Credit Card Fraud Classification.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Dataset From: https://www.kaggle.com/mlg-ulb/creditcardfraud"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Importing Libraries\n",
    "import numpy as np \n",
    "import pandas as pd \n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.metrics import accuracy_score, precision_score, recall_score, confusion_matrix "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fraudulent transactions: 0.17%"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Exploratory Data Analysis\n",
    "data = pd.read_csv('creditcard.csv')\n",
    "plt.bar(['Valid','Fraud'],list(data['Class'].value_counts()))\n",
    "print(\"Fraudulent transactions: \", end='')\n",
    "frauds = data['Class'].value_counts()[1]/sum(data['Class'].value_counts())\n",
    "print(round(frauds*100,2), end='%')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAArUAAALICAYAAABhHAheAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjAsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy8GearUAAAgAElEQVR4nOzdf7SddXnn/fcnSGIoIIqoRJzGKtb6C5qm1D4urIXpJK20tvXXafuoaXRi24fVpj5lbNZyOV2OrtSxxU6ttE+mPUbEHiKRBIsp0VKPggONMSKKaEF0NGJFoxEyKj+v54+zT9kez07Avc++7/vwfrHudfb+7u/9PVc2LtfFlet7f1NVSJIkSV22pOkAJEmSpGGZ1EqSJKnzTGolSZLUeSa1kiRJ6jyTWkmSJHXew5oOYCEs/8lzh36kwxve+pqh4/jREx4+9BoAz3vSSUOv8emv3D70Gh/64jeHXgPg2ac8Yug1vvjt7wy9xuOPG/7fz1333Tf0GgDfu+feoddY9bhHDr3Grpu+NvQaz3jMcUOvAbDsqOH/m/v4ZUcPvcZN37pj6DUAHrVs2dBr3HTw0NBrPGLZ8P+3/8hlS4deA+Dr37tz6DW+evtdQ6/xuONG8+c5funw3+2Pn3j80Gt8YQT/O7nz3tH8f1tbZETrvPj0FaNa6oc2ihxnWN/9xF81/j3Mx0qtJEmSOs+kVpIkSZ23KNsPJEmSFqVYjxzEb0aSJEmdZ1IrSZKkzrP9QJIkqSvSygcPtIKVWkmSJHXeglVqk5wIXNl7+zjgXuDrwJOBC6vq9xbqd0uSJC1KbhQbaMGS2qo6AJwOkORPgENV9WcL9fskSZL00DX2dD/J85Jc3nv9J0nemeQDSb6Y5NeT/Pckn0pyRZKje/N+KsmHk3w8ye4kJ487bkmSJLVXG2rYTwKeD7wAuAj4UFU9E/gu8PxeYvs24EVV9VPAJPCmuYsk2ZBkb5K993zjhvFFL0mSNC5J81dLteHpB/9YVXcn+RRwFHBFb/xTwErgx4FnAB/MzBd5FPDVuYtU1RZgC7TjXGRJkiSNTxuS2jsBquq+JHdX1WxCeh8z8QW4oap+tqkAJUmS1G5taD84ks8BJyX5WYAkRyd5esMxSZIkjV+WNH+1VHsj66mqu4AXAW9O8kngOuD/ajYqSZIktclY2g+q6k/6Xk8D03PHe++PHXDPdcBzFzRISZKktmvxRq2mtb5SK0mSJB2JSa0kSZI6rw1PP5AkSdID0eKNWk3zm5EkSVLn5f7Hwi4eb5m+Zeg/1Ov/8Pyh49g6uWnoNSRJeih76qOPH3qNA//nrhFEAmc99cTGd2kt/9k/bjxx++41f9r49zAfK7WSJEnqPJNaSZIkdZ4bxSRJkrrCjWID+c1IkiSp86zUSpIkdYUnig1kpVaSJEmdZ1IrSZKkzmtFUptkOsmaOWMbk1yQ5IokB5Nc3lR8kiRJrZAlzV8t1ZbIpoCJOWMTvfG3AC8be0SSJEnqjLZsFNsOvDHJsqq6M8lKYAVwdVVVkuc1GZwkSVIruFFsoFZUaqvqALAHWNsbmgC21YM4wzfJhiR7k+y99h+mFiJMSZIktVQrktqe/haE2daDB6yqtlTV6qpa/exf/o2RBydJkqT2akv7AcBO4Pwkq4DlVbWv6YAkSZJapcUbtZrWmm+mqg4B08AkD7JKK0mSpIe21iS1PVPAacDFswNJrgIuAc5Osn/uo78kSZKkNrUfUFU7gMwZO7OhcCRJktrF9oOB/GYkSZLUea2q1EqSJOkwlvic2kGs1EqSJGmkkqxN8rkkNyf543k+/9EkVya5Psl0klOG/Z0mtZIkSRqZJEcBbwd+EXga8BtJnjZn2p8BF1bVs4A3AJuH/b2Lsv3gR094+NBrbJ3cNPQa69YP/e8HGE0skiR10Y+ffNzQa2y59gsjiATO4sSRrDOUbmwUOwO4uapuAUhyMfAC4DN9c54G/GHv9YeYOa9gKJ34ZiRJktQOSTYk2dt3bZgz5fHAl/ve7++N9fsk8MLe618Djksy1H81LMpKrSRJ0qKU5jeKVdUWYMthpswXZM15/0fAXyVZB3wE+ApwzzBxmdRKkiRplPYDT+h7fwpwa/+EqroV+HWAJMcCL6yqbw/zS20/kCRJ0ih9DDg1yROTLAUmgPf1T0jy6OTfG4Q3AZPD/lKTWkmSpK7IkuavI6iqe4Bzgd3AjcB7quqGJG9I8iu9ac8DPpfkX4HHAm8a9qux/UCSJEkjVVW7gF1zxl7f93o7sH2Uv9OkVpIkqStasFGsrVrRftA7SWLNnLGNSXYluSbJDb0TJ17aVIySJElqr7ZUaqeYaSLe3Tc2AbwWuLWqbkqyAvh4kt1VdbCJICVJktRObUlqtwNvTLKsqu5MshJYAXykqgpmHv2Q5DbgJMCkVpIkPfR040SxRrTim6mqA8AeYG1vaALYNpvQAiQ5A1gKfH6+NfpPt/in91600CFLkiSpRdpSqYX7WxAu6/1cP/tBkpOBdwGvqKr75ru5/3SL91x369xTKyRJkrrPjWIDtaJS27MTODvJKmB5Ve0DSHI88H7gdVV1bZMBSpIkqZ1ak9RW1SFgmpkTJaYAeqdQ7AAurKpLmotOkiRJbdam9gOYSWYvZab9AOAlwHOBE5Os642tq6rrGohNkiSpWW4UG6hVSW1V7QDS9/4iwF1fkiRJOqxWJbWSJEk6DDeKDWQNW5IkSZ1nUitJkqTOs/1AkiSpK9woNlD6Du1aNG674+6h/1Afuvm2UYQyEuvWbx56ja2Tm0YQiSRJ4/WURx0/9Br3jSjX+amVxzfe0Lr8l/5H44nbd3f9QePfw3xM9yVJktR5th9IkiR1hU8/GMhKrSRJkjrPSq0kSVJXuFFsIL8ZSZIkdZ5JrSRJkjrP9gNJkqSusP1goFZ8M0mmk6yZM7YxyTuSfDzJdUluSPI7TcUoSZKk9mpLpXYKmAB2941NAK8Frq2qO5McC3w6yfuq6tYmgpQkSWqUj/QaqBWVWmA7cE6SZQBJVgIrgI9U1Z29OctoT7ySJElqkVYkiVV1