3. UNSW-NB15 - Bayesian_enc - Classification before feature selection.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "3de84f4c",
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "train = pd.read_csv('UNSW_NB15_bayesian_enc_training_set.csv')\n",
    "test = pd.read_csv('UNSW_NB15_bayesian_enc_testing_set.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "63dbb67f",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "             dur  spkts  dpkts  sbytes  dbytes           rate  sttl  dttl  \\\n",
      "0       0.121478      6      4     258     172      74.087490   252   254   \n",
      "1       0.649902     14     38     734   42014      78.473372    62   252   \n",
      "2       1.623129      8     16     364   13186      14.170161    62   252   \n",
      "3       1.681642     12     12     628     770      13.677108    62   252   \n",
      "4       0.449454     10      6     534     268      33.373826   254   252   \n",
      "...          ...    ...    ...     ...     ...            ...   ...   ...   \n",
      "175336  0.000009      2      0     114       0  111111.107200   254     0   \n",
      "175337  0.505762     10      8     620     354      33.612649   254   252   \n",
      "175338  0.000009      2      0     114       0  111111.107200   254     0   \n",
      "175339  0.000009      2      0     114       0  111111.107200   254     0   \n",
      "175340  0.000009      2      0     114       0  111111.107200   254     0   \n",
      "\n",
      "               sload          dload  ...  ct_dst_src_ltm  is_ftp_login  \\\n",
      "0       1.415894e+04    8495.365234  ...               1             0   \n",
      "1       8.395112e+03  503571.312500  ...               2             0   \n",
      "2       1.572272e+03   60929.230470  ...               3             0   \n",
      "3       2.740179e+03    3358.622070  ...               3             1   \n",
      "4       8.561499e+03    3987.059814  ...              40             0   \n",
      "...              ...            ...  ...             ...           ...   \n",
      "175336  5.066666e+07       0.000000  ...              24             0   \n",
      "175337  8.826286e+03    4903.492188  ...               2             0   \n",
      "175338  5.066666e+07       0.000000  ...              13             0   \n",
      "175339  5.066666e+07       0.000000  ...              30             0   \n",
      "175340  5.066666e+07       0.000000  ...              30             0   \n",
      "\n",
      "        ct_ftp_cmd  ct_flw_http_mthd  ct_src_ltm  ct_srv_dst  is_sm_ips_ports  \\\n",
      "0                0                 0           1           1                0   \n",
      "1                0                 0           1           6                0   \n",
      "2                0                 0           2           6                0   \n",
      "3                1                 0           2           1                0   \n",
      "4                0                 0           2          39                0   \n",
      "...            ...               ...         ...         ...              ...   \n",
      "175336           0                 0          24          24                0   \n",
      "175337           0                 0           1           1                0   \n",
      "175338           0                 0           3          12                0   \n",
      "175339           0                 0          30          30                0   \n",
      "175340           0                 0          30          30                0   \n",
      "\n",
      "           proto   service     state  \n",
      "0       0.455718  0.547930  0.476401  \n",
      "1       0.455718  0.547930  0.476401  \n",
      "2       0.455718  0.547930  0.476401  \n",
      "3       0.455718  0.603213  0.476401  \n",
      "4       0.455718  0.547930  0.476401  \n",
      "...          ...       ...       ...  \n",
      "175336  0.762473  0.846186  0.912400  \n",
      "175337  0.455718  0.547930  0.476401  \n",
      "175338  0.762473  0.846186  0.912400  \n",
      "175339  0.762473  0.846186  0.912400  \n",
      "175340  0.762473  0.846186  0.912400  \n",
      "\n",
      "[175341 rows x 42 columns]\n",
      "0         0\n",
      "1         0\n",
      "2         0\n",
      "3         0\n",
      "4         0\n",
      "         ..\n",
      "175336    1\n",
      "175337    1\n",
      "175338    1\n",
      "175339    1\n",
      "175340    1\n",
      "Name: label, Length: 175341, dtype: int64\n"
     ]
    }
   ],
   "source": [
    "x1 = train.iloc[:,1:43]\n",
    "y1 = train['label']\n",
    "print(x1)\n",
    "print(y1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "d8a38484",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "            dur  spkts  dpkts  sbytes  dbytes           rate  sttl  dttl  \\\n",
      "0      0.000011      2      0     496       0   90909.090200   254     0   \n",
      "1      0.000008      2      0    1762       0  125000.000300   254     0   \n",
      "2      0.000005      2      0    1068       0  200000.005100   254     0   \n",
      "3      0.000006      2      0     900       0  166666.660800   254     0   \n",
      "4      0.000010      2      0    2126       0  100000.002500   254     0   \n",
      "...         ...    ...    ...     ...     ...            ...   ...   ...   \n",
      "82327  0.000005      2      0     104       0  200000.005100   254     0   \n",
      "82328  1.106101     20      8   18062     354      24.410067   254   252   \n",
      "82329  0.000000      1      0      46       0       0.000000     0     0   \n",
      "82330  0.000000      1      0      46       0       0.000000     0     0   \n",
      "82331  0.000009      2      0     104       0  111111.107200   254     0   \n",
      "\n",
      "              sload        dload  ...  ct_dst_src_ltm  is_ftp_login  \\\n",
      "0      1.803636e+08     0.000000  ...               2             0   \n",
      "1      8.810000e+08     0.000000  ...               2             0   \n",
      "2      8.544000e+08     0.000000  ...               3             0   \n",
      "3      6.000000e+08     0.000000  ...               3             0   \n",
      "4      8.504000e+08     0.000000  ...               3             0   \n",
      "...             ...          ...  ...             ...           ...   \n",
      "82327  8.320000e+07     0.000000  ...               2             0   \n",
      "82328  1.241044e+05  2242.109863  ...               1             0   \n",
      "82329  0.000000e+00     0.000000  ...               1             0   \n",
      "82330  0.000000e+00     0.000000  ...               1             0   \n",
      "82331  4.622222e+07     0.000000  ...               1             0   \n",
      "\n",
      "       ct_ftp_cmd  ct_flw_http_mthd  ct_src_ltm  ct_srv_dst  is_sm_ips_ports  \\\n",
      "0               0                 0           1           2                0   \n",
      "1               0                 0           1           2                0   \n",
      "2               0                 0           1           3                0   \n",
      "3               0                 0           2           3                0   \n",
      "4               0                 0           2           3                0   \n",
      "...           ...               ...         ...         ...              ...   \n",
      "82327           0                 0           2           1                0   \n",
      "82328           0                 0           3           2                0   \n",
      "82329           0                 0           1           1                1   \n",
      "82330           0                 0           1           1                1   \n",
      "82331           0                 0           1           1                0   \n",
      "\n",
      "          proto  service     state  \n",
      "0      0.762473  0.54793  0.912400  \n",
      "1      0.762473  0.54793  0.912400  \n",
      "2      0.762473  0.54793  0.912400  \n",
      "3      0.762473  0.54793  0.912400  \n",
      "4      0.762473  0.54793  0.912400  \n",
      "...         ...      ...       ...  \n",
      "82327  0.762473  0.54793  0.912400  \n",
      "82328  0.455718  0.54793  0.476401  \n",
      "82329  0.000000  0.54793  0.912400  \n",
      "82330  0.000000  0.54793  0.912400  \n",
      "82331  0.762473  0.54793  0.912400  \n",
      "\n",
      "[82332 rows x 42 columns]\n",
      "0        0\n",
      "1        0\n",
      "2        0\n",
      "3        0\n",
      "4        0\n",
      "        ..\n",
      "82327    0\n",
      "82328    0\n",
      "82329    0\n",
      "82330    0\n",
      "82331    0\n",
      "Name: label, Length: 82332, dtype: int64\n"
     ]
    }
   ],
   "source": [
    "x2 = test.iloc[:,1:43]\n",
    "y2 = test['label']\n",
    "print(x2)\n",
    "print(y2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "fd392215",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1    119341\n",
       "0     56000\n",
       "Name: label, dtype: int64"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train[\"label\"].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 258,
