv3779ku35+fsydO5cxY8bQr18/AA4ePEjjxo3tU7Ldcccdl/0HUgRc9zfcq1cvCgsLuemmmwDYu3cvr732Gm+88QY2mw1fX1+aNGnCgAEDeO655zh//jwxMTEOOWaRS1HwVnFeXr+etMjLy+Pll1/m3//+NwBPPvkkNpsNf39/Tp48CcDhw4c5e/YsAKGhoWzfvp2OHTtecQLx22+/ne7du/P6668THR1N3bp1ycnJ4dy5c9SoUYOtW7eW+cdTpCKc8Td8odDQUAYNGkR4eDg5OTls27aN77//nvz8fJYvX86JEyeIiorinnvuceyBivw/Ba+bMJvNhIeH8/DDD1OjRg2CgoI4ceIEPXv2JDAwkL59+9KoUSPq1q0LwLRp0xg9ejQrVqwgODgYf3//y+5/+PDhfP7558Cv185iY2N5/PHH8fLyIiQkhLFjxxp+jOLejP4b/k18fDzTp0+nqKiIwsJCJk2aRIMGDXjllVd477338PX1JS4uzshDFQ+nmatEREScSIOrREREnEjBKyIi4kQKXhERESdS8IqIiDiRgldERMSJdDuRuL1Dhw7RpUsXGjVqhMlkori4mFq1ajF37lzGjRvHsWPHqFGjBgAWi4V69eqxcOFCbrjhBmJiYsq8D3DDDTewYsUKli5dSnJyMjfccAM2m43S0lKioqJ47LHHAEhNTWXr1q3MmzcPgOzsbBYvXkxubi6lpaVEREQwfvx4CgoKGDhwIACnTp2ytwGwcuVK4uLiKt0HEbn2KHjFI9SqVYv333/f/vu8efN48cUXAZg1a5Z9Xt7S0lLi4uJ46623GDdu3EXv/6+oqChiY2MBOH36NE888QT+/v707du3zHp79uxh5MiRvPLKK7Rs2ZLz588zc+ZMpkyZwoIFC+x9++3pT7/t8zeO6IOIXBt0qlk8Urt27fjhhx8uWn7u3Dlyc3Mv+dCIKwkODmbEiBGsXbv2ovdWrFjBI488QsuWLQHw8fFh3Lhx3HfffRXvfCX7ICLXBlW84nGKi4v5+OOPadWqFT/++COTJ0+mevXqnD59muuuu45u3brZT/0CTJ48ucxp3i5dujBixIhL7jssLOyScwb/5z//4f777y+zzGw207lz56vqsyP6ICLXBgWveITfph4EsFqttGjRgjFjxhAbG2s/jfvNN98QFxfH/fffX+b5w5c7zfu/TCYT1apVu+Tyq53S8FIc0QcRuTYoeMUj/O813ksJDw8nJiaGMWPG8Pe//x0fn4r/5/H999/TqFGji5Y3b96cb7/9lk6dOtmXWSwWxo4dy8svv1wm6P+o8vogItcGXeMVucCTTz5Jfn5+pZ4vfOLECV599VUGDBhw0XsDBw4kKSmJnTt3Ar+e7p43bx5ms9mhoXu5PojItUEVr8gF/Pz8GDVqFHPmzOGhhx4CLr6+CtifK5ucnMwnn3yCyWTCZrPx6KOP8uCDD16036ZNm7JgwQJmz55NQUEBxcXF3HXXXUyePPmq+uWIPojItUFPJxIREXEinWoWERFxIgWviIiIEyl4RUREnEjBKyIi4kQKXhERESdS8IqIiDiRgldERMSJFLwiIiJO9H+d/WaFoPsJwQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 576x360 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# KNNeighbors evaluation\n",
    "\n",
    "print(classification_report(y_test, yp_KNN, target_names = labels))\n",
    "f, ax = plt.subplots(figsize=(8,5))\n",
    "sns.heatmap(confusion_matrix(y_test, yp_KNN), annot=True, fmt=\".0f\", ax=ax, cmap = 'Blues')\n",
    "\n",
    "ax.xaxis.set_ticklabels(labels)\n",
    "ax.yaxis.set_ticklabels(labels)\n",
    "\n",
    "plt.title('KNNeighbors Matrix')\n",
    "plt.xlabel('PREDICTED')\n",
    "plt.ylabel('ACTUAL')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> Dalam model **KNNeighbors** ini juga terjadi hal yang sama. Walaupun ketika kita bandingkan, kedua model sama - sama memiliki 98% *accuracy*, tetapi jika di perhatikan lagi pada ***confusion matrix*** di atas, **Logistic Regression** masih memiliki *precision* yang lebih baik daripada model **KNNeighbors**."