ANgDrO0NTQDbqqqSPCHJ9cCXgTcPqtIm2ZBkb5K9F77jb8cTuCRJklqhLe0HcH8LwmW9n+sBqurLwLOSrAB2JtleVV+be3NVbQG2ANx2x92Nn4ssSZI0cm4UG6hN38xO4Owkq4DlVbWv/8NehfYG4MwmgpMkSVJ7tSaprapDwDQwyUzVliSnJFnee/1I4DnA55qKUZIkqVFJ81dLtan9AGaS2UuZaT8A+Angz5MUEODPqupTTQUnSZKkdmpVUltVO5hJXmfffxB4VnMRSZIkqQtaldRKkiTpMNwoNpDfjCRJkjrPSq0kSVJXtHijVtOs1EqSJKnzFmWl9tNfub3pEEZq6+SmoddYt35zK+KQJOnB+PCXvjH0Gt+5674RRAI/tfL4kayjhbEok1pJkqTFKLYfDGT7gSRJkjrPSq0kSVJHWKkdzEqtJEmSOs+kVpIkSZ1n+4EkSVJX2H0wkJVaSZIkdV4rktok00nWzBnbmOSC3uvjk3wlyV81E6EkSZLarBVJLTAFTMwZm+iNA/w34MNjjUiSJKllkjR+tVVbktrtwDlJlgEkWQmsAK5O8lPAY4EPNBadJEmSWq0VSW1VHQD2AGt7QxPANmbaof8cOO9IayTZkGRvkr2Xv+edCxarJElSU5qu0ra5Utumpx/MtiBc1vu5Hvg9YFdVfflIX2JVbQG2APzzZw/UwoYqSZKkNmlTUrsTOD/JKmB5Ve1L8v8CZyb5PeBYYGmSQ1X1x41GKkmSpFZpTVJbVYeSTAOT9DaIVdVvzX6eZB2w2oRWkiQ9VLX5r/+b1oqe2j5TwGnAxU0HIkmSpO5oTaUWoKp2MOCsjKraCmwdZzySJEltYqV2sLZVaiVJkqQHzaRWkiRJndeq9gNJkiQdht0HA1mplSRJUuctykrth774zaHXeMZjjxlBJO2xdXLT0GusW795BJGMJhZJ0kPDi5/5+KHX+Oy/3T6CSNrBjWKDWamVJElS55nUSpIkqfMWZfuBJEnSYmT7wWBWaiVJktR5VmolSZI6wkrtYFZqJUmS1HkmtZIkSeq8ViS1SaaTrJkztjHJBUnuTXJd73pfUzFKkiQ1LUnjV1u1IqkFpoCJOWMTvfHvVtXpvetXxh+aJEmS2q4tSe124JwkywCSrARWAFc3GJMkSZI6ohVJbVUdAPYAa3tDE8C2qirg4Un2Jrk2ya8OWiPJht68vft2XTyGqCVJksYsLbhaqhVJbU9/C8Js6wHAf6iq1cBvAn+R5Enz3VxVW6pqdVWtXvVLczsZJEmStJi16Tm1O4Hzk6wCllfVPoCqurX385Yk08BPAp9vLEpJkqSGtHmjVtNaU6mtqkPANDBJr0qb5JF9fbaPBp4DfKapGCVJktRObarUwkwyeyn3tyH8BPD/JbmPmQT8T6vKpFaSJEnfp1VJbVXtoK8Fuar+F/DM5iKSJElqD9sPBmtN+4EkSZL0w2pVpVaSJEmDWakdzEqtJEmSOs+kVpIkSZ23KNsPnn3KI4Ze49Ddd48gksVl6+Smkayzbv3modcYVSySpHb784/cMvQaP/fEE0YQSUvYfTCQlVpJkiR13qKs1EqSJC1GbhQbzEqtJEmSOs+kVpIkSZ1n+4EkSVJH2H4wmJVaSZIkdZ6VWkmSpI6wUjtYKyq1SaaTrJkztjHJBUn+Q5IPJLkxyWeSrGwmSkmSJLVVK5JaYAqYmDM20Ru/EHhLVf0EcAZw25hjkyRJUsu1pf1gO/DGJMuq6s5eNXYF8E3gYVX1QYCqOtRciJIkSc2y/WCwVlRqq+oAsAdY2xuaALYBpwIHk1ya5BNJ3pLkqPnWSLIhyd4ke6+45MLxBC5JkqRWaEVS29PfgjDbevAw4Ezgj4CfBn4MWDffzVW1papWV9XqtS9++cJHK0mSNG5pwdVSbUpqdwJnJ1kFLK+qfcB+4BNVdUtV3dObs6rJICVJktQ+rUlqe/2y08AkM1VagI8Bj0xyUu/9WcBnxh+dJEmS2qwtG8VmTQGX0mtDqKp7k/wRcGVmOqM/DvzPBuOTJElqjBvFBmtVUltVO5jTrdF78sGzmolIkiRJXdCa9gNJkiTph9WqSq0kSZIGs/1gMCu1kiRJ6jwrtZIkSR1hpXawRZnUfvHb3xl6jUcfc/QIItF8tk5uGnqNdes3tyIOSdLCuvEr3x56jd985ooRRKK2s/1AkiRJnbcoK7WSJEmLkt0HA1mplSRJUudZqZUkSeoIN4oNZqVWkiRJnWdSK0mSpM6z/UCSJKkjbD8YrBWV2iTTSdbMGduY5MYk1/Vd30vyq03FKUmSpHZqS6V2CpgAdveNTQAbquoqgCSPAm4GPjD+8CRJkppnpXawVlRqge3AOUmWASRZCawAru6b8yLgH6tq+OPCJEmStKi0IqmtqgPAHmBtb2gC2FZV1TdtgpmK7rySbEiyN8neqy/7+4ULVpIkSa3TlvYDuL8F4bLez/WzHyQ5GXgm39+e8H2qaguwBeDtH/1iDZonSZLUVbYfDNaKSm3PTuDsJKuA5VW1r++zlwA7quruZkKTJElSm7Umqa2qQ8A0MMkPthn8xjxjkiRJDy1pwdVSrUlqe6aA04CLZwd6m8aeAHy4mZAkSZLUdm3qqaWqdjDnvwGq6ovA4xsJSJIkSZ3QqqRWkiRJg7lRbLC2tR9IkiRJD5qVWkmSpI6wUjuYlVpJkiR13qKs1D7+uIcPvcad9947gki0ULZObhp6jXXrN7ciDknSYOf9/JOHXuOz37x9BJHA6iceP5J1HgqSrAX+B3AU8LdV9afzzHkJ8CdAAZ+sqt8c5ncuyqRWkiRpMepC90GSo4C3A78A7Ac+luR9VfWZvjmnApuA51TVt5I8Ztjfa/uBJEmSRukM4OaquqWq7mLm/IEXzJnzn4G3V9W3AKrqtmF/qUmtJEmSHrAkG5Ls7bs2zJnyeODLfe/384NnDjwFeEqSjya5tteuMBTbDyRJkjqiDU8/qKotwJbDTJkvyJrz/mHAqcDzgFOAq5I8o6oO/rBxWamVJEnSKO0HntD3/hTg1nnmXFZVd1fVF4DPMZPk/tBMaiVJkjoiaf56AD4GnJrkiUmWAhPA++bM2Qn8/MyfKY9mph3hlmG+G5NaSZIkjUxV3QOcC+wGbgTeU1U3JHlDkl/pTdsNHEjyGeBDwHlVdWCY39uKntok08DmqtrdN7aRmaz9EPB8ZhLwDwJ/UFVz+zIkSZLUElW1C9g1Z+z1fa8LeE3vGom2VGqnmClN95sAtgHPAZ4FPAP4aeDnxhuaJElSOyRp/GqrtiS124FzkiwDSLISWAHcBTwcWAosA44GvtZMiJIkSWqrViS1vR6KPcDsM8omgG1VdQ0zfRZf7V27q+rG+dbof2baB7a/axxhS5IkjVXTm8RaXKhtR1Lb09+CMAFMJXky8BPMPAri8cBZSZ47381VtaWqVlfV6v/0opeNJWBJkiS1Q5uS2p3A2UlWAcurah/wa8C1VXWoqg4B/wg8u8kgJUmS1D6tSWp7Ses0MMlM1RbgS8DPJXlYkqOZ2SQ2b/uBJEnSYrdkSRq/2qo1SW3PFHAacHHv/Xbg88CngE8Cn6yqf2goNkmSJLVUK55TO6uqdtB3XnBV3Qu8urmIJEmS2qPNG7Wa1rZKrSRJkvSgmdRKkiSp81rVfiBJkqTB2nyiV9Os1EqSJKnzFmWl9q777ms6BHXA1slNQ6+xbv3mEUQymlgkaTG64+67h17j5B95+AgiaQcLtYNZqZUkSVLnmdRKkiSp8xZl+4EkSdJi5EaxwazUSpIkqfNMaiVJktR5th9IkiR1hO0Hg1mplSRJUue1olKbZBrYXFW7+8Y2Ak8B7gCe3xv+b1W1bfwRSpIkNc9C7WBtqdROARNzxiaArwGrgNOBnwHOS3L8mGOTJElSy7Ulqd0OnJNkGUCSlcAK4DvAh6vqnqr6P8AngbVNBSlJkqR2akVSW1UHgD3cn7BOANuYSWJ/MckxSR4N/DzwhPnWSLIhyd4ke//pvReNI2xJkqSxStL41Vat6KntmW1BuKz3c31V7Uvy08D/Ar4OXAPcM9/NVbUF2ALwnuturbFELEmSpFZoRaW2ZydwdpJVwPKq2gdQVW+qqtOr6heAADc1GaQkSVJTkuavtmpNUltVh4BpYJKZqi1JjkpyYu/1s4BnAR9oKkZJkiS1U5vaD2Ammb2U+5+EcDRwVa9/43bg/66qedsPJEmS9NDVqqS2qnYw02Iw+/57wNOai0iSJKk92rxRq2mtaT+QJEmSflitqtRKkiRpMAu1g1mplSRJUueZ1EqSJKnzFmX7wffuuXfoNZYdZb6vI9s6uWkk66xbv3noNUYViyS1yXFHHz30Go86ZukIImkHN4oNZuYmSZKkzluUlVpJkqTFyELtYFZqJUmS1HkmtZIkSeo82w8kSZI6wo1ig1mplSRJUudZqZUkSeoIC7WDjbVSm2Q6yZo5YxuTXJDkiiQHk1w+5/MnJvmXJDcl2ZZk8TxsTpIkSSMx7vaDKWBizthEb/wtwMvmuefNwFur6lTgW8ArFzRCSZIkdc64k9rtwDlJlgEkWQmsAK6uqiuBO/onZ6Yb+qzefQDvBH51XMFKkiS1SZLGr7Yaa1JbVQeAPcDa3tAEsK2qasAtJwIHq+qe3vv9wOPnm5hkQ5K9SfZ+6NJ3jzJsSZIktVwTTz/ob0GYbT0YZL7/HJg3Aa6qLVW1uqpW//yv/9aQIUqSJKlLmnj6wU7g/CSrgOVVte8wc78BnJDkYb1q7SnAreMIUpIkqW1a/Lf/jRt7pbaqDgHTwCSHr9LSa0v4EPCi3tArgMsWMj5JkiR1T1OHL0wBpwEXzw4kuQq4BDg7yf6+R3+9FnhNkpuZ6bH9u3EHK0mS1AZNbxJr80axRg5fqKodzOmXraozB8y9BThjHHFJkiSpmzwmV5IkSZ3nMbmSJEkd0eK//W+clVpJkiR1npVaSZKkjmjzRq2mLcqkdtXjHjn0Gjd8/dsjiER6YLZObhp6jXXrN7ciDkkapa9/93tDr3H1l781gkjg9P9w3EjW0cKw/UCSJEmdtygrtZIkSYuR7QeDWamVJElS51mplSRJ6ggLtYNZqZUkSVLnmdRKkiSp82w/kCRJ6gg3ig021kptkukka+aMbUxyQZIrkhxMcvmcz89NcnOSSvLoccYrSZKkbhh3pXYKmAB2941NAOcBS4FjgFfPueejwOXA9BjikyRJai0LtYONu6d2O3BOkmUASVYCK4Crq+pK4I65N1TVJ6rqi2OMUZIkSR0z1qS2qg4Ae4C1vaEJYFtV1bBrJ9mQZG+SvZe8e3LY5SRJktQhTWwUm21BuKz3c/0oFq2qLcAWgE/vPzR0kixJktQ2bhQbrIlHeu0Ezk6yClheVfsaiEGSJEmLyNgrtVV1KMk0MMlM1VaSJEkPgIXawZo6fGEKOA24eHYgyVXAJcxUcffPPvorye8n2Q+cAlyf5G+bCFiSJEnt1cjhC1W1A8icsTMHzP1L4C/HEZckSZK6yRPFJEmSOmKJ/QcDNdV+IEmSJI2MSa0kSZI6z/YDSZKkjrD7YLBFmdTuuulrQ6/xoyc8fASRSOOzdXLT0GusW7+5FXFI0qxRHDbwwqedPIJI1HaLMqmVJElajDxRbDB7aiVJktR5JrWSJEnqPNsPJEmSOmKJ3QcDWamVJElS51mplSRJ6gg3ig1mpVaSJEmdN9akNsl0kjVzxjYmuSDJFUkOJrl8zufvTvK5JJ9OMpnk6HHGLEmSpPYbd6V2CpiYMzbRG38L8LJ57nk38FTgmcBy4FULGaAkSVJbJc1fbTXupHY7cE6SZQBJVgIrgKur6krgjrk3VNWu6gH2AKeML1xJkiR1wViT2qo6wExiurY3NAFs6yWsh9VrO3gZcMWAzzck2Ztk779cfvGoQpYkSWqNtOCftmpio1h/C8Js68EDcQHwkaq6ar4Pq2pLVa2uqtU/c87cDgdJkiQtZk0ktTuBs5OsApZX1b4j3ZDkvwInAa9Z6OAkSZLUPWN/Tm1VHUoyDUzyAKq0SV4FrAHOrqr7Fjg8SZKk1vJEscGaek7tFHAa8O/Nr0muAi5hpoq7v+/RX38DPBa4Jsl1SV4/9mglSZLUao2cKFZVO+D7O42r6swBcz31TJIkCU8UOxxPFJMkSVLnmdRKkiSp8/yrfUmSpI6w+2AwK7WSJEnqvEVZqX3GY44beo077rp7BJFI3bJ1ctNI1lm3fvPQa4wqFknddu+RDx09ok/fdnAEkcDTH/8jI1lHC2NRJrWSmjOKhFaSNL8l9h8MZPuBJEmSOs9KrSRJUkdYqB3MSq0kSZI6z6RWkiRJnWf7gSRJUkd4TO5gVmolSZLUeWNNapNMJ1kzZ2xjkguSXJHkYJLL53z+d0k+meT6JNuTHDvOmCVJktoiaf5qq3FXaqeAiTljE73xtwAvm+eeP6yq06rqWcCXgHMXNkRJkiR1zbiT2u3AOUmWASRZCawArq6qK4E75t5QVbf35gZYDgx/tIgkSZIWlbEmtVV1ANgDrO0NTQDbqg5/Bl6SdwD/BjwVeNuAORuS7E2y9x8vuXCEUUuSJLXDkqTxq62a2CjW34Iw23pwWFX128xUdG8EXjpgzpaqWl1Vq3/xxS8fVaySJEnqgCaS2p3A2UlWAcurat8Duamq7gW2AS9cyOAkSZLaKi242mrsSW1VHQKmgUmOUKXNjCfPvgZ+GfjsQscoSZKkbmnq8IUp4FL6noSQ5CpmemaPTbIfeCXwQeCdSY5n5j8OPgn87vjDlSRJUps1ktRW1Q7mVLCr6swB05+z8BFJkiS1nyeKDeaJYpIkSeq8ptoPJEmS9CAtsVA7kJVaSZIkdZ5JrSRJkjpvUbYfLDtq+Fz9B87rlfSAbJ3cNJJ11q3fPPQao4pFUnPuuve+odf42SecOIJI2sGNYoNZqZUkSdJIJVmb5HNJbk7yx/N8/jtJPpXkuiRXJ3nasL/TpFaSJKkjkuavI8eYo4C3A78IPA34jXmS1r+vqmdW1enAfwfOH/a7MamVJEnSKJ0B3FxVt1TVXcDFwAv6J1TV7X1vfwSoYX/pouyplSRJUmMeD3y57/1+4GfmTkry/wCvAZYCZw37S63USpIkdUSSNlwbkuztuzbMDXOe0H+gEltVb6+qJwGvBV437HdjpVaSJEkPWFVtAbYcZsp+4Al9708Bbj3M/IuBvx42Liu1kiRJGqWPAacmeWKSpcAE8L7+CUlO7Xv7fOCmYX/pWJPaJNNJ1swZ25jkgiRXJDmY5PIB974tyaHxRCpJktQ+S9L8dSRVdQ9wLrAbuBF4T1XdkOQNSX6lN+3cJDckuY6ZvtpXDPvdjLv9YIqZbH1339gEcB4zTcLHAK+ee1OS1cAJ4whQkiRJw6mqXcCuOWOv73v9B6P+neNuP9gOnJNkGUCSlcAK4OqqupJ5DvLqPevsLcB/GV+YkiRJ7dOCTWJNfwUDjTWpraoDwB5gbW9oAthWVYd7Ntm5wPuq6quHW7t/J97l77lwNAFLkiSpE5p4+sFsC8JlvZ/rB01MsgJ4MfC8Iy3avxPvys9+Y+gH+EqSJKk7mnj6wU7g7CSrgOVVte8wc38SeDJwc5IvAsckuXkMMUqSJLVOWnC11dgrtVV1KMk0MMlM1fZwc98PPG72fZJDVfXkhY1QkiRJXdPU4QtTwKXMtB8AkOQq4KnAsUn2A6+sqt0D7pckSXrIWdLijVpNaySpraodzKlgV9WZD+C+YxcsKEmSJHWWJ4pJkiSp85pqP5AkSdKDZPfBYFZqJUmS1HlWaiVJkjqizSd6NW1RJrXHLzt6JGvccvDQCKKR9MPYOrlpJOusW7956DVGFYukB+/hRw3/l8qfuPVbI4gEnvLYY0ayjhaG7QcDmNBK3WdCK0kPHYuyUitJkrQY2X0wmJVaSZIkdZ6VWkmSpI7wRLHBrNRKkiSp80xqJUmS1Hm2H0iSJHWE3QeDjbVSm2Q6yZo5YxuTXJDkiiQHk1w+5/OtSb6Q5Lredfo4Y5YkSVL7jbtSOwVMALv7xiaA8xfx3gwAACAASURBVIClwDHAq+e577yq2r7w4UmSJLWXJ4oNNu6e2u3AOUmWASRZCawArq6qK4E7xhyPJEmSFoGxJrVVdQDYA6ztDU0A26qqjnDrm5Jcn+StswmxJEmSNKuJpx/MtiDQ+zl1hPmbgKcCPw08CnjtfJOSbEiyN8neHVNbRxSqJElSeyxpwdVWTTz9YCdwfpJVwPKq2ne4yVX11d7LO5O8A/ijAfO2AFsAPvaFbx+p8itJkqRFZOwJd1UdAqaBSY5cpSXJyb2fAX4V+PRCxidJkqTuaeo5tVPApdzfhkCSq5hpMzg2yX7glVW1G3h3kpOAANcBv9NAvJIkSY3z6QeDNZLUVtUOZpLU/rEzB8w9ayxBSZIkqbM8UUySJKkjllioHajNm9gkSZKkB8SkVpIkSZ1n+4EkSVJH2H4wmJVaSZIkdd6irNTe9K07hl7jKB+ZIXXe1slNQ6+xbv3mEUQymlgk/XBOfeRxTYcwMj7SazArtZIkSeo8k1pJkiR13qJsP5AkSVqM3Cg2mJVaSZIkdZ6VWkmSpI5wn9hgVmolSZLUeWNNapNMJ1kzZ2xjkguSXJHkYJLL53yeJG9K8q9Jbkzy++OMWZIkSe037vaDKWAC2N03NgGcBywFjgFePeeedcATgKdW1X1JHjOGOCVJklpnif0HA427/WA7cE6SZQBJVgIrgKur6kpgvlMTfhd4Q1XdB1BVt40nVEmSJHXFWJPaqjoA7AHW9oYmgG1VVYe57UnAS5PsTfKPSU5d6DglSZLaaEkLrrZqIrbZFgR6P6eOMH8Z8L2qWg38T2ByvklJNvQS373/fOm7RxasJEmS2q+JpHYncHaSVcDyqtp3hPn7gff2Xu8AnjXfpKraUlWrq2r1Wb/+W6OLVpIkSa039qS2qg4B08xUXI9UpYWZJPis3uufA/51YSKTJElqt6T5q62aao2YAk4DLp4dSHIVcAkzVdz9fY/++lPghUk+BWwGXjXuYCVJktRujZwoVlU7gMwZO3PA3IPA88cRlyRJUpv5SK/B2ryJTZIkSXpATGolSZLUeY20H0iSJOnBs/tgMCu1kiRJ6jyTWkmSJHXeomw/eNSyZUOv8e277hpBJJK6buvkppGss2795qHXGFUsUpfcWzX0Go/4kaNHEEk7LLH9YCArtZIkSeq8RVmplSRJWox8Tu1gVmolSZLUeSa1kiRJ6jzbDyRJkjrC7oPBrNRKkiSp88ZaqU0yDWyuqt19YxuBpwA/BjwbuLqqzun7/CrguN7bxwB7qupXxxa0JElSS/hIr8HG3X4wBUwAu/vGJoDzgKXAMcCr+2+oqjNnXyd5L3DZwocpSZKkLhl3+8F24JwkywCSrARWMFOdvRK4Y9CNSY4DzgJ2LnyYkiRJ6pKxJrVVdQDYA6ztDU0A26oe0HEhvwZcWVW3L1R8kiRJbZYW/NNWTWwUm21BoPdz6gHe9xuHm5tkQ5K9SfbuuuTCIUOUJElSlzTxSK+dwPlJVgHLq2rfkW5IciJwBjPV2nlV1RZgC8AVN3x9+IOiJUmSWsaNYoONvVJbVYeAaWCSB16lfTFweVV9b6HikiRJUnc19ZzaKeA04OLZgd6juy4Bzk6yP8mavvkPpk1BkiRJDzGNnChWVTvg+zuN+x/dNc/85y10TJIkSW1n+8FgnigmSZKkzmukUitJkqQHL7FUO4iVWkmSJHWeSa0kSZI6z/YDSZKkjnCj2GCLMqm96eChodd4zDFLRxCJJM3YOrlp6DXWrd/cijikcbr7vuHPU/rSN78zgkjgSSctH8k6Whi2H0iSJKnzFmWlVpIkaTHy4QeDWamVJElS51mplSRJ6ogllmoHslIrSZKkzjOplSRJUueNNalNMp1kzZyxjUkuSHJFkoNJLp/z+dlJ9iW5LsnVSZ48zpglSZLaYkmav9pq3JXaKWBizthEb/wtwMvmueevgd+qqtOBvwdet6ARSpIkqXPGvVFsO/DGJMuq6s4kK4EVwNVVVUmeN889BRzfe/0I4NZxBCpJktQ27hMbbKxJbVUdSLIHWAtcxkyVdltVHe64kFcBu5J8F7gdePbCRypJkqQuaWKjWH8LwmzrweH8IfBLVXUK8A7g/PkmJdmQZG+SvR+97EhLSpIkaTFp4jm1O4Hzk6wCllfVvkETk5wEnFZV/9Ib2gZcMd/cqtoCbAF420e/MPxB0ZIkSS2zBPsPBhl7pbaqDgHTwCRHrtJ+C3hEkqf03v8CcOPCRSdJkqQuaupEsSngUvqehJDkKuCpwLFJ9gOvrKrdSf4z8N4k9zGT5K5vImBJkqSmuVFssEaS2qraAd9fP6+qMw8zd8c44pIkSVI3eaKYJEmSOq+p9gNJkiQ9SG0+0atpVmolSZLUeVZqJUmSOmKJO8UGslIrSZKkzluUldpHLFuUfyxJD3FbJzcNvca69ZtbEYf0QJ348KVDr/GlO74zgkgAThzROloIZn+SJEkdYffBYLYfSJIkqfOs1EqSJHWEG8UGs1IrSZKkzjOplSRJUufZfiBJktQRdh8MZqVWkiRJnTfWpDbJdJI1c8Y2JrkgyRVJDia5fM7nZyXZl+TTSd6ZxOqyJEmSvs+4K7VTwMScsYne+FuAl/V/kGQJ8E5goqqeAfxv4BVjiFOSJKl1lrTgaqtxx7YdOCfJMoAkK4EVwNVVdSVwx5z5JwJ3VtW/9t5/EHjheEKVJElSV4w1qa2qA8AeYG1vaALYVlU14JZvAEcnWd17/yLgCfNNTLIhyd4kez906btHGbYkSVIrJGn8aqsmqsj9LQizrQfz6iW7E8Bbk+xhppJ7z4C5W6pqdVWt/vlf/60RhyxJkqQ2a2LT1U7g/CSrgOVVte9wk6vqGuBMgCT/CXjKwocoSZKkLhl7pbaqDgHTwCSHqdLOSvKY3s9lwGuBv1nI+CRJktoqLbjaqqlNbFPAacDFswNJrgIuAc5Osr/v0V/nJbkRuB74h6r657FHK0mSpFZr5JmvVbWDOcl+VZ05YO55wHnjiEuSJKnNlrR4o1bT2vy4MUmSJOkBMamVJElS53nkrCRJUkfYfDCYlVpJkiR13qKs1D5y2dKh1/jOPfOe8SBJnbZ1ctPQa6xbv3kEkYwmFi1+79z7laHXeMEzHzOCSNrBfWKDWamVJElS55nUSpIkaaSSrE3yuSQ3J/njeT5flmRb7/N/SbJy2N+5KNsPJEmSFqN0oP8gyVHA24FfAPYDH0vyvqr6TN+0VwLfqqonJ5kA3gy8dJjfa6VWkiRJo3QGcHNV3VJVdzFzguwL5sx5AfDO3uvtzJwoO1TGblIrSZLUEUtacCXZkGRv37VhTpiPB77c935/b2zeOVV1D/Bt4MQf8msBbD+QJEnSg1BVW4Ath5kyX8W1fog5D4qVWkmSJI3SfuAJfe9PAW4dNCfJw4BHAN8c5peONalNMp1kzZyxjUl2JbkmyQ1Jrk/y0r7Pn9jbFXdTb5fc8A+hlSRJ6qAkjV8PwMeAU3s53FJgAnjfnDnvA17Re/0i4J+rqlOV2ilm/mD9Zne8vbyqng6sBf4iyQm9z98MvLWqTgW+xcxuOUmSJLVQr0f2XGA3cCPwnqq6IckbkvxKb9rfAScmuRl4DfADj/16sMbdU7sdeGOSZVV1Z++ZZCuAj8xm51V1a5LbgJOSfBs4C/jN3v3vBP4E+Osxxy1JktS49j/Qa0ZV7QJ2zRl7fd/r7wEvHuXvHGultqoOAHuYqcbCTJV2W3+5OckZwFLg88zsgjvYy/hh/t1zs/f9+068K7a/a6H+CJIkSWqhJp5+MNuCcFnv5/rZD5KcDLwLeEVV3TfgeWXz9lv078