   "id": "fd42c621",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAaAAAAFNCAYAAABG7LsFAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuNCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8QVMy6AAAACXBIWXMAAAsTAAALEwEAmpwYAAAtRElEQVR4nO3debxd0/3/8ddNTBGJIa4Kaio+dCCGGEu1tJqitIqvpJTWVFX9tugkKVp0wrdo08HQaIP6lbZfRUpRlSJp1VCCN60khKg036aREBL3/v5Y67Bz3OGcm3uyb859Px+P+7jnrL323mufs/f+7DWcvVva29sxMzNb3gaUXQAzM+ufHIDMzKwUDkBmZlYKByAzMyuFA5CZmZXCAcjMzEqxUlcTI2JTYDpwrKTLC+mnAe+UdHRvFCIiZgAfk3Rfbyyvm3UNBSYBawHjJP2q0evspjxnAetKOrkXlnUZ8AtJty1zwXq2/juB70u6roz5a1j+psAjktaoc74RwPXAPOAQSTN6u2x9WURMIH1u50fEg8Dekub10rI3BH4IHAScSQfHQkQ8Apws6c58rpgs6cjC9J2A6yRtmt/vD4wFVied46YBX5A0KyIeB74k6X9z3v2A3wEnSPpJTtsZuAEYLqm9sJ6zgJOAbSU930X5XgFeJl3gDwQuqiy7MM8ngY9IOrCQdj2wHbAgJ/1B0uer5tubdP5SThoIzAHOq+W4j4hLgR9J+mt3eesREWsCv5b0vvx+InCOpMe7mq/LAJS1ARdExJ8kqdvcfd8I4C2Stii7IL1N0rFll6FJfZh0Muj3n6+kEb28yEuBMyW1R0St8xwaEbdImlg9ISI2AK4EdpQ0M6edAfw/YHfSyfu9wP/mWQ4AfksKgJUg8T7g5mLwKRgK/Cwi9utkOsCYysV0RLwVeCIiJkl6JiLWAc4DxgB/rJpvN2AnSc91s/3/KH4PEbEdcEtEHCRpajfzvh/4cTd5emJtYOfC+68BV0fEbl18TjUFoJeBCwoLe7U4sXh1VP0+Xw1cTfpC1wa+A+wB7AgsBj5c+LA/kz/IVYELJF2Rl3cg6WpmFeAl4DRJ9+arkd2ADYCHJH28qlwHk66oBgAvAl8A/gNcAWyYr+R2k/RyYZ6P5nW1Aa8Bp0u6KyJ2zWVfFRgO/F7Sp/IV9R3A7/M2rUT64E8AtgbuA44ANibtbL8DdgFaSFdMk6vKvCHw/Zx/ZVJt5ryIWAm4JH92i4GngGMkLaia/848/33A7cDNeX1rA1+U9GuqRMTuwLeBwXmbz5Z0Y0QMJl2ZbgkMy5/haEmKiPWBH+VtbCNdUV2cF3lQRJwOrA/cBhwnqa1qnV3NX8nzVdJJYVAu22mSfh0RWwOXA6vlz/EySeM7S6/eXmBArilW9sFTJE3J6zwDOIS0z8wgXe2+N/8fGBGDJI2JiHGk73UJ8ATpu3w+f/7/l7frh8DPgIuAd5G+z9tJ+9SSqm2dAKwDvA24ERiXv5P3kK5wH8jlnB8RnwZOBF4FFpGu3B+NiAOAr5KOk/WAKyWNy1fM3wSeBgJYCHwLOCW/v17S53O+bwMzc/lfBo6W9FhVWduBVtKJ+yP5+9uSdGx+QtJjEbEF6ThbB5idv4+JkiZULWsXYD1Jf+nge+rKGcAlEXG3pOlV09bNn0Gxlvs94KH8ehLpWK44ANgPmBoRgyUtBPYh7Z8dmQjsCpwKnF9DWdcmfeaVY/Uw4DngNKBY+9kMGAJcGhEbk47hUyX9X3crkPRQRFwMfB74ry7OV+eSzpdXRcRRpO+lo3ydnm86O18APwUG5fPqjpKeioh5pIu3SrB/k1r7gM4lfYDn1Zi/aDVJu5JOzD8hVUe3A54Bji7ke1nSDqQI/c2IeEdEbJnX+SFJ2wPHA7/KJ0eATYDtOwg+W5N2oEPyur5G+hBmA8eSryCKwSf7LnCSpJ1IJ4G9c/rngK9J2gV4O/DhiNgxT9sMuCnPcy/phHME8A5gT9LOCjkI5SuXLwPXRsTKVev/OXCFpB1JVxP7RsRhpEC7N7BdnvYUsO2bP+qlbA7cImnnvL7vVWeIiLVJO86R+bM/CPhhPgBGAfMk7SZpK+AvQKVpZDzwhKStc9mOzycdSAfR7sA2eRl7dFC2ruYnIjYB9iU19WxLOuF8PU8+Hfht/hw+BOwVEQO6SK82iHSgbU+62PhlRKySD8h3ATvn7+hmUhC7irQvXZuDzzF5u0bmsj0CTCgs/9+S3i7pEuB/gL/mMm1POjl+oYMyAawu6R2SvkT6vpaQDuTtSCesb0XEQNL3+EFJI0nH07sjooV0QvxE3g93Bb4SEevmZY8EvpW3az7wFWB/YAfShd8GOd9OwCV5u35K2h+78h7gs5LeCUzN5SbPd01OP4X0HXfkUFLArdcfSfvQ1flk+TpJfyPVqh6IiEdzk9OBpIu/yrxvi4h1IuJdpO/riVz+D0TEqqRj7/edrHsR6fgeFxE7dJLnqoh4MDf3PUC6wPp3Lt+PJH2d1ExXtB7pgu1EUivNAlIQr9VDpP0XOjlfSTqDtC+NyTWlzs5rHZ5vujlfHEM6h4+Q9Foux63AR7sqdE0BKF/Bfhw4JiLeX8s8Bdfn//8Anpf0UOH9OoV8P87reo5U8H1IwWg4cHuOrFeRrrgqJ6sp1VeT2fuA2yU9lZd5B/AC6aq3K78Afp2vkCs1NoBPAGvlq/LxpJNY5QprMakKX9mmeyTNl7SI9GVXtvHfkq7O5ZlEunp4PYjkoPoe4Bt5W6eQgtYI4OGcf2pEfIN01XpPN9uymHQSBbifpT/rit1In+9v8jpvBtpJbdzXARMi4rMRcRFph6xs877k5gpJ/5H0Tkl/z9OulfSapJeAJ0kHVrWu5kep6eQoYExEfIt0UFbW/WvgixHxK9LOfUrePztLrzZP0rV5PbfmtK1JV8K7Avflz+KzpBpCtVHAT/OVMqQLjn0iYpX8vlirPQA4IS/vr6QT27vo2J+q5juIdBJ9EDgYeHs+sH8J3BMR3yf1SV2u1MRxILBjRJwJXEi6uq1cqE2X9EB+/Q9Sc+Krkv5FCkiVfeOhQq38CmD7iBjWSXkhBddZ+fX9wDr5JLUzcBlArkHd3sn8WwN/L7zv6PuCdJ56rSrtzLyNZ1VnlnQqab8eR6rJfRf4Y0QMlPQKKQjtRfrMKgHwRuADpGPiPknzOykLkh4mXbxcXbgYLhqTT8Rbk85VoyPiiM6Wl5c5VdJHJD2Tv+ezgP0L+1V32km1UOj6fFXUWb7Ozjedni86KdN0Oj6GXlfzKDhJz5Calq4kXclVtJN2hIrqD6wY6Rd3sYriDjYg5x1ICiQjKn+kk8QjOd8COjYwl6toAKkZpFP5CuHdpOrv0cBdedJdpKvqx0lX4s/yxja/qqXbODvbxupAWX1QDczL3L1qW89T6vDdjlRtf41Uezqpq23J5aoc0NXfUXGdj3Xw+d6Sm3ouJ+3UVwPXFJaxhMLnGxGbRxrcAUtvf2fr7Wp+8pXlvaT29ltJVf4WgFzd35LUpr898HBEbNRZegfrrj6RFfe1bxc+h53ouPZWvW8NIDW9VrZzQVXeQwvL3IU3apHVquf7XGG+nYGP5e3/OOnE+XdSjeOafBJ8gFSjuZ9UG1xcKFP11XYt+2hl3urPq6jYglD5rivLKH7vnS2jnaXPQf8iNfdWewswt5iQLzxHA58hBRMAIuLDEXGMpLmSrpd0Cqk2/g7SfgGpGW4vUqCvBKCbSMf++/LrLuUa7t9JFyBd5ZtOGtCwV1f5ImLPiPhwIamFN7oCajGSFDig6/NVUYf5ujjfdHq+6KRMi7srf13DsPNV8STgvwvJc0gHa6UD8D31LLPg6LyMjUlXyLfnvw/kJjUi4kPA30iRuiu3A/tFxOZ5vvcBbyVVszsUEStF6rNaXdKPyKNdIuItpC/3S0oj5jYiXdUMrHP7WiPig3ldB5K+nMoOQ77imkJuoomItYC7SX0qB+RtukfSWaS+hZF1rr8jU4AtI2KvvM4RpFrLhqR28QlKox9FOulVtvk2UpW7MvrldtLJv1bdzb8X6Sr0QtLV6sGVdUfE1cDhkn5B+o7mk5pUOkzvYN3D8udZ+R5eztt8C3BsIRB+nY6boH4HfLJw5XsKcFe+sq52C/D5iGjJTTs30HkAqp7v5Nw0OIDUpPTNiFg3Ip4B5kr6HukqfCTpsxsKjJX0W1JtdVXq30dHRETlavZ40v42r54FSHqRtN9Wvt/NSK0ZHXVEi6W/o9tItcmtKwm5afQ/pJNk9bqeIn3+xa6BF0mf1dsLaZuTAuM/8vtJpNrOJqSm5UqggLSvdRuAsmNITZmdDmgqtGz8uZtlrUHq16rURk8njezrNgBFGrX3aeCifN7o6ny1BFi5q3xdnG+6Ol8syfMWA91mdPC9FfXkd0CnkDoqKy4BhkeESO2Dd/RgmQCrRcT9pGrdZyU9IelR0oHwi4h4CPgGaeBCZzUfAPJ8J5H6ix4hdboeKOk/XcyzhBRYr87l+CXwSUn/JHXi3p+X9WXSAVbvKLpFwJF5O84ADu5g5xoN7BoRD5OC5TVKfRCTSENJH4mI+0h9LGfXuf43kTSH1On+3Vyun5Pad2eQOlhPiIi/kZqV7ueNbT4Z2CZPuxv4puob1tnd/NcA60bEY8CjpNrBOhExhLQPjMnlnUpqeruri/RqLwCH5CaEr5D6CZeQmoxuBKZExDRSs8LRHcx/OelE+edcvh1II5o6cgqpGexh0oXTwyzdAd6Zb5AGQTyQt7+F1CH9L+AcUpP0X0n79XF52TcCj+cyHZjnq3cffR44N+9/BwNHdp29U0cBh+Xv4gekppiXOsh3HfDBypvcF3M88PPch/IY8F/AqE6aU5H087ycyvs/kPavKyPiybyM75H6kSv9ME+RWkNuqWq9mAQMlmob7ZuPn0/w5lafSh/QA6Tv8EZJP+1mWZOAi4G787n0bXR+sfK2vPwH87nq26QBQg/lC4auzle/Ig2k2LmLfB2eb7o5X8wmBdlphWbbD5LOo51q8eMYGi96+PsTs+Ul0ii47ysNHFjWZZ1B6jd4PNdw/0YKIo92kPcW0u/xuqsh2AokIt5G6rPfTV0Mw/adEMystz1B6jd4gDS44lsdBZ/sBOBrVU03tuI7h3QDgy5rOK4BmZlZKVwDMjOzUjgAmZlZKWq5FU9/syppyOFsah+Db2bWnYGkH3L+hTf/NqtfcgB6s5Es/Wt2M7PetCdL3/mi33IAerPZAP/+90La2vr2AI1hw9Zg7twufxJl1i+sCMfCgAEtrL32YMjnGHMA6shrAG1t7X0+AAErRBnNlocV6Fhw037mQQhmZlYKByAzMyuFA5CZmZXCAcjMzErhAGRmZqVwADIzs1I4AJmZWSkcgMzMrBQOQGZmVgoHIDMzK4UDkJmZlcIByMzMSuEAZGZmpXAAMjOzUiyXxzFExFDgHuAASTMi4njgFKAduA84QdKrETECuAwYCtwFnChpSURsDEwE1gMEjJG0ICLWAq4CNgfmAIdJej4iVgEuB3YCXgZGS3p8eWyrmZnVpuE1oIjYhfT0v63y+62A04HdgW1zGT6Ts08ETpa0FdACHJfTxwPjJW1NCljjcvo5wGRJ2wCXAhfl9FOAhTn9v4EJDdo8MzProeXRBHccKcA8l9+/Apwkab6kduBhYOOI2AQYJGlKzjcBODQiVgb2Aq4rpufX+5NqQADXAKNy/tfTJd0FtOZalJmZ9RENb4KTdCxARFTezwRm5rRW4GTgaGADln5U7WxgI2BdYL6kJVXpFOfJTXXzgdYulvV0reUeNmyNWrOWqrV1SNlFMOsTfCyseEp7JHdEbAhMAi6XdGdE7EHqE6poAdpItbTqZ+22FfIUdTZPS2Gemsydu6DPP+K3tXUIc+a8WHYxzEq3IhwLAwa0rDAXtstLKQEoIrYGbgEulnRBTp4FDC9kW5/UbPcCsGZEDJT0Ws5Tac57NuebFRErAUOAuYVl/aNqWQ0zePBgVl99+Q8qXN5XfS+91MbChQuX6zrNrDkt9zNmRAwBbgXGFoJPpWluUa4JARwJTJK0GJgMHJ7TjyLVnABuzu/J0yfn/K+nR8S7gUWSam5+64nVVx9ASwtN/1dGkDWz5lRGDehY4C3AqRFxak67QdLXgDHApXnY9v3AxXn6ScCVETGW1I9zRE4fB0yIiGnAvDw/wCXAj3P6K6RgZmZmfUhLe3vf7ucowabA9Hr7gFpbh9BS3SPVhNrb6fNt7db/rGB9QJsBM8otTd/g9hQzMyuFA5CZmZXCAcjMzErhAGRmZqVwADIzs1I4AJmZWSkcgMzMrBQOQGZmVgoHIDMzK4UDkJmZlcIByMzMSuEAZGZmpXAAMjOzUjgAmZlZKRyAzMysFA5AZmZWCgcgMzMrhQOQmZmVwgHIzMxK4QBkZmalcAAyM7NSOACZmVkpHIDMzKwUDkBmZlYKByAzMyuFA5CZmZXCAcjMzErhAGRmZqVwADIzs1I4AJmZWSkcgMzMrBQOQGZmVgoHIDMzK8VKy2MlETEUuAc4QNKMiNgXuBAYBFwraWzONwK4DBgK3AWcKGlJRGwMTATWAwSMkbQgItYCrgI2B+YAh0l6PiJWAS4HdgJeBkZLenx5bKuZmdWm4TWgiNgF+BOwVX4/CLgCOAjYBhgZEaNy9onAyZK2AlqA43L6eGC8pK2B+4BxOf0cYLKkbYBLgYty+inAwpz+38CERm2fmZn1zPJogjsO+AzwXH6/M/CkpOmSlpCCzqERsQkwSNKUnG9CTl8Z2Au4rpieX+9PqgEBXAOMyvlfT5d0F9Caa1FmZtZHNLwJTtKxABFRSdoAmF3IMhvYqIv0dYH5OVgV05daVm6qmw+0drGsp2st97Bha9Satd9pbR1SdhHM3sT75YpnufQBVRkAtBfetwBtdaST0yt5irpbVs3mzl1AW1v1ajvXn3b+OXNeLLsIZktpbR3S5/fLAQNafGFbpYxRcLOA4YX365Oa5zpLfwFYMyIG5vThvNGc92zOR0SsBAwB5naxLDMz6yPKCEBTgYiILXJQGQ1MkjQTWBQRe+R8R+b0xcBk4PCcfhQwKb++Ob8nT5+c87+eHhHvBhZJqrn5zczMGm+5ByBJi4CjgeuBR4HHeWOAwRjgfyLicWAN4OKcfhJwfEQ8CuwJjM3p44BdI2JazvOZnH4JsGpOv5gUzMzMrA9paW+vvZ+jn9gUmN6TPqCW6h6pJtTe7j4g63tWsD6gzYAZ5Zamb/CdEMzMrBQOQGZmVgoHIDMzK4UDkJmZlcIByMzMSuEAZGZmpXAAMjOzUjgAmZlZKRyAzMysFA5AZmZWCgcgMzMrhQOQmZmVwgHIzMxK4QBkZmalqDkARcTqEbFrfn1iRFweERs3rmhmZtbM6qkB/RQ4KCJGAl8EngEubUipzMys6dUTgDaX9BXgQGCCpLOAdRpSKjMza3r1BKCV8//9gDsiYiDpsdlmZmZ1W6mOvPdExKPAEuAe4HbgtoaUyszMml49NaDPAscDe0pqA84HPteQUpmZWdOrOQBJeg1YHzg9IlYHhuZAZGZmVrd6hmF/Gfg0cBgwCDgzIsY1qmBmZtbc6mmC+y/gQ8BCSXOBXYHRDSmVmZk1vXoC0GJJr1TeSJoHLO71EpmZWb9Qzyi4ZyJif6A9IlYFTgNmNqZYZmbW7OoJQCcDPwe2BRYCU4AxjSiUmZk1v3pGwT0HHAQMJd0B4aOSXAMyM7MeqWcU3GHAA5JeIg3HfiQiDmxYyczMrKnVMwjhDOC9AJKeAHYEzm5EoczMrPnVE4AGSppVeSPpmTrnNzMze109gxBeiIgTgMuBduATwD8bUiozM2t69dRgTiDdC25R/jseOKkRhTIzs+ZXcw1I0pPAjhGxNrBE0ouNK5aZmTW7mgNQRLwFOJE0BLslIgCQdEpPVhwRHwe+kt9OknRaROwLXEi619y1ksbmvCOAy0hDwO8CTpS0JD8SfCKwHiBgjKQFEbEWcBWwOTAHOEzS8z0pp5mZNUY9TXATgR2AfwNzC391y3fTvhh4D7AdsGce0n0F6bdG2wAjI2JUYd0nS9oKaAGOy+njgfGStgbuAyo3Rz0HmCxpG9Jjwy/qSTnNzKxx6hmEsFE+ofeGgaTgN5h0V4WVgfnAk5KmA0TERODQ/BC8QZKm5HknAGdHxGXAXsDBhfQ/Al8C9s/TAK4BfhARK0vyvevMzPqIegLQzIgYLGnhsq5U0ov5UQ6PAy+RAscGwOxCttnARl2krwvMl7SkKp3iPLmpbj7QCjxXaxmHDfPTxjvT2jqk7CKYvYn3yxVPPQFoNvBgRNwJvFxJ7EkfUERsC3wS2AT4D6mJbSvS8O6KFqCNVFOqJZ2cXslT1FKYVpO5cxfQ1la9+M71p51/zhyPP7G+pbV1SJ/fLwcMaPGFbZV6+oBmkALFLJaxDwjYD7hd0gv5EQ8TgL2B4YU865NqLLM6SX8BWDMiBub04bxRw3k25yMiVgKGLENZzcysAeoZhv2m2+5ExOAervch4Dt5/peAA4GpwJiI2AKYTnrY3RWSZkbEoojYQ9LdwJGkUXOLI2IycDhwNXAUMCkv/+b8/rw8fbL7f8zM+pZ6hmEfBHwdWIPUpDWQNCS77rYnSbdGxPbAX0kPtfszcBbwe+B6YDVSELkuzzIGuDQihgL3k0bQQfoh7JURMRZ4Gjgip48DJkTENGAefmyEmVmf09LeXls/R0Q8CYwl/Rbo28BHSIMATm9c8UqxKTC9J31ALdU9T02ovd19QNb3rGB9QJuRujT6vXr6gBZKupb0ILpFwKeBAxpSKjMza3r1BKBF+VHcfwdGSGrjzaPQzMzMalLPMOwbgJtId8G+NyL2BP7VkFKZmVnTq+eR3OcBn5T0LOnuA3cBhzSoXGZm1uTqeST3JElPA0i6X9JFwG8bVjIzM2tq3TbBRcR1pLsUvC0i/laYtDLwSqMKZmZmza2WPqDTSEOTLwU+W0hfAjzagDKZmVk/0G0AkjQDmBERkUe+EREbAG+T9O8Gl8/MzJpUPcOwj4+IqyNiXdLdCC6LiG82qFxmZtbk6glAxwKfBw4lDcl+B/D+RhTKzMyaXz0BqF3SP4F9gdvyc3gGdjOPmZlZh+oJQK9ExBdJj9H+fUR8mvQ0UzMzs7rVE4A+RRqO/Yk8+ODdpGY5MzOzutXzPCBRCDiS/IgDMzPrsXqfB/Q90jOAXn/wgKShvV8sMzNrdvXcjPQ7wKmkIdi+C7aZmS2TegLQPEm/alhJzMysX6lnEMLUiBjVsJKYmVm/Uk8N6EPAyRHxKvAqqR+o3X1AZmbWE/UEoH0aVgozM+t3ankcw/sk3QHs2EmWmb1bJDMz6w9qqQEdAdzB0o9iqGgHPDDBzMzqVsvjGI7L/9/b+OKYmVl/Uc8oODMzs17jAGRmZqXoNgBFxMH5/6oNL42ZmfUbtdSAvpH/39vIgpiZWf9Syyi4+RHxBLBhRPyteqKkbXu/WGZm1uxqCUAfBLYHLqfjodhmZmZ1q2UY9ovAXRGxP/Ac6QepKwNT8zQzM7O61TMKbk3gCdIzgS4EZkbE7o0olJmZNb96AtAFwBhJ2+d+n4+RApGZmVnd6glAQyT9ofIm3x9u9d4vkpmZ9Qf13A27PSI2kTQTICI2BV7r6Yoj4kDgTGAwcKukz0XEvqRa1SDgWkljc94RwGXAUOAu4ERJSyJiY2AisB4gUg1tQUSsBVwFbA7MAQ6T9HxPy2pmZr2vnhrQ14EpEfHziPgZMBU4pycrjYjNgR8BBwPbAjvkh91dARwEbAOMLDwAbyJwsqStSM8hOi6njwfGS9oauA8Yl9PPASZL2ga4FLioJ+U0M7PGqTkASfoNsDdwD/BnYG9J1/dwvR8h1XBmSVoMHA68BDwpabqkJaSgc2hEbAIMkjQlzzshp68M7AVcV0zPr/cn1YAArgFG5fxmZtZH1NMEhySRmrqW1RbAqxFxA7AxcCMwDZhdyDMb2AjYoJP0dYH5OVgV0ynOk5vq5gOtpGHkNRk2bI06N6n/aG0dUnYRzN7E++WKp64A1Mvr3YtUo1oA3AC8THq+UEUL0EaqpdWSTk6v5ClqKUyrydy5C2hrq1585/rTzj9njn/+ZX1La+uQPr9fDhjQ4gvbKmXdDft54DZJcyS9DPwa2BcYXsizPqnGMquT9BeANSNiYE4fzhs1nGdzPiJiJWAIMLcxm2JmZj1RcwDKAw96y43AfhGxVg4go0h9ORERW+S00cCkPOpuUUTskec9MqcvBiaT+o8AjgIm5dc35/fk6ZNzfjMz6yPqqQGNiIjqpq0ekTQV+A7wJ+BRYCbwQ+Bo4Pqc9jhvDDAYA/xPRDwOrAFcnNNPAo6PiEeBPYGxOX0csGtETMt5PtMb5TYzs97T0t5eWz9HRPyONGBgCqnfBgBJpzSmaKXZFJjekz6gll4Jz31be7v7gKzvWcH6gDYDZpRbmr6hnkEI9+JnApmZWS+pOQBJOjsiBpGGUE8DVpP0UsNKZmZmTa3mABQRu5BGqy0BdgceiogDJd3TqMKZ2Ypn8ODBrL768h9gu7x/CvHSS20sXLhwua6z2dSzl5xPGio9V9Is0mg03+LGzJay+uoDaGmh6f/KCLLNpp5PcHVJj1beSLqZ8n7IamZmK7h6AtDiiFibfPeBiIjGFMnMzPqDemow5wB/BIZHxDXAB4DjG1IqMzNrevWMgrsx/xD0/cBA4OuSHmtYyczMrKnV24u2Min4LM5/ZmZmPVLPveCOAf4AjCTd9mZyRBzSqIKZmVlzq6cP6AvA9pJmA+THYd9IunebmZlZXeppgnu1EnwAJD2Nm+HMzKyHuq0BRcQO+eVDEfF94MfAa6Q7V9/duKKZmVkzq6UJrrqJbf/C63ag2e6GbWZmy0G3AUjSZsujIGZm1r/UczPS9UnNbusU0yV9sZfLZGZm/UA9gxBuAHYGWqr+zMzM6lbPMOxVJH20YSUxM7N+pZ4a0F8j4p0NK4mZmfUr9dSA7gYejIjZFH7/I2nzXi+VmZk1vXoC0OnAaOAfDSqLmZn1I/UEoHmS/l/DSmJmZv1KPQHojog4n/TD1FcqiZLu7/VSmZlZ06snAI3O/4t3wG4H3AdkZmZ1q+eBdL4jgpmZ9Zp67oTwhY7SJV3Ye8UxM7P+op4muHcVXq8CvAe4vXeLY2Zm/UU9TXDHFN9HxAbA5b1eIjMz6xfquRPCUiQ9B2zae0UxM7P+pKd9QC3ATsALvV4iMzPrF3raB9QOPE26O4KZmVndetwHZGZmtiy6DUAR8VNSjacj7ZI+tSwFyHdXWFfS0RGxL3AhMAi4VtLYnGcEcBkwFLgLOFHSkojYGJgIrAcIGCNpQUSsBVxF+pHsHOAwSc8vSznNzKx31TII4RFgWtXfP4EjgN2XZeURsQ/wifx6EHAFcBCwDTAyIkblrBOBkyVtRep/Oi6njwfGS9oauA8Yl9PPASZL2ga4FLhoWcppZma9r9sakKQLiu9zLeVKUg3jlJ6uOCLWAc4FzgO2Iz1t9UlJ0/P0icChEfEoMEjSlDzrBODsiLgM2As4uJD+R+BLwP55GsA1wA8iYmVJrz9GwszMylXPKLiVgG8CR5OawK5fxnX/GDgDeGt+vwEwuzB9NrBRF+nrAvMlLalKX2pZualuPtAKPLeMZTYzs15SUwCKiC1JNYkFwPaSZi3LSiPiWOAZSbdHxNE5eQBL9zW1AG11pJPTK3mKWgrTajJs2Br1ZO9XWluHlF0Esz7Bx8KyqWUQwjHABcAFks7tpfUeDgyPiAeBdYA1gE2A1wp51ifVWGYBwztIfwFYMyIGSnot56nUcJ7N+WblmtsQYG49BZw7dwFtbZ2NvXiz/rQjzpnzYtlFsD7Mx0LHBgxo8YVtlVoGIVxOGn325YiYX/h7MTdt1U3S+yW9U9II4GvADcAoICJii4gYSHr8wyRJM4FFEbFHnv3InL4YmEwKZgBHAZPy65vze/L0ye7/MTPrW2ppglsuj2GQtCg3x10PrEYKItflyWOASyNiKHA/cHFOPwm4MiLGkn4Ye0ROHwdMiIhpwLw8v5mZ9SEt7e21NzP1E5sC03vSBNdS3fPUhNrb3QRnXfOx0LFCE9xmwIwGFWuF0uObkZqZmS0LByAzMyuFA5CZmZXCAcjMzErhAGRmZqVwADIzs1I4AJmZWSkcgMzMrBQOQGZmVgoHIDMzK4UDkJmZlcIByMzMSuEAZGZmpXAAMjOzUjgAmZlZKRyAzMysFA5AZmZWCgcgMzMrhQOQmZmVwgHIzMxK4QBkZmalcAAyM7NSOACZmVkpHIDMzKwUDkBmZlYKByAzMyuFA5CZmZXCAcjMzErhAGRmZqVwADIzs1I4AJmZWSkcgMzMrBQOQGZmVoqVylpxRJwJHJbf3iTpixGxL3AhMAi4VtLYnHcEcBkwFLgLOFHSkojYGJgIrAcIGCNpQUSsBVwFbA7MAQ6T9Pxy2zgzM+tWKTWgHGg+AGwPjAB2jIgjgCuAg4BtgJERMSrPMhE4WdJWQAtwXE4fD4yXtDVwHzAup58DTJa0DXApcFHDN8rMzOpSVhPcbOBUSa9KWgw8BmwFPClpuqQlpKBzaERsAgySNCXPOyGnrwzsBVxXTM+v9yfVgACuAUbl/GZm1keU0gQnaVrldURsSWqKu4QUmCpmAxsBG3SSvi4wPwerYjrFeXJT3XygFXiu1jIOG7ZGHVvUv7S2Dim7CGZ9go+FZVNaHxBARLwDuAk4HVhCqgVVtABtpFpaew3p5PRKnqKWwrSazJ27gLa26sV3rj/tiHPmvFh2EawP87HQsQEDWnxhW6W0UXARsQdwO/BlSVcCs4DhhSzrk2osnaW/AKwZEQNz+nDeqOE8m/MRESsBQ4C5jdkSMzPribIGIbwV+A0wWtIvcvLUNCm2yEFlNDBJ0kxgUQ5YAEfm9MXAZODwnH4UMCm/vjm/J0+fnPObmVkfUVYT3GnAasCFEVFJ+xFwNHB9nnYzbwwwGANcGhFDgfuBi3P6ScCVETEWeBo4IqePAyZExDRgXp7fzMz6kJb29tr7OfqJTYHpPekDaqnueWpC7e3uA7Ku+VjoWKEPaDNgRoOKtULxnRDMzKwUDkBmZlYKByAzMyuFA5CZmZXCAcjMzErhAGRmZqVwADIzs1I4AJmZWSkcgMzMrBQOQGZmVgoHIDMzK4UDkJmZlcIByMzMSuEAZGZmpXAAMjOzUjgAmZlZKRyAzMysFA5AZmZWCgcgMzMrhQOQmZmVwgHIzMxK4QBkZmalcAAyM7NSOACZmVkpHIDMzKwUDkBmZlYKByAzMyuFA5CZmZXCAcjMzErhAGRmZqVwADIzs1I4AJmZWSkcgMzMrBQrlV2ARoqI0cBYYGXge5J+UHKRzMwsa9oaUERsCJwLvBsYARwfEW8vtVBmZva6Zq4B7QvcIen/ACLiOuBjwNe7mW8gwIABLXWvcJNN6p5lhdSTz8b6Fx8LXeYd2JDCrICaOQBtAMwuvJ8N7FzDfMMB1l57cN0rnDGj7llWSMOGrVF2EayP87HQpeHAP3q5KCukZg5AA4D2wvsWoK2G+f4C7EkKWK81oFxm1j8NJAWfv5RdkL6imQPQLFIgqVgfeK6G+V4B/tSQEplZf+eaT0EzB6DbgLMiohVYCBwCHF9ukczMrKJpR8FJehY4A/gD8CBwtaQ/l1ooMzN7XUt7e3v3uczMzHpZ09aAzMysb3MAMjOzUjgAmZlZKRyAzMysFA5AZmZWCgeg5Sgiuh1yGBF/6KV1XRYRO/XGsswaKSI2jYj2iHh/VfqMiNh0OZXBw4FL0Mw/RF1R7d0bC5F0bG8sx2w5WQxcGhHvkvRi2YWx5cMBqAQRsTfwVeAlYBvgYWA0cH6ePlXSLhExB7iPdP+o04EzJO2d80wA7gR+BVxDutUQwNmSboiIO4GzJN0ZEV8FPk66t92twBeBtwK/Bh4Btgf+CRxauXu42XL2HPB74AKq7ljSxf77O+BfwMvAVcD+wDDSjYh/DGwCvA+YC4yStCgizgX2AdbJ6zxc0j8bvXHWMTfBlWd34GRSANoY2E/SKQCSdsl51gW+LWkE6QqxIx8BZkjaEfgUS9//jogYBXwY2IkUaLYATsyTtwMulPROYB4wpjc2zKyHTgX2q2qK62r/DeDjkir5dwYOBvYDLgQmSdo2T9svIrYAtgZ2l7QV8DQpsFlJHIDK84ikWZLagMdIV2QdmdrNcu4BDo6I3wAjgW9UTd8HuEbSS5KWAFfkNIAXJD1QKU8XZTBrOEnzgeNITXFDcnJ3+++MwiLuljRf0sz8/vb8fyawtqS/k4LcsRFxAbAb4GeLlMgBqDyLCq/bSY+LeBNJL3eSZ+U8/UnSVd1VpNrPnyOi+L1Wf8ctvNH0WlMZzJYXSbfyRlMcdL3/vlw17dWqZS0pvo+IHUlNeAOA60hN0N7nS+QA1Pe8FhEd9c39C9g8IlaLiHXITW0RcTKp3+eXwEnAesDQwnx3AEdExKC83GNIN2g166tOJTWjDad399/3AHdK+hHwBHAAfjppqRyA+p7/BR6KiNWKiZKmATcB04BfApPzpJ8BEREP57TTJc0rzHcjcCNpMMM0Urv3JQ3eBrMeKzTFrULad3tr/70W2C4fK3fmZW62rOW1nvPdsM3MrBSuAZmZWSkcgMzMrBQOQGZmVgoHIDMzK4UDkJmZlcL3grOmEhEDgc+R7q23Emko72+Br0l6Jd9D7xFJ5zewDO1Aq6R/dZHnTuD7kq6rY7lHAx+TdMAyF9KsD3ANyJrND0m3WNkn30NvJOmeYZeVWSgzezPXgKxp5GfHjAGG5x8zImlhRJwI7NFB/k8CJ5BqSesA35L0w4hYn/QD33Vz1pskjessvYvyDCYFxC1Jd2l+ERgtSTnLRyLiy8DqwFWSzs3z7Q58GxhMugP02fkHxWZNxTUgayY7AtMqwadC0vOSri+mRcQapF/bf0jS9sDhwHfy5OOApyTtQLrl0ZYRsWYX6Z0ZBcyTtFu++/JfSHdArxgK7Jr/Ph4RoyJibeCnwJF5PQcBP4yIjev+NMz6ONeArJm0UeNFlaQFEXEAsH9EbAmM4I07I/8OuDmf9G8DvizpPxHRYXoX67guIp6KiM+SHiOwN3BvIctl+YaZ8yPiOuD9pJtjDgd+ExGVfO3Atpg1GdeArJlMBbYp3MofgIjYMCJuiohBhbSNgAdJDy37EzC2Mk3SX0j3CPsJsCnpDuM7dpbeWWEi4tPA5aQHD15NenBg8e7LrxVeDyA982kg8JikEZU/Ug3plpo/BbMVhAOQNQ1Jz5EeS3FFRAwFyP/HA3MLj7aA9ICzOcA5pFv0H5DzD4yIbwHjJP2GNKJuGvDOztK7KNJ+wARJlwMCDmTpuy8fFREtudntMFLNawqpaW+vXJ4RwJPAhj35TMz6MjfBWbM5CRgH3BMRS4BVgd8AZ1bluxX4JCkwtAF/JAWkLYDvAVdGxCPAK8BDwC+AtTtJ78z5wE8i4lOkms+9wLsK0/8D/BUYBFwi6Q8AEXEI8N18R/QBpP6gGYUmObOm4Lthm5lZKdwEZ2ZmpXAAMjOzUjgAmZlZKRyAzMysFA5AZmZWCgcgMzMrhQOQmZmV4v8DCwWzwBpinukAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "data1 = {'Intrusion':119341, 'Normal':56000}\n",
    "Class_label = list(data1.keys())\n",
    "Number_of_instances = list(data1.values())\n",
    "  \n",
    "fig = plt.figure(figsize = (5, 5))\n",
    " \n",
    "# creating the bar plot\n",
    "plt.bar(Class_label, Number_of_instances, color ='blue',\n",
    "        width = 0.3)\n",
    "plt.ylim(0, 130000)\n",
    "plt.xlabel(\"Class label\")\n",
    "plt.ylabel(\"Number of instances\")\n",
    "plt.title(\"Number of samples in each class before resampling (UNSW NB15 Dataset)\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "d6051b4b",
   "metadata": {},
   "outputs": [],
   "source": [
    "# RESAMPLING\n",
    "\n",
    "from imblearn import under_sampling, over_sampling\n",
    "from imblearn.over_sampling import SMOTE\n",
    "sm = SMOTE()\n",
    "x1, y1 = sm.fit_sample(x1, y1) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "id": "61267d8e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(238682, 42)"
      ]
     },
     "execution_count": 102,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x1.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 259,
   "id": "7c5abea0",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "data2 = {'Intrusion':119341, 'Normal':119341}\n",
    "Class_label = list(data2.keys())\n",
    "Number_of_instances = list(data2.values())\n",
    "  \n",
    "fig = plt.figure(figsize = (5, 5))\n",
    " \n",
    "# creating the bar plot\n",
    "plt.bar(Class_label, Number_of_instances, color ='blue',\n",
    "        width = 0.3)\n",
    "plt.ylim(0, 130000)\n",
    "plt.xlabel(\"Class label\")\n",
    "plt.ylabel(\"Number of instances\")\n",
    "plt.title(\"Number of samples in each class after resampling (UNSW NB15 Dataset)\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "66865894",
   "metadata": {},
   "outputs": [],
   "source": [
    "# NORMALIZATION\n",
    "\n",
    "from sklearn.preprocessing import MinMaxScaler\n",
    "model = MinMaxScaler()\n",
    "model.fit(x1)\n",
    "x1 = model.transform(x1)\n",
    "x2 = model.transform(x2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "1b18843b",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\admin\\anaconda3\\lib\\site-packages\\xgboost\\sklearn.py:1224: UserWarning: The use of label encoder in XGBClassifier is deprecated and will be removed in a future release. To remove this warning, do the following: 1) Pass option use_label_encoder=False when constructing XGBClassifier object; and 2) Encode your labels (y) as integers starting with 0, i.e. 0, 1, 2, ..., [num_class - 1].\n",
      "  warnings.warn(label_encoder_deprecation_msg, UserWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[13:39:26] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.5.1/src/learner.cc:1115: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "XGBClassifier(alpha=0.05, base_score=0.5, booster='gbtree', colsample_bylevel=1,\n",
       "              colsample_bynode=1, colsample_bytree=0.6,\n",
       "              enable_categorical=False, gamma=0.3, gpu_id=-1,\n",
       "              importance_type=None, interaction_constraints='',\n",
       "              learning_rate=0.01, max_delta_step=0, max_depth=4,\n",
       "              min_child_weight=12, missing=nan, monotone_constraints='()',\n",
       "              n_estimators=1484, n_jobs=4, num_parallel_tree=1,\n",
       "              predictor='auto', random_state=27, reg_alpha=0.0500000007,\n",
       "              reg_lambda=1, scale_pos_weight=1, seed=27, subsample=0.6,\n",
       "              tree_method='exact', validate_parameters=1, verbosity=None)"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 1. XGBOOST CLASSIFIER MODEL\n",
    "\n",
    "from xgboost import XGBClassifier\n",
    "import xgboost as xgb\n",
    "params = {\n",
    "            'objective':'binary:logistic',\n",
    "            'max_depth': 4, 'min_child_weight': 12, 'gamma': 0.3, 'subsample': 0.6,\n",
    "            'colsample_bytree': 0.6, 'scale_pos_weight': 1,\n",
    "            'alpha': 0.05,\n",
    "            'learning_rate': 0.01,\n",
    "            'n_estimators':1484,'seed': 27\n",
    "         }\n",
    "xgb_clf = XGBClassifier(**params)\n",
    "xgb_clf.fit(x1, y1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "faadf962",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[1 1 1 ... 0 0 1]\n"
     ]
    }
   ],
   "source": [
    "y_pred_xgb=xgb_clf.predict(x2)\n",
    "print(y_pred_xgb)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "38558965",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ROC AUC: 0.893511\n"
     ]
    }
   ],
   "source": [
    "# ROC AUC\n",
    "from sklearn.metrics import roc_auc_score\n",
    "auc = roc_auc_score(y2, y_pred_xgb)\n",
    "print('ROC AUC: %f' % auc)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "8e3bb34d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from matplotlib import pyplot as plt\n",
    "from sklearn import metrics\n",
    "fpr, tpr, _ = metrics.roc_curve(y2,  y_pred_xgb)\n",
    "#create ROC curve\n",
    "plt.plot(fpr,tpr)\n",
    "plt.ylabel('True Positive Rate')\n",
    "plt.xlabel('False Positive Rate')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "1a6d64ee",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy in XGBoost classifier model =  90.1229169703153\n",
      "Confusion Matrix = [[30238  6762]\n",
      " [ 1370 43962]]\n",