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## OVERSAMPLING - SMOTE"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> Pada dasarnya, **SMOTE** atau *Synthetic Minority Oversampling Technique*, merupakan teknik oversampling dengan cara menggunakan synthetic data yang nantinya data tersebut akan di oversampling sehingga jumlah data minoritas sama dengan data mayoritas. **SMOTE** menggunakan teknik yang sama dengan KNN - *euclidean distance based* - namun karena alasan sebelumnya, maka kita hanya akan melakukan oversampling pada model **Logistic Regression** saja. Oversampling ini di lakukan karena data imbalanced, sehingga menyulitkan model dalam proses prediksi."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Oversampling dengan SMOTE\n",
    "\n",
    "X_train_sm, y_train_sm = SMOTE(random_state=101).fit_sample(X_train, y_train)\n",
    "\n",
    "# Model best porameter - oversampled\n",
    "\n",
    "LRG_sm = LogisticRegression(solver = 'lbfgs', penalty = 'l2', max_iter = 200).fit(X_train_sm, y_train_sm)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Predict dengan Logistic Regression - oversampled\n",
    "\n",
    "yp_LRG_sm = LRG_sm.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "    Fraud No       0.99      0.82      0.90    124548\n",
      "   Fraud Yes       0.04      0.42      0.07      2222\n",
      "\n",
      "    accuracy                           0.81    126770\n",
      "   macro avg       0.51      0.62      0.48    126770\n",
      "weighted avg       0.97      0.81      0.88    126770\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x360 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Model evaluation\n",
    "\n",
    "print(classification_report(y_test, yp_LRG_sm, target_names = labels))\n",
    "f, ax = plt.subplots(figsize=(8,5))\n",
    "sns.heatmap(confusion_matrix(y_test, yp_LRG_sm), annot=True, fmt=\".0f\", ax=ax, cmap = 'Blues')\n",
    "\n",
    "ax.xaxis.set_ticklabels(labels)\n",
    "ax.yaxis.set_ticklabels(labels)\n",
    "\n",
    "plt.title('Logistic Regression Oversampled')\n",
    "plt.xlabel('PREDICTED')\n",
    "plt.ylabel('ACTUAL')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> Setelah dilakukan *oversampling*, memang jelas tampak bahwa model menjadi lebih baik daripada sebelumnya. Dan pada akhirnya, model **Logistic Regression** dengan *scaling* menggunakan **Robust Scaler** dan teknik *oversampling* dengan **SMOTE**, menjadi model yang paling baik untuk digunakan untuk memprediksi dalam dataset ini. \n",
    "\n",
    "> Pada intinya, model ini memiliki 81% *accuracy*, serta 99% *precision* pada kategori yang **tidak fraud**. Terlebih lagi, model ini juga memiliki 82% *recall* pada kategori yang **tidak fraud**, artinya model dapat memprediksi dengan benar sekitar 82% pada kategori ini. Walaupun pada kategori yang **fraud** masih hanya sekitar 42%, tetapi bagaimanapun juga versi model ini sudah lebih baik dan karena lebih seimbang di bandingkan dengan versi sebelum di lakukan *balancing*."
   ]
  }
 ],
 "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": "3.7.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}