T7h099baieDEmSJHVLE0ntTuD8JKuA5VW1DyDJ8cD7gddV1bW9ud8ATkjysF61dr7dc5IkSQ8JXThRrCljf6RXVR0CpoFJZqq29HbG7QAurKpL+uYW8CFmdsXBzC65y8YZryRJktqvqefUTgGnMXNsGsBLgOcC65Jc17tO7332WuA1vd1xJzKzW06SJEn6d42cKFZVO+jbwFdVFwEXDZh7CzNnCEuSJD2keWrWYH43kiRJ6rxGKrWSJEl68NwoNpiVWkmSJHWeSa0kSZI6b1G2H3z9e3cOvcaPPOyoEUQiSYvP1slNI1ln3frNQ68xqljUXj/7YycMvcY99903gkjaweaDwazUSpIkqfMWZaVWkiRpMXKf2GBWaiVJktR5JrWSJEnqPNsPJEmSOmKJW8UGslIrSZKkzrNSK0mS1BFuFBtsrJXaJNNJ1swZ25hkV5JrktyQ5PokL+37/NwkNyepJI8eZ7ySJEnqhnG3H0wBE3PGJoA3Ay+vqqcDa4G/SDL7tOWPAv8R+N9ji1KSJEmdMu72g+3AG5Msq6o7k6wEVgAfqaoCqKpbk9wGnAQcrKpPAMR6uyRJeoiLG8UGGmultqoOAHuYqcbCTJV222xCC5DkDGAp8PkHs3aSDUn2Jtn74R1/P6qQJUmS1AFNbBSbbUG4rPdz/ewHSU4G3gW8oqoe1EHNVbUF2AIw+bEv1RGmS5IkdY5/cT1YE4/02gmcnWQVsLyq9gEkOR54P/C6qrq2gbgkSZLUUWNPaqvqEDANTDJTtSXJUmAHcGFVXTLumCRJktRtTR2+MAWcBlzce/8S4LnAuiTX9a7TAZL8fpL9wCnA9Un+tpGIJUmSGraENH61VSOHL1TVDrj/W6mqi4CLBsz9S+AvxxSaJEmSOshjciVJktR5HpMrSZLUET79YDArtZIkSeo8K7WSJEkdYaV2sEWZ1H719ruGXuPJj1o+gkgkSYNsndw09Brr1m9uRRxaOD/+qGOHXuPQ3XePIBK1ne0HkiRJ6rxFWamVJElajNLi58Q2zUqtJEmSOs9KrSRJUkcssVA7kJVaSZIkdZ5JrSRJkjrP9gNJkqSOcKPYYGOt1CaZTrJmztjGJLuSXJPkhiTXJ3lp3+fvTvK5JJ9OMpnk6HHGLEmSpPYbd/vBFDAxZ2wCeDPw8qp6OrAW+IskJ/Q+fzfwVOCZwHLgVWOKVZIkqVWS5q+2GndSux04J8kygCQrgRXAR6rqJoCquhW4DTip935X9QB7gFPGHLMkSZJabqxJbVUdYCYxXdsbmgC29RJWAJKcASwFPt9/b6/t4GXAFfOtnWRDkr1J9n7s8osXInxJkiS1VBNPP+hvQZjovQcgycnAu4Dfrqr75tx3ATMV3avmW7SqtlTV6qpa/dPnzO1wkCRJ6r604J+2aiKp3QmcnWQVsLyq9gEkOR54P/C6qrq2/4Yk/5WZdoTXjDtYSZIktd/YH+lVVYeSTAOT9Kq0SZYCO4ALq+qS/vlJXgWsAc6ep3orSZL0kOGJYoM1dfjCFHAaMNv8+hLgucC6JNf1rtN7n/0N8Fjgmt7468cfriRJktqskcMXqmoH3N+UUVUXARcNmOsBEZIkSTosE0ZJkqSOaPNGraY11X4gSZIkjYyVWkmSpI5o84leTbNSK0mSpM5blJXaxx23tOkQJEljsHVy09BrrFu/eQSRjCYW/aBjjj5q6DUO3X33CCJR2y3KpFaSJGkxsvtgMNsPJEmS1HkmtZIkSeo82w8kSZI6YomPPxjISq0kSZI6z0qtJElSR1inHcxKrSRJkjpvrEltkukka+aMbUyyK8k1SW5Icn2Sl/Z9/ndJPtkb357k2HHGLEmSpPYbd6V2CpiYMzYBvBl4eVU9HVgL/EWSE3qf/2FVnVZVzwK+BJw7tmglSZLaJC24WmrcSe124JwkywCSrARWAB+pqpsAqupW4DbgpN7723tzAywHaswxS5IkqeXGmtRW1QFgDzPVWJip0m6rqn9PVJOcASwFPt839g7g34CnAm+bb+0kG5LsTbL3wzv+foH+BJIkSc1JC/5pqyY2ivW3IEz03gOQ5GTgXcBvV9V9s+NV9dvMVHRvBF7KPKpqS1WtrqrVP/drv7lQsUuSJKmFmkhqdwJnJ1kFLK+qfQBJjgfeD7yuqq6de1NV3QtsA144zmAlSZLUfmN/Tm1VHUoyDUzSq9ImWQrsAC6sqktm5/b6aJ9UVTf3Xv8y8NlxxyxJktQGHig2WFOHL0wBl3J/G8JLgOcCJyZZ1xtbB1wPvLNXxQ3wSeB3xxqpJEmSWq+RpLaqdtD3UIiqugi4aMD054wlKEmSpJazUDuYJ4pJkiSp80xqJUmS1HlN9dRKkiTpwbL/YCArtZIkSeq8RVmpPX7p8H+s+8rTeCXpoWDr5KaRrLNu/eah1xhVLIvJv33ne0OvcdQieg5Wm0/0apqVWkmSJHWeSa0kSZI6b1G2H0iSJC1Gi6iTYuSs1EqSJKnzrNRKkiR1hIXawazUSpIkqfNMaiVJktR5Y01qk0wnWTNnbGOSXUmuSXJDkuuTvHSee9+W5ND4opUkSWqZtOBqqXH31E4BE8DuvrEJ4LXArVV1U5IVwMeT7K6qgwBJVgMnjDlWSZIkdcS42w+2A+ckWQaQZCWwAvhIVd0EUFW3ArcBJ/XmHAW8BfgvY45VkiRJHTHWpLaqDgB7gLW9oQlgW9X9Z9ImOQNYCny+N3Qu8L6q+urh1k6yIcneJHv/6b0XjT54SZKkhqUF/7RVE4/0mm1BuKz3c/3sB0lOBt4FvKKq7uu1IrwYeN6RFq2qLcAWgEuuu7WOMF2SJEmLSBNJ7U7g/CSrgOVVtQ8gyfHA+4HXVdW1vbk/CTwZuDkzR2gck+TmqnpyA3FLkiQ1yhPFBht7UltVh5JMA5PMVG1JshTYAVxYVZf0zX0/8LjZ90kOmdBKkiRprqaeUzsFnAZc3Hv/EuC5wLok1/Wu0xuKTZIkSR3TyDG5VbWDviedVdVFwBF3d1XVsQsZlyRJUpvZfTCYJ4pJkiSp8xqp1EqSJOmHYKl2ICu1kiRJ6jyTWkmSJHXeomw/+PETjx96jRu/8e0RRCJJeqjYOrlp6DXWrd/cijja5NiHDZ+qHLrnnhFE0g5tPtGraVZqJUmS1HmLslIrSZK0GHmi2GBWaiVJktR5JrWSJEnqPNsPJEmSOsLug8Gs1EqSJKnzrNRKkiR1haXagcZaqU0ynWTNnLGNSXYluSbJDUmuT/LSvs+3JvlCkut61+njjFmSJEntN+5K7RQwAezuG5sAXgvcWlU3JVkBfDzJ7qo62JtzXlVtH3OskiRJ6ohxJ7XbgTcmWVZVdyZZCawAPlJVBVBVtya5DTgJODhwJUmSpIcYTxQbbKztB1V1ANgDrO0NTQDbZhNagCRnAEuBz/fd+qZeW8Jbkyybb+0kG5LsTfL/t3fvcXJUdd7HP19CAgQIQmAhXCSA4A0CxIgKyiUB5VkvgNyCXAxB2YsiuC9ZwWXd5/HZFXB1HwVcXS4hkGC4hQRcQBAkREQuIQyXCCaCyCUBBGEloCFhfs8fdcY0w/RMdbqnq6rn+86rXlN9qurXv9NT6Tl9+tSpBVdfNm2QamBmZmZmZVTE7Ac9QxBIP2f1bJA0BpgBHB8R3an4dOBdwPuBTciGKrxFRJwfERMiYsJhR08drNzNzMzMrISKaNTOBSZJGg+sFxELASSNAq4HzoiIu3p2johlkVkBXAzsUUDOZmZmZoWTil/Kqu2N2ohYDswDppF6aSWNAOYAl0bEVbX7p95bJAk4GHi4nfmamZmZWfkVNU/tLOAaVg9DOALYGxgtaUoqmxIRXcBlkjYjm5mtC/jbNudqZmZmVgol7igtXCGN2oiYQ83vJSJmAjPr7DuxXXmZmZmZWTX5NrlmZmZmVnm+Ta6ZmZlZVXj8QV3uqTUzMzOzynNPrZmZmVlF+I5