      "Recall = 0.901229169703153\n",
      "Precision = 0.9071211418448522\n",
      "Classification Report =\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.96      0.82      0.88     37000\n",
      "           1       0.87      0.97      0.92     45332\n",
      "\n",
      "    accuracy                           0.90     82332\n",
      "   macro avg       0.91      0.89      0.90     82332\n",
      "weighted avg       0.91      0.90      0.90     82332\n",
      "\n",
      "F1 Score =  0.8984062997805073\n"
     ]
    }
   ],
   "source": [
    "from sklearn import metrics\n",
    "from sklearn.metrics import f1_score\n",
    "\n",
    "print('Accuracy in XGBoost classifier model = ', metrics.accuracy_score(y2, y_pred_xgb)*100)\n",
    "print(\"Confusion Matrix =\", metrics.confusion_matrix(y2, y_pred_xgb, labels=None, \n",
    "                                              sample_weight=None))\n",
    "print(\"Recall =\", metrics.recall_score(y2, y_pred_xgb, labels=None, \n",
    "                                             pos_label=1, average='weighted', \n",
    "                                             sample_weight=None))\n",
    "print(\"Precision =\", metrics.precision_score(y2, y_pred_xgb, labels=None, \n",
    "                                             pos_label=1, average='weighted', \n",
    "                                             sample_weight=None))\n",
    "print(\"Classification Report =\\n\", metrics.classification_report(y2, y_pred_xgb, \n",
    "                                                                 labels=None, \n",
    "                                                                 target_names=None, \n",
    "                                                                 sample_weight=None, \n",
    "                                                                 digits=2, \n",
    "                                                                 output_dict=False))\n",
    "print(\"F1 Score = \",f1_score(y2, y_pred_xgb, average='macro'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "11ced96a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:>"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "cm_xgb = metrics.confusion_matrix(y_pred_xgb, y2, labels=None, \n",
    "                                              sample_weight=None)\n",
    "import seaborn as sns\n",
    "sns.heatmap(cm_xgb, annot=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "56251b63",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:>"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.heatmap(cm_xgb/np.sum(cm_xgb), annot=True, \n",
    "            fmt='.2%', cmap='Blues')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 220,
   "id": "d53a70af",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_16\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " dense_58 (Dense)            (None, 42)                1806      \n",
      "                                                                 \n",
      " dense_59 (Dense)            (None, 84)                3612      \n",
      "                                                                 \n",
      " dropout_21 (Dropout)        (None, 84)                0         \n",
      "                                                                 \n",
      " dense_60 (Dense)            (None, 168)               14280     \n",
      "                                                                 \n",
      " dropout_22 (Dropout)        (None, 168)               0         \n",
      "                                                                 \n",
      " dense_61 (Dense)            (None, 336)               56784     \n",
      "                                                                 \n",
      " dropout_23 (Dropout)        (None, 336)               0         \n",
      "                                                                 \n",
      " dense_62 (Dense)            (None, 672)               226464    \n",
      "                                                                 \n",
      " dense_63 (Dense)            (None, 2)                 1346      \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 304,292\n",
      "Trainable params: 304,292\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "# 2. DNN CLASSIFIER MODEL\n",
    "\n",
    "from tensorflow import keras\n",
    "model_DNN=keras.models.Sequential()\n",
    "model_DNN.add(keras.layers.Dense(units=42, activation = \"relu\",input_shape = x1.shape[1:]))\n",
    "model_DNN.add(keras.layers.Dense(units=84,activation = \"relu\"))\n",
    "model_DNN.add(keras.layers.Dropout(0.25))\n",
    "model_DNN.add(keras.layers.Dense(units=168,activation = \"relu\" ))\n",
    "model_DNN.add(keras.layers.Dropout(0.25))\n",
    "model_DNN.add(keras.layers.Dense(units=336,activation = \"relu\" ))\n",
    "model_DNN.add(keras.layers.Dropout(0.25))\n",
    "model_DNN.add(keras.layers.Dense(units=672,activation = \"relu\" ))\n",
    "model_DNN.add(keras.layers.Dense(units=2, activation = \"sigmoid\"))\n",
    "model_DNN.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 221,
   "id": "0d7eae25",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/50\n",
      "14918/14918 [==============================] - 99s 6ms/step - loss: 0.3200 - accuracy: 0.8697 - val_loss: 0.4966 - val_accuracy: 0.7508\n",
      "Epoch 2/50\n",
      "14918/14918 [==============================] - 202s 14ms/step - loss: 0.3817 - accuracy: 0.8650 - val_loss: 0.5064 - val_accuracy: 0.7662\n",
      "Epoch 3/50\n",
      "14918/14918 [==============================] - 139s 8ms/step - loss: 0.5603 - accuracy: 0.8078 - val_loss: 0.6688 - val_accuracy: 0.7660\n",
      "Epoch 4/50\n",
      "14918/14918 [==============================] - 134s 9ms/step - loss: 0.3492 - accuracy: 0.8681 - val_loss: 0.4878 - val_accuracy: 0.7664\n",
      "Epoch 5/50\n",
      "14918/14918 [==============================] - 115s 8ms/step - loss: 0.5747 - accuracy: 0.8607 - val_loss: 0.4933 - val_accuracy: 0.7664\n",
      "Epoch 6/50\n",
      "14918/14918 [==============================] - 90s 6ms/step - loss: 0.4583 - accuracy: 0.8450 - val_loss: 0.4979 - val_accuracy: 0.7664\n",
      "Epoch 7/50\n",
      "14918/14918 [==============================] - 186s 12ms/step - loss: 0.3853 - accuracy: 0.8729 - val_loss: 0.5023 - val_accuracy: 0.7665\n",
      "Epoch 8/50\n",
      "14918/14918 [==============================] - 118s 8ms/step - loss: 0.3683 - accuracy: 0.8729 - val_loss: 0.4884 - val_accuracy: 0.7669\n",
      "Epoch 9/50\n",
      "14918/14918 [==============================] - 115s 8ms/step - loss: 0.3981 - accuracy: 0.8581 - val_loss: 0.5011 - val_accuracy: 0.7666\n",
      "Epoch 10/50\n",
      "14918/14918 [==============================] - 127s 9ms/step - loss: 0.4028 - accuracy: 0.8738 - val_loss: 0.5113 - val_accuracy: 0.7665\n",
      "Epoch 11/50\n",
      "14918/14918 [==============================] - 124s 8ms/step - loss: 0.3715 - accuracy: 0.8758 - val_loss: 0.5086 - val_accuracy: 0.7665\n",
      "2573/2573 [==============================] - 13s 5ms/step - loss: 0.4966 - accuracy: 0.7508\n",
      "Test loss =  0.49659720063209534\n",
      "Test accuracy =  0.7508016228675842\n"
     ]
    }
   ],
   "source": [
    "opt = keras.optimizers.Adam(learning_rate=0.01)\n",
    "model_DNN.compile(loss=\"sparse_categorical_crossentropy\",optimizer=opt,metrics=\"accuracy\")\n",
    "es=keras.callbacks.EarlyStopping(monitor='loss',patience=10,restore_best_weights=True)\n",
    "cp=keras.callbacks.ModelCheckpoint(\"SaranyaDNNNet.h5\",monitor='val_loss')\n",
    "history = model_DNN.fit(x1,y1,epochs=50,batch_size=16,callbacks=[es,cp],validation_data = [x2, y2])\n",
    "testloss,testaccuracy=model_DNN.evaluate(x2,y2)\n",
    "print(\"Test loss = \",testloss)\n",
    "print(\"Test accuracy = \",testaccuracy)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 222,
   "id": "cf406e54",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[1 1 1 ... 0 0 1]\n",
      "Test Loss = 0.49659720063209534\n",
      "Test Accuracy = 0.7508016228675842\n"
     ]
    }
   ],
   "source": [
    "y_predicted_DNN = model_DNN.predict(x2)\n",
    "y_predicted_class_DNN=np.argmax(y_predicted_DNN,axis=1)\n",
    "print(y_predicted_class_DNN)\n",
    "print(\"Test Loss =\",testloss)\n",
    "print(\"Test Accuracy =\",testaccuracy)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 223,
   "id": "43797400",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ROC AUC: 0.722756\n"
     ]
    }
   ],
   "source": [
    "# ROC AUC\n",
    "from sklearn.metrics import roc_auc_score\n",
    "auc = roc_auc_score(y2, y_predicted_class_DNN)\n",
    "print('ROC AUC: %f' % auc)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 224,
   "id": "34920626",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn import metrics\n",
    "fpr, tpr, _ = metrics.roc_curve(y2,  y_predicted_class_DNN)\n",
    "#create ROC curve\n",
    "plt.plot(fpr,tpr)\n",
    "plt.ylabel('True Positive Rate')\n",
    "plt.xlabel('False Positive Rate')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 225,
   "id": "ad828970",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy with DNN classifier model =  75.08016324150998\n",
      "Confusion Matrix = [[16488     5]\n",
      " [20512 45327]]\n",
      "Recall = 0.7508016324150998\n",
      "Precision = 0.8888570123088142\n",
      "Classification Report =\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.45      1.00      0.62     16493\n",
      "           1       1.00      0.69      0.82     65839\n",
      "\n",
      "    accuracy                           0.75     82332\n",
      "   macro avg       0.72      0.84      0.72     82332\n",
      "weighted avg       0.89      0.75      0.78     82332\n",
      "\n",
      "F1 Score =  0.7159504818613832\n"
     ]
    }
   ],
   "source": [
    "from sklearn import metrics\n",
    "from sklearn.metrics import f1_score\n",
    "print('Accuracy with DNN classifier model = ', metrics.accuracy_score(y_predicted_class_DNN, y2)*100)\n",
    "print(\"Confusion Matrix =\", metrics.confusion_matrix(y_predicted_class_DNN, y2, labels=None, \n",
    "                                              sample_weight=None))\n",
    "print(\"Recall =\", metrics.recall_score(y_predicted_class_DNN, y2, labels=None, \n",
    "                                             pos_label=1, average='weighted', \n",
    "                                             sample_weight=None))\n",
    "print(\"Precision =\", metrics.precision_score(y_predicted_class_DNN, y2, labels=None, \n",
    "                                             pos_label=1, average='weighted', \n",
    "                                             sample_weight=None))\n",
    "print(\"Classification Report =\\n\", metrics.classification_report(y_predicted_class_DNN, y2, \n",
    "                                                                 labels=None, \n",
    "                                                                 target_names=None, \n",
    "                                                                 sample_weight=None, \n",
    "                                                                 digits=2, \n",
    "                                                                 output_dict=False))\n",
    "print(\"F1 Score = \",f1_score(y_predicted_class_DNN, y2, average='macro'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 226,
   "id": "496355f8",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:>"
      ]
     },
     "execution_count": 226,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "cm_DNN = metrics.confusion_matrix(y_predicted_class_DNN, y2, labels=None, \n",
    "                                              sample_weight=None)\n",
    "import seaborn as sns\n",
    "sns.heatmap(cm_DNN, annot=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 227,
   "id": "6bf538d9",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:>"
      ]
     },
     "execution_count": 227,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.heatmap(cm_DNN/np.sum(cm_DNN), annot=True, \n",
    "            fmt='.2%', cmap='Blues')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 208,
   "id": "bd3df090",
   "metadata": {},
   "outputs": [],
   "source": [
    "# 3. CNN CLASSIFIER MODEL\n",
    "\n",
    "x1_CNN = np.reshape(x1, (x1.shape[0],x1.shape[1],1))\n",
    "x2_CNN = np.reshape(x2, (x2.shape[0],x2.shape[1],1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 209,
   "id": "08328a4d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_14\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " conv1d_8 (Conv1D)           (None, 40, 42)            168       \n",
      "                                                                 \n",
      " conv1d_9 (Conv1D)           (None, 38, 84)            10668     \n",
      "                                                                 \n",
      " batch_normalization_4 (Batc  (None, 38, 84)           336       \n",
      " hNormalization)                                                 \n",
      "                                                                 \n",
      " max_pooling1d_4 (MaxPooling  (None, 19, 84)           0         \n",
      " 1D)                                                             \n",
      "                                                                 \n",
      " dropout_19 (Dropout)        (None, 19, 84)            0         \n",
      "                                                                 \n",
      " flatten_4 (Flatten)         (None, 1596)              0         \n",
      "                                                                 \n",
      " dense_56 (Dense)            (None, 168)               268296    \n",
      "                                                                 \n",
      " dropout_20 (Dropout)        (None, 168)               0         \n",
      "                                                                 \n",
      " dense_57 (Dense)            (None, 2)                 338       \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 279,806\n",
      "Trainable params: 279,638\n",
      "Non-trainable params: 168\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "from tensorflow import keras\n",
    "from tensorflow.keras.layers import Input\n",
    "from keras.layers import Convolution1D, Dense, Dropout, MaxPooling1D, LSTM\n",
    "from keras.layers import Dense, Dropout, Activation, Lambda, Flatten\n",
    "convnet=keras.models.Sequential()\n",
    "convnet.add(keras.layers.Convolution1D(42,3,activation=\"relu\",input_shape=(42,1)))\n",
    "convnet.add(keras.layers.Convolution1D(84,3,activation=\"relu\"))\n",
    "convnet.add(keras.layers.BatchNormalization())\n",
    "convnet.add(keras.layers.MaxPooling1D(2,2))\n",
    "convnet.add(keras.layers.Dropout(0.25))\n",
    "convnet.add(keras.layers.Flatten())\n",
    "convnet.add(keras.layers.Dense(168,activation=\"relu\"))\n",
    "convnet.add(keras.layers.Dropout(0.25))\n",
    "convnet.add(keras.layers.Dense(2,activation=\"softmax\"))\n",
    "convnet.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 210,
   "id": "cc0367a5",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/50\n",
      "14918/14918 [==============================] - 194s 12ms/step - loss: 0.1959 - accuracy: 0.8984 - val_loss: 0.3032 - val_accuracy: 0.7868\n",
      "Epoch 2/50\n",
      "14918/14918 [==============================] - 166s 11ms/step - loss: 0.2076 - accuracy: 0.8942 - val_loss: 0.3294 - val_accuracy: 0.7870\n",
      "Epoch 3/50\n",
      "14918/14918 [==============================] - 172s 12ms/step - loss: 0.2197 - accuracy: 0.8925 - val_loss: 0.3516 - val_accuracy: 0.7869\n",
      "Epoch 4/50\n",
      "14918/14918 [==============================] - 167s 11ms/step - loss: 0.2218 - accuracy: 0.8949 - val_loss: 0.3518 - val_accuracy: 0.7871\n",
      "Epoch 5/50\n",
      "14918/14918 [==============================] - 168s 11ms/step - loss: 0.2342 - accuracy: 0.8947 - val_loss: 0.3294 - val_accuracy: 0.7871\n",
      "Epoch 6/50\n",
      "14918/14918 [==============================] - 164s 11ms/step - loss: 0.2230 - accuracy: 0.8949 - val_loss: 0.3655 - val_accuracy: 0.7869\n",
      "Epoch 7/50\n",
      "14918/14918 [==============================] - 162s 11ms/step - loss: 0.2299 - accuracy: 0.8944 - val_loss: 0.3454 - val_accuracy: 0.7870\n",
      "Epoch 8/50\n",
      "14918/14918 [==============================] - 147s 10ms/step - loss: 0.2208 - accuracy: 0.8952 - val_loss: 0.3742 - val_accuracy: 0.7865\n",
      "Epoch 9/50\n",
      "14918/14918 [==============================] - 148s 10ms/step - loss: 0.2239 - accuracy: 0.8951 - val_loss: 0.3485 - val_accuracy: 0.7869\n",
      "Epoch 10/50\n",
      "14918/14918 [==============================] - 156s 10ms/step - loss: 0.2229 - accuracy: 0.8950 - val_loss: 0.3590 - val_accuracy: 0.7871\n",
      "Epoch 11/50\n",
      "14918/14918 [==============================] - 166s 11ms/step - loss: 0.2292 - accuracy: 0.8950 - val_loss: 0.3516 - val_accuracy: 0.7868\n",
      "2573/2573 [==============================] - 10s 4ms/step - loss: 0.3032 - accuracy: 0.7868\n",
      "Test Loss = 0.3032172620296478\n",
      "Test Accuracy = 0.7867657542228699\n"
     ]
    }
   ],
   "source": [
    "opt = keras.optimizers.Adam(learning_rate=0.01)\n",
    "convnet.compile(loss=\"sparse_categorical_crossentropy\",optimizer=opt,metrics=\"accuracy\")\n",
    "es=keras.callbacks.EarlyStopping(monitor='loss',patience=10,restore_best_weights=True)\n",
    "cp=keras.callbacks.ModelCheckpoint(\"SaranyaConvNet.h5\",monitor='val_loss')\n",
    "history = convnet.fit(x1_CNN,y1,epochs=50,batch_size=16,callbacks=[es,cp],validation_data = [x2_CNN, y2])\n",
    "testloss,testaccuracy=convnet.evaluate(x2_CNN,y2)\n",
    "print(\"Test Loss =\",testloss)\n",
    "print(\"Test Accuracy =\",testaccuracy)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 211,
   "id": "59edbb3a",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[1 1 1 ... 0 0 1]\n"
     ]
    }
   ],
   "source": [
    "y_predicted_CNN = convnet.predict(x2_CNN)\n",
    "y_predicted_class_CNN=np.argmax(y_predicted_CNN,axis=1)\n",
    "print(y_predicted_class_CNN)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 212,
   "id": "ecbed5d4",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ROC AUC: 0.762829\n"
     ]
    }
   ],
   "source": [
    "# ROC AUC\n",
    "from sklearn.metrics import roc_auc_score\n",
    "auc = roc_auc_score(y2, y_predicted_class_CNN)\n",
    "print('ROC AUC: %f' % auc)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 213,
   "id": "0fe4f03b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fpr, tpr, _ = metrics.roc_curve(y2,  y_predicted_class_CNN)\n",
    "#create ROC curve\n",
    "plt.plot(fpr,tpr)\n",
    "plt.ylabel('True Positive Rate')\n",
    "plt.xlabel('False Positive Rate')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 214,
   "id": "a7c37ebb",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy with CNN classifier model =  78.67657775834427\n",
      "Confusion Matrix = [[19473    29]\n",
      " [17527 45303]]\n",
      "Recall = 0.7867657775834427\n",
      "Precision = 0.8873057376127149\n",
      "Classification Report =\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.53      1.00      0.69     19502\n",
      "           1       1.00      0.72      0.84     62830\n",
      "\n",
      "    accuracy                           0.79     82332\n",
      "   macro avg       0.76      0.86      0.76     82332\n",
      "weighted avg       0.89      0.79      0.80     82332\n",
      "\n",
      "F1 Score =  0.7634866238039846\n"
     ]
    }
   ],
   "source": [
    "from sklearn import metrics\n",
    "from sklearn.metrics import f1_score\n",
    "print('Accuracy with CNN classifier model = ', metrics.accuracy_score(y_predicted_class_CNN, y2)*100)\n",
    "print(\"Confusion Matrix =\", metrics.confusion_matrix(y_predicted_class_CNN, y2, labels=None, \n",
    "                                              sample_weight=None))\n",
    "print(\"Recall =\", metrics.recall_score(y_predicted_class_CNN, y2, labels=None, \n",