i9XVko/a3Ly8vOgUzM7OGTZ92etMxpkw9sxR5tMprq1YVnYJVhIcfmJmZmVnldWRPrZmZmVknKvMdvYrmnlozMzMzqzw3as3MzMwqQiVYmspf2kTSTyUtST837mOfbSXdJ6lL0iJJue4m60atmZmZmbXLacCtEbEjcGt63NsyYM+I2A34AHCapC0HCuxGrZmZmZm1y0HAJWn9EuDg3jtExOsRsSI9XIec7VU3as3MzMyqouixBwJJJ0paULOc2EANNo+IZQDp51/1WU1pG0kPAk8BZ0fE0oECt3X2A0nzgDMj4qaaslOAjwIbA6OAN4B/i4gr0nYB/wocnrb9ICLOaWfeZmZmZpaJiPOB8+ttl3QLsEUfm/6pged4ChiXhh3MlXR1RDzX3zHtntJrFjAZuKmmbDLwVWBpRCxJyd8n6aaIeBmYAmwDvCsiuiX12aI3MzMz63RVuKNYROxfb5uk5ySNiYhlksYAzw8Qa6mkRcBHgKv727fdww+uBj4haR0ASWOBLYH5EbEEsuTJKrhZOubvgG9ERHfa3m/lzczMzKy0rgM+m9Y/C1zbewdJW0taL61vDOwF/HqgwG1t1EbEi8A9wIGpaDJwRUREzz6S9gBGAI+loh2AI9OYjRsl7dhX7NrxHTddPWPwKmFmZmZma+os4ABJS4AD0mMkTZB0Ydrn3cDdkh4Abge+HREPDRS4iDuK9QxBuDb9nNqzIXVDzwA+29MzS3bV258jYoKkTwPTyLqg36R2fMe1Dz0bvbebmZmZVV3V7yiWOjgn9VG+APhcWv8pMK7R2EXMfjAXmCRpPLBeRCwEkDQKuB44IyLuqtn/aWB2Wp/DGlTSzMzMzDpb2xu1EbEcmEfW4zoLQNIIsgbrpRFxVa9D5gIT0/o+wOL2ZGpmZmZWLiWY0au0ipqndhawK3B5enwEsDcwJd0SrUvSbmnbWcChkh4CziR1TZuZmZmZ9ShiTC0RMYeaxn5EzARm1tn3ZeDjbUrNzMzMzCqokEatmZmZma2BMn//XzDfJtfMzMzMKs+NWjMzMzOrPA8/MDMzM6uIKtwmtygd2ahd8Ub3wDuZmZl1oOnTTm86xpSpZ7Ygk9bk0oq/6cOqfscCy6UjG7VmZmZmncjt8/o8ptbMzMzMKs+NWjMzMzOrPA8/MDMzM6sIjz6ozz21ZmZmZlZ57qk1MzMzqwp31dblnlozMzMzq7y2NmolzZP0sV5lp0i6QdIvJS2S9KCkI2u2/1xSV1qWSprbzpzNzMzMrPzaPfxgFjAZuKmmbDLwVWBpRCyRtCVwn6SbIuLliPhIz46SZgPXtjVjMzMzs5LwHcXqa/fwg6uBT0haB0DSWGBLYH5ELAGIiKXA88BmtQdK2hCYCLin1szMzMzepK2N2oh4EbgHODAVTQauiIjo2UfSHsAI4LFehx8C3BoRf+wrtqQTJS2QtOCW2TNbn7yZmZlZwaTil7Iq4kKxniEIpJ+zejZIGgPMAI6PiN43ez6qdt/eIuL8iJgQERP2P/SYFqdsZmZmZmVWRKN2LjBJ0nhgvYhYCCBpFHA9cEZE3FV7gKTRwB5pu5mZmZnZm7R9ntqIWC5pHjCN1PMqaQQwB7g0Iq7q47DDgf+OiD+3LVEzMzOzkinxt/+FK2qe2lnArsDl6fERwN7AlJrpu3ar2f9NwxTMzMzMzGoVckexiJhDzYeNiJgJ1L26KyL2bUNaZmZmZqVW5gu1iuY7ipmZmZlZ5blRa2ZmZmaVV8jwAzMzMzNbEx5/UI97as3MzMys8txTa2ZmZm8yfdrpLYkzZeqZTce46KLTWpBJ5/CFYvW5p9bMzMzMKs+NWjMzMzOrPA8/MDMzM6sIjz6ozz21ZmZmZlZ5btSamZmZWeV5+IGZmZlZRXj2g/rcU2tmZmZmldfWRq2keZI+1qvsFEk3SPqlpEWSHpR0ZM32SZIWSuqSdIekd7QzZzMzM7OyUAn+lVW7e2pnAZN7lU0GzgaOi4j3AgcC35X0trT9B8DREbEb8CPgjHYla2ZmZmbV0O5G7dXAJyStAyBpLLAlMD8ilgBExFLgeWCzdEwAo9L6RsDSNuZrZmZmZhXQ1kZtRLwI3EPWGwtZL+0VERE9+0jaAxgBPJaKPgfcIOlp4FjgrL5iSzpR0gJJC26ZPXOwqmBmZmZWHJVgKakiLhSrHYIwOT0GQNIYYAZwfER0p+IvA38dEVsDFwP/0VfQiDg/IiZExIT9Dz1m0JI3MzMzs/IpolE7F5gkaTywXkQsBJA0CrgeOCMi7kplmwG7RsTd6dgrgD0LyNnMzMyscEV30pa4o7b9jdqIWA7MA6aRemkljQDmAJdGxFU1u78EbCRpp/T4AOCR9mVrZmZmZlVQ1M0XZgHXsHoYwhHA3sBoSVNS2ZSI6JL0eWC2pG6yRu7UdidrZmZmZuVWSKM2IuZQ04MdETOBPq/uSvvOaVNqZmZmZqXlO4rV5zuKmZmZmVnlFTX8wMzMzMwaVOY7ehXNPbVmZmZmVnlu1JqZmZlZ5Xn4QR3v2nTUwDsN4J1jNmxBJvDIM680HeP2J19oOsbhu2zVdAyA78x/vOkYjzzzP03HOHW/dzQd45WVK5uOAbDh8OFNx/j9n/7cdAy14AqEN1bfILApr7/RPfBOA1h3WHk+t7fidVnZ3XyM0euOaDrGJQueaToGwIe2f1vTMd65yQZNxxg5fFjTMQCefa35/4MbrN38n+XXVq1qOsaKFvz/A7jootOajnHCCX3eSLQh06ad3nSM0vDog7rK845vZmZmZraG3FNrZmZmVhHuqK3PPbVmZmZmVnlu1JqZmZlZ5Xn4gZmZmVlF+I5i9bmn1szMzMwqzz21ZmZmZhXhO4rVl6unVtIhkkLSuwY7oX5yOEXSyKKe38zMzMzKK+/wg6OAO4DJg5jLQE4B3Kg1MzMzs7cYsFEraQNgL+AEUqNW0r6Sbpd0paTFks6SdLSkeyQ9JGmHtN+2km6V9GD6+fZUPl3SYTXPsbwm7jxJV0t6VNJlynwJ2BK4TdJtLX8VzMzMzCpAKn4pqzw9tQcDP4mIxcAfJI1P5bsCJwO7AMcCO0XEHsCFwElpn/OASyNiHHAZcE6O59udrFf2PcD2wF4RcQ6wFNgvIvbr6yBJJ0paIGnBLbNnuQTkDAAAE7pJREFU5ngaMzMzM+sUeRq1RwGXp/XL02OAeyNiWUSsAB4Dbk7lDwFj0/qHgB+l9RnAh3M83z0R8XREdANdNbH6FRHnR8SEiJiw/6HH5DnEzMzMzDpEv7MfSBoNTAR2lhTAMCCAG4AVNbt21zzu7idupJ+rSA1qSQJG1OxTG/eNgXI0MzMzMxuop/YwsuED20bE2IjYBvgt+XpcAe5k9cVlR5NdbAbwBPC+tH4QMDxHrFeADXM+r5mZmZkNIQM1ao8C5vQqmw18Jmf8LwHHS3qQbNztyan8AmAfSfcAHwBezRHrfOBGXyhmZmZmQ1XRF4mV+UKxfr/aj4h9+yg7h14XfNXuFxHzgHlp/Qmy4Qu9YzwHfLCm6PTex6bHX6xZPxc4t798zczMzGxo8m1yzczMzKzyfBGWmZmZWUX4Nrn1uafWzMzMzCrPPbVmZmZmFVHmC7WKpogYeK+KuapradOVGr3uOk3n8fALf2w6BsBeW49uOsZNj/2+6Rgf2GqjpmMALF+5qukYW20wsukYj/6h+d/PmPXXbToGwOiRzZ9v1/36uaZjHPqeMU3HePj5l5uOAbD7lhs3HeP+pS81HWPHjVszk+BG6+eZubB/T/7hteZjvNJ8jGEt+qu6qru76RjrrT2sBZm0hlrwurxRkr/J3d2tyaMVp0orXtepU89sPhHgT/efV3iT8o9/btEvpwmj1l2r8NehLx5+YGZmZmaV5+EHZmZmZhVRyi7SknBPrZmZmZlVnntqzczMzKrCXbV1uafWzMzMzCrPjVozMzMzqzwPPzAzMzOrCN9RrL5B6amVtIWkyyU9JulXkm6QtJOkhwfj+czMzMxsaGt5T62yWZLnAJdExORUthuweaufy8zMzGwo8R3F6huMntr9gJUR8cOegojoAp7qeSxprKSfS1qYlj1T+RhJ8yV1SXpY0kckDZM0PT1+SNKXByFnMzMzM6uwwWjU7gzcN8A+zwMHRMR44EjgnFT+GeCmiNgN2BXoAnYDtoqInSNiF+DivgJKOlHSAkkLbpk9sxX1MDMzM7OKKOpCseHAeWlYwhvATqn8XmCapOHA3IjokvQ4sL2kc4HrgZv7ChgR5wPnA1zVtbTw+yKbmZmZtZpHH9Q3GD21i4D3DbDPl4HnyHpjJwAjACJiPrA38AwwQ9JxEfFS2m8e8AXgwkHI2czMzMwqbDAatT8D1pH0+Z4CSe8Htq3ZZyNgWUR0A8cCw9J+2wLPR8QFwEXAeEmbAmtFxGzgn4Hxg5CzmZmZmVVYy4cfRERIOgT4rqTTgD8DTwCn1Oz2n8BsSYcDtwGvpvJ9gVMlrQSWA8cBWwEXS+ppgJ/e6pzNzMzMKsHjD+oalDG1EbEUOKKPTTun7UuAcTXlp6fyS4BL+jjOvbNmZmZmVpfvKGZmZmZWEb6jWH2DckcxMzMzM7N2cqPWzMzMzCrPww/MzMzMKsK3ya3PPbVmZmZmVn0RMSQX4MQyxChTLmWJUaZcyhKjTLm4Pn5NqpZLWWKUKRfXZ/BeEy/FLUO5p/bEksRoVZxOitGqOJ0Uo1VxyhKjVXE6KUar4pQlRqvidFKMVsUpS4xWxemkGFagodyoNTMzM7MO4UatmZmZmVXeUG7Unl+SGK2K00kxWhWnk2K0Kk5ZYrQqTifFaFWcssRoVZxOitGqOGWJ0ao4nRTDCqQ0ONrMzMzMrLKGck+tmZmZmXUIN2rNzMzMrPLcqDUzMzOzyhsyjVpJm0u6SNKN6fF7JJ1QdF5mZmZm1rwh06gFpgM3AVumx4uBU5oNKumABvcfJWmHPsrHNRBjC0lbpPXNJH1a0nsbyaOPmN9s8vjtUh7vavC4t0taN61L0vGSzpX0d5LWzhnjUz0xmiFpb0nvTOsflvQVSR9fgzgbSDpM0pclnSTpQEm5/69JWlvS30j6iaQHJT0g6UZJfytpeKP59BE/1xW+koalPP6vpL16bTujgecbKekfJZ0qaV1JUyRdJ+lbkjZoNP+auIsb3H9czfpwSWekPL4paWQDcb4oadO0/g5J8yW9LOluSbvkjHGNpGOarP/2kqZJ+td0zl0g6WFJV0kamzPGWpKmSro+nWf3Sbpc0r4N5tIx5+xgna8pdtvP2Vacr+nYUpyzNbFuzVOWI87Jyv4uS1nH10JJH200jpXDkJn9QNK9EfF+SfdHxO6prCsidmsy7pMR8fac+x4BfBd4HhgOTImIe9O2hRExPkeMvwFOAwScDUwBFgF7Ad+KiItyxDindxFwLHApQER8KUeMuRFxcFo/KNVrHrAncGZETB8oRjr2YWCPiHhN0tnADsBcYGLKZWqOGH8CXgVuBGYBN0XEG3mevybGd4E9gLXJPvxMSvH2Ae6PiFNzxjkCOBV4ANgPuJPsw+MuwNER8VCOGLOAl4FLgKdT8dbAZ4FNIuLIHDE2qbcJeCAits4R40JgJHAP2flxe0T8Q9qW63xN+14JPAWsB7wTeAS4EvgksEVEHJsjxitAz5uV0s+RwGtARMSoHDH+krOk7wCjgYuBg4HREXFczvosioj3pvXrgQsjYk5qCP5bROzVb4DsuGeAX5Kd57eQnbfXR8TreXJIMean4zYCjkl1uRL4KNm5NjFHjIuB36UcDgP+CPwc+CpwbUScmzOXjjlnW3G+pjilOGdbcb6mY8tyzq5L9jreBuzL6td2FHBjRLw7bz4p3gMRsaukjwFfAP4ZuDjv+5uVTNH36W3XQtbgGg0sTI8/SPaGl+fY6+osPwZebSCHLmBMWt8DeBT4dHp8f84YD5H9hx4NLCd7kwXYGOjKGeNpYCZwHNkfnc8Cv+9Zzxnj/pr1O4Ht0vqmZH+A8r4mv6pZvw9Yq+ZxrjjA/an+nwduBZ4Dfgjs00Aei8jeHEcCLwEjU/lw4OEG4jxYc+ymZA1sgHHAnTlj/LqfbYtzxngDeBz4bc3S8/j1vHWpWV+bbA7Ha4B18p6vPed9+ingWVZ/mFbtcwwQ41yyD12b15T9Nm8OfZyzXcDwRvPo/fsB7q33muXJBdiQrPF1Q/o/eDHw0TWoz5P1tuX9HafHd6Wf6wCPrMlrUvVzthXna5nO2VacryU7Z09O58SKXufLA8AXG3l9a18D4HvAIY3k4qV8S66vdzvEP5A1RHeQ9AtgM7KeiTw+QvapcnmvcpE1TvNaOyKWAUTEPZL2A/5b0tas/kQ/kFUR8RrwmqTHIuLZFO8lSXljvAf4BnAgcGpEPCPpXyLikgbqUvtca0fEb1MeL0jqbiDOU5ImRsTPgCeAbYDfSRrdSC4R8RJwAXCBsqEZRwBnSdo6IrbJGSNqcu+pXzeNDdMR8Ke0/irwVyn4g5IG7JlJXpJ0ODA7Iroh+5oYOJyswZ3H48CkiHjyLQlKT+WMMaJnJSJWASdK+jrwM6DhryDT63tDRPZXIz3Odc5GxEmS3gfMkjQXOI/8/2d6bCTpELLf5zoRsbLRPJKrJU0n+z80R9IpZA2nScBbXu86el6DV4AZwIzUU3kE2TcxN+eI0S1pJ7Jer5GSJkTEAknvAIblzGOlpB0i4jFJ44HXU14rGnxNOu6cbeZ8TfuX5ZxtxfkKJTlnI+J7wPcknRQ5v0kYwH2Sbga2A06XtCHZ+75VUdGt6nYuZJ/c3wvsTPrEm/O4G4H96myb30CcO4EdepVtSNa7uCJnjAWs/rS+dU35ujTQQ5qOeR/ZVzhfAZ5o8NhVZF9VvgKsZHWP8Qga+/S/TcphPlnP90tkf4DuJ/sDlyfGwn62bZszxtnAHcC9wL+nXP6J7I36hw3U5yyy4QtfI/sa92upfBNgUc4YY4EryHpBFqfl+VS2Xc4YXwB2rbPtpJwxZgIH9lH+OWBlA6/JhcAGfZTvANzR4Hm3FvCl9NoubfDYi3stm6fyLYBbG4w1BbgbeCH9H/gV8E1go5zH537f6CfGJODXZF+PfxiYDfwmnSsH5Ywxkaxhs5ist+sDqXwzsuFMeXPpmHO2ledrWc7ZZs/XMp2zveLtCXyG7FvH44Dj1vD3Mx54W3q8CTCu2bp6KWYZSmNqhwEfJ3vz/UsPdUT8R45jvw/8KCJ+0WQO1wNnRcTPe5UPB46IiMtyxJgGTIuIO3qVbwW8OyJuyRHjPLL63ClJwN8DH4qIYxqoS5+viaS3pTx+mTPOeWRjrF4CdiT73TxN9jVZrk/Lkn4FfC4i7sybfx8xvg9cTvY1593KLuY7hOwP/tUN5PJ9sq8sXyP7kHFLKl+L7MPIigbzGk329ecLjRxXFZIUa/AmJGkMsHtE3DAIaVWWsguCXooGxpSn94DRrTrHOvmcXdPzNR3rc7YPa3LOpuNmkH3Q6CIbugJZJ/aA14T0irMX2ZCTVyUdQ9bA/V5E/K6ROFYOQ2n2gx+TfVodTdY72rPksRj4tqQnJJ0taU0vLrsZ+FbvOBGxMk+DNnkA+Pc+YjyTp0GbLAG+I+kJsp7FXzTSoE36fE0i4uW8DdqaXL5NNj5rT+CxiLg7byMy+S9SfZr4/SwGvgVcoeyCtQ0j4tsRcWWDuSwG/pqsZ+aAmtelu9EGbTruxdrGgRqcbaMvZYmR7L8mB0XEsp7GQZnqU3QuEfFCRLzRSIzIvKUB2mgeSjO79HHONjKzSytmhxm0GGQXfOZWG6fXOVuK+jQSY7ByqTlnG8oFmADsFRF/HxEnpaWhBm3yA7LhfLsC/0h24eSlaxDHyqDoruJ2LTTwlXg/MbYluyr4frKvT74O7NSiODu2O5eyxChTLmWqT524T3ZKjDLl4vo0