    "                                             pos_label=1, average='weighted', \n",
    "                                             sample_weight=None))\n",
    "print(\"Precision =\", metrics.precision_score(y_predicted_class_CNN, y2, labels=None, \n",
    "                                             pos_label=1, average='weighted', \n",
    "                                             sample_weight=None))\n",
    "print(\"Classification Report =\\n\", metrics.classification_report(y_predicted_class_CNN, y2, \n",
    "                                                                 labels=None, \n",
    "                                                                 target_names=None, \n",
    "                                                                 sample_weight=None, \n",
    "                                                                 digits=2, \n",
    "                                                                 output_dict=False))\n",
    "print(\"F1 Score = \",f1_score(y_predicted_class_CNN, y2, average='macro'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 215,
   "id": "9e04ed77",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:>"
      ]
     },
     "execution_count": 215,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "cm_CNN = metrics.confusion_matrix(y_predicted_class_CNN, y2, labels=None, \n",
    "                                              sample_weight=None)\n",
    "import seaborn as sns\n",
    "sns.heatmap(cm_CNN, annot=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 216,
   "id": "e26bdfdf",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:>"
      ]
     },
     "execution_count": 216,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAWIAAAD7CAYAAABQQp5FAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuNCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8QVMy6AAAACXBIWXMAAAsTAAALEwEAmpwYAAAYOUlEQVR4nO3deXhU1f3H8ffMJCFhE5fIEmSR5biyi5TFBbFYpFYElFL3JUXtz9Jat4IKCm4IjQpigQJFiWKhaEHUgol1AUUUsG6nAiqyiGwBsm/z+yMhHWIkM9Lk5F4/L5/7PJy595577vPoxy9nzr0TCIfDiIiIO0HXAxAR+bFTEIuIOKYgFhFxTEEsIuKYglhExDEFsYiIY3G1ebFHMjdqrZwc4pZ+7VwPQeqoxDgCR3J+UtffRJ03eWunHtG1jlStBrGISK0JeOcv/ApiEfGngNMiNyYKYhHxJ1XEIiKOqSIWEXEsGHI9gqgpiEXEnzQ1ISLimKYmREQcU0UsIuKYKmIREcdUEYuIOKZVEyIijqkiFhFxLKg5YhERt1QRi4g4plUTIiKO1dCXdcaYkcBYIB5Is9ZOq7T/XuBaYG/5RzMrH1OZglhE/KkGpiaMMSnARKA7UACsNMZkWms/iTisBzDCWrsq2n69M4kiIhKLQCD6LXoDgAxr7R5rbQ6wEBhW6ZgewB+NMR8aY6YaYxKr61QVsYj4UwwVsTGmCdCkil1Z1tqsiHYLYHtEezvQM6KfhsBa4DZgAzAXuBsYc7jrK4hFxJ9iq3RHA/dW8fl4YFxEOwhE/hZeACg92LDWZgODDraNMZOB2SiIReRHKbY54jTKqtfKsiq1twD9ItrNgG0HG8aYVsAAa+3sg6MAiqq7uIJYRPwphlUT5dMPWVEcugIYZ4xJBnKAoUBqxP484BFjTCbwJXAzsLjaoUY9UhERLwkEo9+iZK3dStk0QyawDki31q42xiwzxvSw1u4Efg0sASxlFfHk6vpVRSwi/lRDD3RYa9OB9EqfDYr48yJgUSx9KohFxJ/0iLOIiGN6xFlExDFVxCIibgWCCmIREacCmpoQEXHMOzmsIBYRf1JFLCLimIJYRMSxoL6sExFxzDsFsYJYRPxJUxMiIo4piEVEHFMQi4g4piAWEXEsEFQQi4g4pYpYRMQxBbGIiGveyWEFsYj4kypiERHHFMQiIo7pXRMiIq55pyBWEIuIP2lqQkTEMQWxiIhjCmKfKy0p5o15fyJ797eUFBXRZdAIGie34O35jxMOhzmm5Yn8ZMQogsHQIeetf2UBX61/l9KSYk4++0JMn4Hs2ryB5dPG0/j4FgCcfPYgTuxxNm/Nf4I9WzZx8tmD6dDrPArzclj57JOcc+1tLm5ZYlRaWsrE+8fxH2tJSEjg3vETaNW6dcX+1zMzmDF9GqG4OC4eMpShwy+t2Ld7925+eekl/HnmbNqe2I6333yDaVMfp3nzFkyakkYwGOSBCfdx1TXXkpLS0sXteYIecfa5De9mkNigMedccxv52ft5YeL/cWyrdnS/+CqadzidN+ZOYfP6d2nTtXfFOdvth+zY+Ck/v+1RigsL+PfyRQDs3ryB0wYM4fTzL6k4Nj97P3n79/Lz2yaz7E930aHXeax/5Xk6DRxe6/cqP0zGaysoLCjk6fQFfLh+HZMnPcRjU6cDUFRUxKMPP0j6goUkJSVx1eW/5OxzzuW45GSKioq4f/w91KuXWNHXgufSeWrmbKZPexxrPyMuFKJhw4YK4Wr4qiI2xpwEDANaAqXANuAVa+2aGh5bndW2Wz/adOtb0Q6GQpz36zEEgyFKiovI3b+XpMZNDjlnyyfvc0xKG1Y8NYGi/FzOGHotALs2b2Dfji18tX4VRx2fwpmXphKKT6C0pISS4kJC8Qkc2PUNxQX5HJPSphbvUo7E2g/ep3fffgB06tyFjz/+qGLfF5s2ckKrVjQ+6igAunbrzgcfrOGnA3/GlEcfZvilI5g9a0bF8fXrNyAvL4+8vDySkpL485NT+ePd42r1frzIS0F82IV2xpibgOfKm+8BH5T/eaYx5taaHFhdFp+YREJifQrzc8mY8QDdL7qCYDDEgd07+Pv4GynI3sdRTQ+tVvKz97Pzq8/pn3oXvUf+hn/NnkQ4HCa5TUd6XnIdg/8wiUbHNWPt0nTi6yXSutOZZM56hK6DR7J22bOc2v8XrFrwFO88P4OignxHdy7RysnJplGjhhXtUDBEcXExANnZ2TRs1KhiX/0GDcg+kM2Li//O0UcfQ5/yAD8oddRNPPzABFJSWvL15s107tKNl5ct5f7x97B+3drauSEPCgQCUW+uVbfi+bdAb2vtBGvtX6y1s6y1E4A+wA01P7y6K3vPTl6eciftz+xPu57nAtDo2KYMv38WJ501iHcXzjzk+MSGjWl5SndCcfE0adaSUFwC+Qf20bpLb45r3QGA1l16s/vrjQCcdNYgzr/pHgiHaXxcc7Z9to5m7U+jabtT2Lj69Vq9V4ldgwYNycnJqWiXhkuJiyv7C2jDhg3JjdiXm5NDo0aNeGHxIt5ZtZLrrr4C+9mnjLnrDnbt3MmJ7dox5bEnuPb6VBYvWsigCwez8u23uGvMPcx46slavzfPCMSwOVZdEBcD8VV8ngQU/e+H4w15+/fyyuNjOeOSa+nY56cALH9yPPt2bAXKKubK/5dt2u4Utny8hnA4TE7WbooK86nXsBGvPj6WnV9YALbZdRzXqv0h5320YjGnDbiY4sICAsEgBKC4IK8W7lKORNeu3XjrjTcA+HD9Ojp06Fixr+2J7dj81Vfsy8qiqLCQ999fQ6cuXZkzbz6z//oMf5n7NOakk5n44MMcl5xccd6ivy3goouHABAuLSUQCJCXp38Xvo+XKuLq5ognAmuNMa8B24Ew0ALoD4yp4bHVWeteXkBhbjZrX3qWtS89C0CPX1zJG3+dQigunlBCPfpd8VsA/jXnUbr/4kpadTqTbz7/iH88NJpwOEzvETcRDIboPfI3rHpuOsFQHElHHU3fX91ScZ2N7/2LVp3OJC4hkbbd+5Ex80ECwSDnXneHk/uW6PUfcD6rVr3Nlb8aQTgc5r4JD7Bs6RJyc3MZdull3Hr7ndyYeh2l4TAXDxlK06ZND9tfdnY27723mkmT0wA49rhkrrr8l1w2YmQt3I03BWto1YQxZiQwlrIiNc1aO+17jrsQmGqtbVtdn4FwOFzdRVsAAygL4CCwBVhhrd0W2/DhkcyNh7+Y/Ojc0q+d6yFIHZUYd2STBh1ueyXqvPl80gVRXcsYkwK8BXQHCoCVwC+ttZ9UOq4p8DqQZK1tU12/1a6aKA/cedEMUkSkrohlxsEY0wRoUsWuLGttVkR7AJBhrd1Tft5CylaV3VfpvFnAeOChaK7vndcTiYjEIMY54tHAF1Vsoyt124KyadqDtlO2tLeCMeYWylaYvRPtWPVAh4j4UiwVcRjSgLlV7Mqq1A6WHf7fy1D2fAUAxpjTgKHAeVQK6MNREIuIL8XyZd0nZdMPWVEcugWIXOjdjLKH3A4aDjQH1gAJQAtjzJvW2kMXh1eiIBYRX6qhVRMrgHHGmGQgh7LqN/XgTmvtvcC9AMaYNsDr1YUwaI5YRHwqEIh+i5a1ditlS3czgXVAurV2tTFmmTGmxw8dqypiEfGlmnpQw1qbDqRX+mxQFcd9CbSJpk8FsYj4Ul14Yi5aCmIR8SUP5bCCWET8qaYeca4JCmIR8SVNTYiIOOahHFYQi4g/qSIWEXHMQzmsIBYRf1JFLCLimFZNiIg45qGCWEEsIv6kqQkREcc8lMMKYhHxJ1XEIiKOKYhFRBzTqgkREcc8VBAriEXEnzQ1ISLimIdyWEEsIv4U9FASK4hFxJf0ZZ2IiGMeymEFsYj4k76sExFxzEM5rCAWEX8K4J0kVhCLiC9pjlhExDGtmhARcUzriEVEHPNQDiuIRcSftHxNRMQxD+WwglhE/ClUQ0lsjBkJjAXigTRr7bRK+4cA44EQ8B6Qaq0tPFyfwRoZqYiIY4FAIOotWsaYFGAi0BfoAqQaY06J2N8AmAqcb609FUgErq6uXwWxiPhSMBD9FoMBQIa1do+1NgdYCAw7uLP8szbW2h3GmPrA8cDe6jrV1ISI+FKMlW4ToEkVu7KstVkR7RbA9oj2dqBn5AnW2iJjzM+AZ4CtwD+ru74qYhHxpUAg+g0YDXxRxTa6UrdBIBx5GaC08rWttS9ba48FlgLTqxurKmIR8aUYl6+lAXOr+DyrUnsL0C+i3QzYdrBhjDkG6GGtPVgFzwcWVHdxBbGI+FIohsnf8umHrCgOXQGMM8YkAznAUCA1Yn8AeMYY08NauxkYDrxVXaeamhARXwrEsEXLWrsVGANkAuuAdGvtamPMsvLw3U1ZMC81xqwHDHBHdf2qIhYRX6qpd01Ya9OB9EqfDYr48wvAC7H0qSAWEV/Sk3UiIo7pXRMiIo55KIcVxCLiT7GsmnCtVoP46PrKfTnU0Wf8xvUQpI7KWzv1iM7X1ISIiGNeWpurIBYRX1JFLCLimIemiBXEIuJP+rJORMQxD+WwglhE/MlDU8QKYhHxp5p610RNUBCLiC9p+ZqIiGMeKogVxCLiT1o1ISLimIdyWEEsIv6kL+tERBzzUA4riEXEnzQ1ISLiWCCmnwV1S0EsIr4U56GFxApiEfElvQZTRMQxzRGLiDjmoYJYQSwi/qR1xCIijoX0ZZ2IiFtBLV8TEXHLQzMTCmIR8SetmhARcUxf1omIOOahHFYQi4g/1dSL4Y0xI4GxQDyQZq2dVmn/L4DxQAD4ArjGWrv3cH16aIGHiEj0gjFs0TLGpAATgb5AFyDVGHNKxP7GwHTgQmttZ+BDYFx1/aoiFhFfiuVdE8aYJkCTKnZlWWuzItoDgAxr7Z7y8xYCw4D7yvfHAzdba7eWtz8EflXd9VURi4gvBWLYgNGUTSNU3kZX6rYFsD2ivR1oebBhrd1trV0MYIxJAu4EXqhurKqIRcSXYlw1kQbMreLzrMrdAuGIdgAorXySMeYoYDGw3lr71+ouriAWEV+KJYbLpx+yojh0C9Avot0M2BZ5gDGmOfAqkAH8LprrK4hFxJeCNbNqYgUwzhiTDOQAQ4HUgzuNMSFgCfC8tXZCtJ0qiEXEl2riCzBr7VZjzBggE0gAZllrVxtjlgH3ACcA3YA4Y8yw8tPWWGuvP1y/CmIR8aWa+oUOa206kF7ps0Hlf1zDD/h/gIJYRHzJQw/WKYhFxJ/0m3UiIo6FFMQiIm55J4YVxCLiUx4qiBXEIuJP+qkkERHHVBGLiDgWUEUsIuKWVk2IiDjmoRxWEIuIPymIRUQc0xyxiIhjNfTboTVCQSwivhTjL3Q4pSAWEV/S1ITPlRQX8+qsyezbtYOS4iJ6XTSS9t1+AkDm/Okc3fwEuvQfHNU5O778nOVzHycUF8/xrdrR//IbCQSD/HNOGjs3b6LLeT/n1L7nU5Cbw4p5T3DhqDtd3LJEadWzd7A/Ox+AL7fuZvpzr7PosVFs2LwTgJl/e5OF//yg4vhAIMBjf7yMTh1TKCgs5sb75rPp61106pjClDuGU1IapqCwmOvvnse3ew7wxJgRdOqYwp//9ibpS1fTuGEiaXdeyrVj5zm537pMUxM+98nK10hs2JhBo+4g78B+5t19Iy3an8yyGY+w95utnNH8hKjOad/tJ/xzdhr9r7iJlA6n8tbCOXy6KoO2nXqSu28vI+9OY8FDt3Nq3/N5d8lz9Bx8mYO7lWjVSyj7z2ngDY9VfHb1kJ/w+DMZPPZ0RpXnXHRuJxIT4jjnqsn0PL0ND/3+Ei793QwevX0Yv3/4b3z4n61cN7QPt15zPg/PepWmxzbinKun8MqMW0hfuprbrv0pj85ZXiv35zWqiH3O9DwLc8Z/fz8wGApRWJBHnyFXsGn9e1GfA3Bg7y5SOpwKQIsOp7Lxg1W079GHkpISiosKiYtPIGvndooK8klu2bYG70qOVKeOKdRPTGDJkzcTFwpy79QldD25FR1bH8/gczqxYfO33DZpEdm5BRXn9O7ajuUrPwVg9b+/pPsprQC48s45fLNrPwBxoRD5BUXkFxQRFxcisV4c+QVFtG5xLA2S6vHJxu3fHYz4Z/maMabV4fZbazf/b4fjDQmJSQAU5uXyj6n303fo1TRJbk6T5ObfG8RVnQPQJLk5X3/2ISec1ImNa9+hqCCfhHpJtO/ai6VPPkjviy/nnRfSOfOiEbz29DSCwSB9hl1NQr2kWrlXiV5ufhFp815jzuKVtG91PC9OvZFH5yxn7uKVrP30a26/biBjfj2Iu/60uOKcRg0S2ZedV9EuKSklFApWhHCvzm0ZddlZnH99Grn5hbz0r3/z1wev4YEZL3PXDRfwyOxXmXz7MEpKShk3bSm5+YW1ft91lYdyuNqK+CWgA2U/F135vsLAiTUxKC/Yv/tbXnxsPF3O+zkn9+7/g8+54IZbyXhmOqtfep5mbTtSGB8PQOf+g+ncfzBbP/+YJk2bs/njtbQ0pwPw2cpMOp076HuvI258/tW3bPy6bC54w+Zv2bMvh+UrP2HLjiwA/pG5nim3Dz/knAM5+TSqX6+iHQwGKCkpBWDYT7tx+3UDGXLLdHbtzQbgL4ve5i+L3qZX57Zs2rKLc3sa3vpgAwCX/awHcxavrOnb9AwvPeJc3Y/c9QEscIW1tm2l7Ucbwjn79rLwkbs467LrOf3sC47onE3r3uWC629l6K0TyM8+QOvTuh9y3ppXFtF94CUUFRYQDAYJBAIUFuRV7l7qgKsu7sVDvx8CQPPko2jUIJEFU1LpcWprAM7taVj76aF/iVy1bhMD+5ZNTfU8vQ0fbdgGwIhBZzDqsrMYeMNjfLl193eudcvl/Xlifgb1ExMoKSklHIaGEYEulJWO0W6OHbYittbuN8bcAFwPvF07Q6r73l3yLPm52ax6cT6rXpwPwNA/TCQ+4bv/ISz78yP0HXo1a15ZWOU5RzdLYdHkMcQnJHLCyZ05sXPPinM/eyeTdl16EV8vEdPzLJZMm0ggEGTwTX+snRuVmMxdvIqZ913Ba7N/RzgcZtT4+eQXFPGnOy+lsKiEHbv3c/P9zwIw6/4rGD9tKS9mrKd/r5PInPt7AoEAqfc+QzAYYPLtw/j6m708N/kGAN58/3MmPLUMgOEDu/PSGx+Rl1/E35ev5emHr6G0NMyVd85xdu91kZe+rAuEw+Fau9jMd7+qvYuJJ9wyapLrIUgdlbd26hEl6epN+6LOm54nHuU0tbVqQkR8yTv1sIJYRPzKQ0msIBYRX9K7JkREHPNODCuIRcSvPJTECmIR8SUvLV9TEIuIL3loilhBLCL+VFNBbIwZCYwF4oE0a+207zluHpBhrZ1bXZ/VPeIsIuJJgRj+iZYxJgWYCPQFugCpxphTKh3TwhizBBgWbb8KYhHxpUAg+i0GAyircvdYa3OAhXw3cH8FvAg8H22nmpoQEV+KJV+NMU2AJlXsyrLWZkW0WwCRL4DeDvSMPMFaO6m8z77RXl8VsYj4U2xvXxsNfFHFNrpSr0HKXgEceZXSIx2qKmIR8aUYl6+lAXOr+DyrUnsL0C+i3Yyy97UfEQWxiPhSLD8eWj79kBXFoSuAccaYZCAHGAqkxj66Q2lqQkT8qQZeDG+t3QqMATKBdUC6tXa1MWaZMabHDx2qKmIR8aWaerLOWpsOpFf67Du/XWatvTraPhXEIuJLerJORMQxD+WwglhEfMpDSawgFhFf0ovhRUQc804MK4hFxK88lMQKYhHxJb0YXkTEMQ9NESuIRcSfFMQiIo5pakJExDFVxCIijnkohxXEIuJPqohFRJzzThIriEXEl2J5MbxrCmIR8SVNTYiIOKblayIirnknhxXEIuJPHsphBbGI+JPmiEVEHAt4KIkVxCLiS96JYQWxiPiUhwpiBbGI+JOWr4mIOKaKWETEMQWxiIhjmpoQEXFMFbGIiGMeymEFsYj4lIeSWEEsIr6kOWIREcf0YngREdcUxCIibnlpaiIQDoddj0FE5Ect6HoAIiI/dgpiERHHFMQiIo4piEVEHFMQi4g4piAWEXFMQSwi4piCWETEMQWxiIhjesS5lhljRgJjgXggzVo7zfGQpI4wxjQGVgKDrbVfOh6O1CJVxLXIGJMCTAT6Al2AVGPMKU4HJXWCMeZM4C2go+uxSO1TENeuAUCGtXaPtTYHWAgMczwmqRtuAG4GtrkeiNQ+TU3UrhbA9oj2dqCno7FIHWKtvR7AGON6KOKAKuLaFQQiX3cXAEodjUVE6ggFce3aAjSPaDdDfxUV+dHT1ETtWgGMM8YkAznAUCDV7ZBExDVVxLXIWrsVGANkAuuAdGvtaqeDEhHn9AsdIiKOqSIWEXFMQSwi4piCWETEMQWxiIhjCmIREccUxCIijimIRUQcUxCLiDj2/4rLkbsG3MXiAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.heatmap(cm_CNN/np.sum(cm_CNN), annot=True, \n",