F4NsXOVSsl6zRcD7a7bVHR7UqTHKlEun1adm/6tIF143s7D64vGvAyesSS5eyrMMpeEHZ5ONQcozmD1PvN2BaWRjb/JelDEocTopRplyKao+kq6rtwmYGBHrVyVGmXJxfQYnRorTBfyviFgmaQ+ynq6vRcQ1qplGcajEKFMunVafmli3AbuRTd/2l2/BIuJTeWOkOLcDPwGOB/YmGxfeFREN9cpbOQyl2Q/uIrvycy2yC5tEzrkCeygb+3ogMJlswPvtwP9pNJFWxOmkGGXKpST1acVsG2WJUaZcXJ/BiQGtmdmlk2KUKZdOq0+P/93g/vUcSXax2QkR8aykt5NdLGxVVHRXcbsWsulixpHmHGzw2APIetqeIxubezSwfhFxOilGmXIpWX2anm2jLDHKlIvrM6ivSStmdumYGGXKpdPq48VLf8tQ6qldQjaJ/pqMt/ga8CPgKxHxhyZyaEWcTopRplzKVJ/HSfOF9hYRe1csRplycX0GJwZks5dsCTxWc/wrkg4kG0c51GKUKZdOqw9A7zu2jSC7Uc6r0cC3rynOB8lulPHuFGcYsDwiNmokjpXDUBpTOx3Ynqxnonb8zYBTepm1k6STyYYtjCGb53NWRHRVMUaZcnF9BidGmXIpS4wy5dJp9ekn9sFkt1z/WoPHLUg5XUU2o8JxZBduNxTHymEoNWr/pa/yiGh4vKRZO0jaluzNdjLZzTVmAZdHxOKqxShTLq7P4MToJ86siFgyFGOUKZdOq0+duHdFxAcbPGZBREyQ9GBEjEtld0bEns3kYsUYMo1asyory2wOrYhRplxcn8GJUaZcyhKjTLl0Qn0kfbrm4Vpkvaz7RMSHGnz++WTzZV9IduOcZcCUiNi1kThWDh1/8wVld6xC0o8lXdd7KTo/s3okDZf0SUmXkQ2bWQwcWsUYZcrF9RmcGGXKpSwxypRLp9UH+GTN8jGy2/8e1GAMgGPJxtF+EXiV7NbtDb8uVhLtuBqtyAX4Y/q5T19L0fl58dJ7oSSzObQiRplycX38mvg1qX59vHjpb+n44QdqcEJns6Ipm1T8R8DsWMMZFMoSo0y5uD6DE6NMuZQlRply6bT61MTammzWgr3IZkG4Azg5Ip7OefxD9DM3bqTxtVYtQ6FR+zRQd4aD8OwHZmZmlSLpp2QN5Bmp6Bjg6Ig4IOfxOwKbA0/12rQtsDQiftOqXK19On5MLdlYmQ3IJnjuazEzM7Nq2SwiLo6IVWmZDmzWwPH/j2x44u9qF+C1tM0qaCjcfGFZRHyj6CTMzMysZV6QdAzZdGAARwEvNnD82Ih4sHdhRCyQNLb59KwIQ6GnVkUnYGZmZi01lewuZD3TcB2WyvJat59t6zWRlxVoKIyp3aTZAelmZmbWOSTNAn4WERf0Kj8B+GhEHFlMZtaMjm/UmpmZWWeRtB1wEjCWmqGUEfGpnMdvDswBXgfuS8UTgBHAIRHxbCvztfZwo9bMzMwqRdIDwEXAQ0B3T3lE3N5gnP2AndPDRRHxs5YlaW3nRq2ZmZlViqS7I+IDRedh5eJGrZmZmVWKpM8AOwI3Ayt6yiNiYWFJWeGGwpReZmZm1ll2AY4FJrJ6+EGkxzZEuafWzMzMKkXSo8C4iHi96FysPIbCPLVmZmbWWR4A3lZ0ElYuHn5gZmZmVbM58Kike1k9pjYi4qACc7KCefiBmZmZVYqkfWofAh8GjoqI9xaUkpWAhx+YmZlZpaT5aP8H+DgwHZgE/LDInKx4Hn5gZmZmlSBpJ2AycBTwInAF2bfO+xWamJWChx+YmZlZJUjqBn4OnBARv0llj0fE9sVmZmXg4QdmZmZWFYcCzwK3SbpA0iSyMbVm7qk1MzOzapG0PnAw2TCEicAlwJyIuLnQxKxQbtSamZlZZUnaBDgcODIifEexIcyNWjMzMzOrPI+pNTMzM7PKc6PWzMzMzCrPjVozMzMzqzw3as3MzMys8v4/mXgzOxx/8iMAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 864x864 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Correlation Matrix\n",
    "fig = plt.figure(figsize = (12, 12)) \n",
    "sns.heatmap(data.corr(), cmap='Blues')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Data Pre-Processing\n",
    "X = data.iloc[:,:-1]\n",
    "Y = data.iloc[:,-1]\n",
    "X_train, X_test, Y_train, Y_test = train_test_split(X, Y, test_size=0.2, random_state=42)\n",
    "\n",
    "print(X_train.shape)\n",
    "print(X_test.shape)\n",
    "print(Y_train.shape)\n",
    "print(Y_test.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\sklearn\\ensemble\\forest.py:245: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n"
     ]
    }
   ],
   "source": [
    "# Classification Model\n",
    "classifier = RandomForestClassifier() \n",
    "classifier.fit(X_train, Y_train) \n",
    "Y_pred = classifier.predict(X_test) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model Accuracy: 0.9996\n",
      "Model Precision: 0.963\n",
      "Model Recall: 0.7959\n"
     ]
    }
   ],
   "source": [
    "# Model Evaluation\n",
    "print(\"Model Accuracy:\", round(accuracy_score(Y_test, Y_pred),4))\n",
    "print(\"Model Precision:\", round(precision_score(Y_test, Y_pred),4))\n",
    "print(\"Model Recall:\", round(recall_score(Y_test, Y_pred),4))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Model Evaluation - Confusion Matrix\n",
    "labels = ['Valid', 'Fraud'] \n",
    "conf_matrix = confusion_matrix(Y_test, Y_pred) \n",
    "plt.figure(figsize =(6, 6)) \n",
    "sns.heatmap(conf_matrix, xticklabels = labels, yticklabels = labels, annot = True, fmt =\"d\"); \n",
    "plt.title(\"Random Forest Classifier - Confusion Matrix\") \n",
    "plt.ylabel('True Value') \n",
    "plt.xlabel('Predicted Value') \n",
    "plt.show() "
   ]
  }
 ],
 "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": "3.7.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}