    "            fmt='.2%', cmap='Blues')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "5c074ce2",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "238682\n",
      "82332\n"
     ]
    }
   ],
   "source": [
    "# 4. TCN CLASSIFIER MODEL\n",
    "\n",
    "print(x1.shape[0])\n",
    "print(x2.shape[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "c2bb9864",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(238682, 42) (238682,) (82332, 42) (82332,)\n",
      "(238682, 42, 1) (238682,) (82332, 42, 1) (82332,)\n"
     ]
    }
   ],
   "source": [
    "print(x1.shape, y1.shape, x2.shape, y2.shape)\n",
    "x1_TCN = np.reshape(x1, (x1.shape[0],x1.shape[1],1))\n",
    "x2_TCN = np.reshape(x2, (x2.shape[0],x2.shape[1],1))\n",
    "print(x1_TCN.shape, y1.shape, x2_TCN.shape, y2.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "7c221b21",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(238682, 42, 1) (238682, 2) (82332, 42, 1) (82332, 2)\n"
     ]
    }
   ],
   "source": [
    "from tensorflow.keras.utils import to_categorical\n",
    "num_classes = 2\n",
    "y1_TCN = to_categorical(y1, num_classes)\n",
    "y2_TCN = to_categorical(y2, num_classes)\n",
    "print(x1_TCN.shape, y1_TCN.shape, x2_TCN.shape, y2_TCN.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "523678f3",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(238682, 42, 1) (238682, 2, 1) (82332, 42, 1) (82332, 2, 1)\n"
     ]
    }
   ],
   "source": [
    "y1_TCN = np.expand_dims(y1_TCN, axis=2)\n",
    "y2_TCN = np.expand_dims(y2_TCN, axis=2)\n",
    "print(x1_TCN.shape, y1_TCN.shape, x2_TCN.shape, y2_TCN.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "c7f85482",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Requirement already satisfied: keras-tcn in c:\\users\\admin\\anaconda3\\lib\\site-packages (3.4.2)\n",
      "Requirement already satisfied: tensorflow-addons in c:\\users\\admin\\anaconda3\\lib\\site-packages (from keras-tcn) (0.16.1)\n",
      "Requirement already satisfied: tensorflow in c:\\users\\admin\\anaconda3\\lib\\site-packages (from keras-tcn) (2.7.0)\n",
      "Requirement already satisfied: numpy in c:\\users\\admin\\anaconda3\\lib\\site-packages (from keras-tcn) (1.20.1)\n",
      "Requirement already satisfied: absl-py>=0.4.0 in c:\\users\\admin\\anaconda3\\lib\\site-packages (from tensorflow->keras-tcn) (1.0.0)\n",
      "Requirement already satisfied: termcolor>=1.1.0 in c:\\users\\admin\\anaconda3\\lib\\site-packages (from tensorflow->keras-tcn) (1.1.0)\n",
      "Requirement already satisfied: typing-extensions>=3.6.6 in c:\\users\\admin\\anaconda3\\lib\\site-packages (from tensorflow->keras-tcn) (3.7.4.3)\n",
      "Requirement already satisfied: tensorflow-estimator<2.8,~=2.7.0rc0 in c:\\users\\admin\\anaconda3\\lib\\site-packages (from tensorflow->keras-tcn) (2.7.0)\n",
      "Requirement already satisfied: tensorflow-io-gcs-filesystem>=0.21.0 in c:\\users\\admin\\anaconda3\\lib\\site-packages (from tensorflow->keras-tcn) (0.23.1)\n",
      "Requirement already satisfied: wrapt>=1.11.0 in c:\\users\\admin\\anaconda3\\lib\\site-packages (from tensorflow->keras-tcn) (1.12.1)\n",
      "Requirement already satisfied: keras-preprocessing>=1.1.1 in c:\\users\\admin\\anaconda3\\lib\\site-packages (from tensorflow->keras-tcn) (1.1.2)\n",
      "Requirement already satisfied: protobuf>=3.9.2 in c:\\users\\admin\\anaconda3\\lib\\site-packages (from tensorflow->keras-tcn) (3.19.1)\n",
      "Requirement already satisfied: astunparse>=1.6.0 in c:\\users\\admin\\anaconda3\\lib\\site-packages (from tensorflow->keras-tcn) (1.6.3)\n",
      "Requirement already satisfied: h5py>=2.9.0 in c:\\users\\admin\\anaconda3\\lib\\site-packages (from tensorflow->keras-tcn) (2.10.0)\n",
      "Requirement already satisfied: wheel<1.0,>=0.32.0 in c:\\users\\admin\\anaconda3\\lib\\site-packages (from tensorflow->keras-tcn) (0.36.2)\n",
      "Requirement already satisfied: google-pasta>=0.1.1 in c:\\users\\admin\\anaconda3\\lib\\site-packages (from tensorflow->keras-tcn) (0.2.0)\n",
      "Requirement already satisfied: gast<0.5.0,>=0.2.1 in c:\\users\\admin\\anaconda3\\lib\\site-packages (from tensorflow->keras-tcn) (0.4.0)\n",
      "Requirement already satisfied: opt-einsum>=2.3.2 in c:\\users\\admin\\anaconda3\\lib\\site-packages (from tensorflow->keras-tcn) (3.3.0)\n",
      "Requirement already satisfied: six>=1.12.0 in c:\\users\\admin\\anaconda3\\lib\\site-packages (from tensorflow->keras-tcn) (1.15.0)\n",
      "Requirement already satisfied: grpcio<2.0,>=1.24.3 in c:\\users\\admin\\anaconda3\\lib\\site-packages (from tensorflow->keras-tcn) (1.43.0)\n",
      "Requirement already satisfied: flatbuffers<3.0,>=1.12 in c:\\users\\admin\\anaconda3\\lib\\site-packages (from tensorflow->keras-tcn) (2.0)\n",
      "Requirement already satisfied: libclang>=9.0.1 in c:\\users\\admin\\anaconda3\\lib\\site-packages (from tensorflow->keras-tcn) (12.0.0)\n",
      "Requirement already satisfied: tensorboard~=2.6 in c:\\users\\admin\\anaconda3\\lib\\site-packages (from tensorflow->keras-tcn) (2.7.0)\n",
      "Requirement already satisfied: keras<2.8,>=2.7.0rc0 in c:\\users\\admin\\anaconda3\\lib\\site-packages (from tensorflow->keras-tcn) (2.7.0)\n",
      "Requirement already satisfied: google-auth-oauthlib<0.5,>=0.4.1 in c:\\users\\admin\\anaconda3\\lib\\site-packages (from tensorboard~=2.6->tensorflow->keras-tcn) (0.4.6)\n",
      "Requirement already satisfied: google-auth<3,>=1.6.3 in c:\\users\\admin\\anaconda3\\lib\\site-packages (from tensorboard~=2.6->tensorflow->keras-tcn) (2.3.3)\n",
      "Requirement already satisfied: tensorboard-data-server<0.7.0,>=0.6.0 in c:\\users\\admin\\anaconda3\\lib\\site-packages (from tensorboard~=2.6->tensorflow->keras-tcn) (0.6.1)\n",
      "Requirement already satisfied: tensorboard-plugin-wit>=1.6.0 in c:\\users\\admin\\anaconda3\\lib\\site-packages (from tensorboard~=2.6->tensorflow->keras-tcn) (1.8.0)\n",
      "Requirement already satisfied: werkzeug>=0.11.15 in c:\\users\\admin\\anaconda3\\lib\\site-packages (from tensorboard~=2.6->tensorflow->keras-tcn) (1.0.1)\n",
      "Requirement already satisfied: setuptools>=41.0.0 in c:\\users\\admin\\anaconda3\\lib\\site-packages (from tensorboard~=2.6->tensorflow->keras-tcn) (52.0.0.post20210125)\n",
      "Requirement already satisfied: requests<3,>=2.21.0 in c:\\users\\admin\\anaconda3\\lib\\site-packages (from tensorboard~=2.6->tensorflow->keras-tcn) (2.25.1)\n",
      "Requirement already satisfied: markdown>=2.6.8 in c:\\users\\admin\\anaconda3\\lib\\site-packages (from tensorboard~=2.6->tensorflow->keras-tcn) (3.3.6)\n",
      "Requirement already satisfied: pyasn1-modules>=0.2.1 in c:\\users\\admin\\anaconda3\\lib\\site-packages (from google-auth<3,>=1.6.3->tensorboard~=2.6->tensorflow->keras-tcn) (0.2.8)\n",
      "Requirement already satisfied: cachetools<5.0,>=2.0.0 in c:\\users\\admin\\anaconda3\\lib\\site-packages (from google-auth<3,>=1.6.3->tensorboard~=2.6->tensorflow->keras-tcn) (4.2.4)\n",
      "Requirement already satisfied: rsa<5,>=3.1.4 in c:\\users\\admin\\anaconda3\\lib\\site-packages (from google-auth<3,>=1.6.3->tensorboard~=2.6->tensorflow->keras-tcn) (4.8)\n",
      "Requirement already satisfied: requests-oauthlib>=0.7.0 in c:\\users\\admin\\anaconda3\\lib\\site-packages (from google-auth-oauthlib<0.5,>=0.4.1->tensorboard~=2.6->tensorflow->keras-tcn) (1.3.0)\n",
      "Requirement already satisfied: importlib-metadata>=4.4 in c:\\users\\admin\\anaconda3\\lib\\site-packages (from markdown>=2.6.8->tensorboard~=2.6->tensorflow->keras-tcn) (4.10.0)\n",
      "Requirement already satisfied: zipp>=0.5 in c:\\users\\admin\\anaconda3\\lib\\site-packages (from importlib-metadata>=4.4->markdown>=2.6.8->tensorboard~=2.6->tensorflow->keras-tcn) (3.4.1)\n",
      "Requirement already satisfied: pyasn1<0.5.0,>=0.4.6 in c:\\users\\admin\\anaconda3\\lib\\site-packages (from pyasn1-modules>=0.2.1->google-auth<3,>=1.6.3->tensorboard~=2.6->tensorflow->keras-tcn) (0.4.8)\n",
      "Requirement already satisfied: certifi>=2017.4.17 in c:\\users\\admin\\anaconda3\\lib\\site-packages (from requests<3,>=2.21.0->tensorboard~=2.6->tensorflow->keras-tcn) (2020.12.5)\n",
      "Requirement already satisfied: urllib3<1.27,>=1.21.1 in c:\\users\\admin\\anaconda3\\lib\\site-packages (from requests<3,>=2.21.0->tensorboard~=2.6->tensorflow->keras-tcn) (1.26.4)\n",
      "Requirement already satisfied: idna<3,>=2.5 in c:\\users\\admin\\anaconda3\\lib\\site-packages (from requests<3,>=2.21.0->tensorboard~=2.6->tensorflow->keras-tcn) (2.10)\n",
      "Requirement already satisfied: chardet<5,>=3.0.2 in c:\\users\\admin\\anaconda3\\lib\\site-packages (from requests<3,>=2.21.0->tensorboard~=2.6->tensorflow->keras-tcn) (4.0.0)\n",
      "Requirement already satisfied: oauthlib>=3.0.0 in c:\\users\\admin\\anaconda3\\lib\\site-packages (from requests-oauthlib>=0.7.0->google-auth-oauthlib<0.5,>=0.4.1->tensorboard~=2.6->tensorflow->keras-tcn) (3.1.1)\n",
      "Requirement already satisfied: typeguard>=2.7 in c:\\users\\admin\\anaconda3\\lib\\site-packages (from tensorflow-addons->keras-tcn) (2.13.3)\n"
     ]
    }
   ],
   "source": [
    "!pip install keras-tcn"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "id": "9c3317e6",
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "from tensorflow.keras import Model, Input\n",
    "from tensorflow.keras.layers import Dense, Dropout, Embedding\n",
    "from tensorflow.keras.preprocessing import sequence\n",
    "from tcn import compiled_tcn"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "id": "c9a06ff2",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "x.shape= (None, 64)\n",
      "model.x = (None, 42, 1)\n",
      "model.y = (None, 2)\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\admin\\anaconda3\\lib\\site-packages\\keras\\optimizer_v2\\adam.py:105: UserWarning: The `lr` argument is deprecated, use `learning_rate` instead.\n",
      "  super(Adam, self).__init__(name, **kwargs)\n"
     ]
    }
   ],
   "source": [
    "model_TCN = compiled_tcn(return_sequences = False,num_feat = 1, num_classes = 2, nb_filters = 64,kernel_size=6,\n",
    "                    dilations=[1,2,4,8,16,32,64],nb_stacks=1, opt='adam',lr=0.002,activation = 'relu',kernel_initializer = \n",
    "                    'he_normal',use_batch_norm=True,use_layer_norm = False,use_weight_norm = False,max_len=x1_TCN[0:1].shape[1],\n",
    "                     use_skip_connections=True) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "b522934f",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "42\n"
     ]
    }
   ],
   "source": [
    "print(x1_TCN[0:1].shape[1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "2f285c44",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "x_train.shape = (238682, 42, 1)\n",
      "y_train.shape = (238682, 2, 1)\n",
      "x_test.shape = (82332, 42, 1)\n",
      "y_test.shape = (82332, 2, 1)\n"
     ]
    }
   ],
   "source": [
    "print(f'x_train.shape = {x1_TCN.shape}')\n",
    "print(f'y_train.shape = {y1_TCN.shape}')\n",
    "print(f'x_test.shape = {x2_TCN.shape}')\n",
    "print(f'y_test.shape = {y2_TCN.shape}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "id": "5b0cd3be",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"model\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " input_1 (InputLayer)        [(None, 42, 1)]           0         \n",
      "                                                                 \n",
      " tcn (TCN)                   (None, 64)                324480    \n",
      "                                                                 \n",
      " dense_8 (Dense)             (None, 2)                 130       \n",
      "                                                                 \n",
      " activation (Activation)     (None, 2)                 0         \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 324,610\n",
      "Trainable params: 322,818\n",
      "Non-trainable params: 1,792\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model_TCN.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "id": "921693aa",
   "metadata": {},
   "outputs": [],
   "source": [
    "y1_TCN = y1_TCN.squeeze().argmax(axis=1)\n",
    "y2_TCN = y2_TCN.squeeze().argmax(axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "id": "3f2b4124",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/10\n",
      "7459/7459 [==============================] - 1536s 205ms/step - loss: 0.1380 - accuracy: 0.9333 - val_loss: 0.2139 - val_accuracy: 0.9071\n",
      "Epoch 2/10\n",
      "7459/7459 [==============================] - 1536s 206ms/step - loss: 0.1156 - accuracy: 0.9428 - val_loss: 0.2201 - val_accuracy: 0.8992\n",
      "Epoch 3/10\n",
      "7459/7459 [==============================] - 1536s 206ms/step - loss: 0.1109 - accuracy: 0.9450 - val_loss: 0.1951 - val_accuracy: 0.9099\n",
      "Epoch 4/10\n",
      "7459/7459 [==============================] - 4071s 546ms/step - loss: 0.1086 - accuracy: 0.9463 - val_loss: 0.2296 - val_accuracy: 0.8901\n",
      "Epoch 5/10\n",
      "7459/7459 [==============================] - 1601s 215ms/step - loss: 0.1069 - accuracy: 0.9471 - val_loss: 0.1824 - val_accuracy: 0.9147\n",
      "Epoch 6/10\n",
      "7459/7459 [==============================] - 1565s 210ms/step - loss: 0.1055 - accuracy: 0.9478 - val_loss: 0.2024 - val_accuracy: 0.9071\n",
      "Epoch 7/10\n",
      "7459/7459 [==============================] - 1615s 217ms/step - loss: 0.1041 - accuracy: 0.9482 - val_loss: 0.2090 - val_accuracy: 0.9055\n",
      "Epoch 8/10\n",
      "7459/7459 [==============================] - 1542s 207ms/step - loss: 0.1032 - accuracy: 0.9489 - val_loss: 0.2064 - val_accuracy: 0.9107\n",
      "Epoch 9/10\n",
      "7459/7459 [==============================] - 1699s 228ms/step - loss: 0.1028 - accuracy: 0.9493 - val_loss: 0.2082 - val_accuracy: 0.9077\n",
      "Epoch 10/10\n",
      "7459/7459 [==============================] - 1671s 224ms/step - loss: 0.1019 - accuracy: 0.9497 - val_loss: 0.1948 - val_accuracy: 0.9094\n"
     ]
    }
   ],
   "source": [
    "history = model_TCN.fit(x1_TCN, y1_TCN, epochs = 10, validation_data = [x2_TCN, y2_TCN], workers=4, use_multiprocessing=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "id": "81bc2ee4",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2573/2573 [==============================] - 123s 48ms/step - loss: 0.1948 - accuracy: 0.9094\n",
      "Accuracy : 90.94 %\n",
      "Loss : 19.48 %\n"
     ]
    }
   ],
   "source": [
    "result = model_TCN.evaluate(x2_TCN, y2_TCN)\n",
    "print(f\"Accuracy : {result[1] * 100:.2f} %\")\n",
    "print(f\"Loss : {result[0] * 100:.2f} %\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "id": "66dcfa46",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[1 1 1 ... 0 0 0]\n",
      "[1 1 1 ... 0 0 0]\n"
     ]
    }
   ],
   "source": [
    "y_pred_TCN = model_TCN.predict(x2_TCN, verbose=0)\n",
    "print(np.argmax(y_pred_TCN, axis=1))\n",
    "y_pred_class_TCN=np.argmax(y_pred_TCN,axis=1)\n",
    "print(y_pred_class_TCN)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "id": "1d1e5dd4",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ROC AUC: 0.903312\n"
     ]
    }
   ],
   "source": [
    "# ROC AUC\n",
    "from sklearn.metrics import roc_auc_score\n",
    "auc = roc_auc_score(y2_TCN, y_pred_class_TCN)\n",
    "print('ROC AUC: %f' % auc)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "id": "93109700",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYIAAAEGCAYAAABo25JHAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuNCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8QVMy6AAAACXBIWXMAAAsTAAALEwEAmpwYAAAe80lEQVR4nO3df5xddX3n8dc7M/lNEn4k/EoICTH8CCsoDiHRYkGrArJLrVZBqw9p+6AoWH1YXVjt2lZbWxfXrdQf2YgU7VrZqmijRandLWIlAwkQQgJCwyBkCCxhBgLhR5KZ+ewf59zJzZ07MyfJnHvvuef9fDzuY+75cc/9HMLj+znn+7nn+1VEYGZm5TWp2QGYmVlzORGYmZWcE4GZWck5EZiZlZwTgZlZyXU2O4D9NXfu3Fi0aFGzwzAzK5S77rrr6YiYV29b4RLBokWLWL9+fbPDMDMrFEmPjrbNXUNmZiXnRGBmVnJOBGZmJedEYGZWck4EZmYll1sikHS9pKckbRpluyRdK2mLpI2SzsgrFjMzG12edwQ3AOeNsf18YGn6ugz4ao6xmJnZKHJ7jiAibpO0aIxdLgK+Gck42N2SDpV0TEQ8kVdMZmaNFhHsGQx2Dw6xeyB57RoYTP8OsXtwiF17hvbZvntwcJ91u9JX1/GH8foT6z4TdlCa+UDZfGBr1XJvum5EIpB0GcldAwsXLmxIcGZWXENDScO7q6bhrW5Y9zbKlcZ4sKaxHhrxmeHjjNqA7z1O9ecnatqXD5yzpO0Sgeqsq/ufKyJWA6sBurq6PJOOWYuJCAaGok4jOjh8Nbt7oKbBrGpUd9VpeEdrwHeNcZzKPgNDE9NMTBJM6ZzElI5JTJ3ckfztnJSs60zeT5s8idnTOtPljuFtlX2r95/SMYkpnR37ruus2q+jY591+3yuYxKTJtVrNg9eMxNBL3Bc1fICYFuTYjErnNqr3roN6MAQu+pcBe8eGBzlird63eCIRrf6OLuqjzOBV72TO7RPwzuiYeyYxKxpnWnj2VHVwE4a0cBWjjO1o3bdpJpjd+xz/MqxOjvK8cPKZiaCNcCVkm4EzgJ2uD5grW5gcGSXQL0ugxHdD3W6Hup3P+zbQI/V9bBncGJaXon0anTfq9XahnFWetU7pbNjn8Zyat0GduSV7dSahrduw5zjVa+NLrdEIOnbwDnAXEm9wJ8AkwEiYhVwM3ABsAV4Ebg0r1isuCJqrnpH6XoYccVbr8+3TsNa2/Uw4oq35jgT1OMwfNU7omGsWnfI1E6OmFndzdBR52p30j7dFfWOs0/DnnY9VF85d04SkhvfMsvzV0OXjLM9gCvy+n47cJWr3rH6b0cUzer0A4/WyO4aXjdY06jXb9AngkTNVWz9LoWZUzur1nXUaUhr+3Y7qhrq0bsm9vkeX/VaiyncMNRlteOlPWzY+uw+V6cji2b1C2f1GvBdYzS8gxN02ds5SfWLXlXdDzOndnLYjJF9tlM6OvZZN7LoNvLKdkpHUrirLrgN9/X6qtdsVE4EBfFnazZz0z2Pj7tfvavX2v7YGTM6R/zyYawuherj1Da8oxXdJndMosNXvWaF4ERQABHBv215mjecfCQffdOJNY343gZ8coeves1s/zkRFEDP0y/w1PO7eNOyo/gP8+c0OxwzazPl+JFswXX39AGw4oQjmhyJmbUjJ4IC6O7p56jZU1l0xIxmh2JmbciJoMVFBN09faw44Qj3/5tZLpwIWlzP0y+w/fld7hYys9w4EbS4Sn1gpROBmeXEiaDFdff0c/TsaRzv+oCZ5cSJoIVFBGsf7mPFCYe7PmBmuXEiaGEPb3+Bp3e6PmBm+XIiaGF+fsDMGsGJoIV19/S5PmBmuXMiaFHJ8wP9rg+YWe6cCFqU6wNm1ihOBC1q+PmBJU4EZpYvJ4IWtbanj2PmTGPh4a4PmFm+nAhaUERwh8cXMrMGcSJoQQ9v38nTO3ez4oTDmx2KmZWAE0ELWtvTD/j5ATNrDCeCFtTt+oCZNZATQYtxfcDMGs2JoMVU6gMedtrMGsWJoMWsfdjjC5lZYzkRtJjunn6OnTON4w6f3uxQzKwknAhaiOcnNrNmcCJoIVue2knfC7vdLWRmDeVE0EI8/4CZNYMTQQtxfcDMmsGJoEUM1weWuD5gZo2VayKQdJ6kByVtkXR1ne1zJP1Q0r2SNku6NM94Wtm/uz5gZk2SWyKQ1AF8GTgfWAZcImlZzW5XAPdHxOnAOcB/lzQlr5ha2fD8A04EZtZged4RLAe2RERPROwGbgQuqtkngFlK+kIOAfqBgRxjalndPX3MP3Q6Cw5zfcDMGivPRDAf2Fq13Juuq/Yl4BRgG3Af8OGIGKo9kKTLJK2XtH779u15xds0lfmJz/L8xGbWBHkmgnotWtQsvwXYABwLvAr4kqTZIz4UsToiuiKia968eRMdZ9P9+1M76Xd9wMyaJM9E0AscV7W8gOTKv9qlwE2R2AI8ApycY0wtyfUBM2umPBPBOmCppMVpAfhiYE3NPo8BbwSQdBRwEtCTY0wtyfUBM2umzrwOHBEDkq4EbgE6gOsjYrOky9Ptq4DPADdIuo+kK+mqiHg6r5ha0dBQUh8496QjXR8ws6bILREARMTNwM0161ZVvd8GvDnPGFrd3vqA5yc2s+bwk8VN5vGFzKzZnAiarFIfOM7zE5tZkzgRNNHQUHDHI/2+GzCzpnIiaCLXB8ysFTgRNJHrA2bWCpwImmjtw30sOMz1ATNrLieCJknqA32+GzCzpnMiaJKHnnqeZ17c40RgZk3nRNAk3Q8n9YGzFrtQbGbN5UTQJN09/a4PmFlLyJwIJM3MM5AycX3AzFrJuIlA0msl3Q88kC6fLukruUfWxlwfMLNWkuWO4H+QTCDTBxAR9wKvzzOodlepD/hBMjNrBZm6hiJia82qwRxiKY21PX0cd/h0Fhzm+oCZNV+WRLBV0muBkDRF0sdIu4ls/w2PL7TY3UJm1hqyJILLgStIJp7vJZlb+IM5xtTWHvx/z/Os6wNm1kKyTExzUkS8p3qFpNcBv8gnpPZWGV/oLNcHzKxFZLkj+JuM6yyDbtcHzKzFjHpHIGkl8FpgnqSPVm2aTTIHse2nSn3gTacc1exQzMyGjdU1NAU4JN1nVtX654B35BlUu3J9wMxa0aiJICJ+BvxM0g0R8WgDY2pbayvPDyxxIjCz1pGlWPyipGuAU4FplZUR8YbcompT3T19LDx8BvMPnd7sUMzMhmUpFn8L+CWwGPgz4FfAuhxjakt75yf2r4XMrLVkSQRHRMTXgT0R8bOI+F1gRc5xtZ1fPvk8O15yfcDMWk+WrqE96d8nJL0V2AYsyC+k9rT3+QEnAjNrLVkSwZ9LmgP8EcnzA7OBj+QZVDtyfcDMWtW4iSAifpS+3QGcC8NPFltGlfrAW0718wNm1nrGeqCsA3gnyRhDP4mITZIuBD4BTAde3ZgQi++BJ59jx0t7WOmfjZpZCxrrjuDrwHHAncC1kh4FVgJXR8QPGhBb2+ju6QfgLI84amYtaKxE0AWcFhFDkqYBTwOviIgnGxNa++ju6eP4I2ZwrOsDZtaCxvr56O6IGAKIiJeBh/Y3CUg6T9KDkrZIunqUfc6RtEHSZkk/25/jF8HQUHCn5x8wsxY21h3ByZI2pu8FLEmXBUREnDbWgdMaw5eBN5HMY7BO0pqIuL9qn0OBrwDnRcRjko488FNpTZX6wIolfpDMzFrTWInglIM89nJgS0T0AEi6EbgIuL9qn3cDN0XEYwAR8dRBfmfLcX3AzFrdWIPOHexAc/OB6rmOe4GzavY5EZgs6VaSEU6/GBHfrD2QpMuAywAWLlx4kGE1lusDZtbqMk1ef4BUZ13ULHcCrwHeCrwF+K+SThzxoYjVEdEVEV3z5s2b+EhzMjgU3NHTx0o/TWxmLSzLk8UHqpfk56cVC0iGp6jd5+mIeAF4QdJtwOnAQznG1TAPPPEcz7084PGFzKylZbojkDRd0kn7eex1wFJJiyVNAS4G1tTs84/A2ZI6Jc0g6Tp6YD+/p2V5fmIzK4JxE4Gk/whsAH6SLr9KUm2DPkJEDABXAreQNO7/EBGbJV0u6fJ0nwfS424keXDtuojYdIDn0nK6e/pZdMQMjpnj+oCZta4sXUN/SvILoFsBImKDpEVZDh4RNwM316xbVbN8DXBNluMVyeBQcOcjfVzwymOaHYqZ2ZiydA0NRMSO3CNpM64PmFlRZLkj2CTp3UCHpKXAHwK35xtW8bk+YGZFkeWO4EMk8xXvAv6eZDjqj+QYU1vo7ulzfcDMCiHLHcFJEfFJ4JN5B9MuBtP5By48zfUBM2t9We4IviDpl5I+I+nU3CNqAw888RzPuz5gZgUxbiKIiHOBc4DtwGpJ90n647wDK7Lh+oDHFzKzAsj0QFlEPBkR1wKXkzxT8Kk8gyq67p4+Fs+dydFzpjU7FDOzcWV5oOwUSX8qaRPwJZJfDC3IPbKCqtQHVvjXQmZWEFmKxX8LfBt4c0TUjhVkNVwfMLOiGTcRRMSKRgTSLtY+nNQHnAjMrChGTQSS/iEi3inpPvYdPjrTDGVl1d3TxwlzZ3LUbNcHzKwYxroj+HD698JGBNIOBtP5iS88/dhmh2JmltmoxeKIeCJ9+8GIeLT6BXywMeEVy/3bnuP5XQMuFJtZoWT5+eib6qw7f6IDaQeV5wdcHzCzIhmrRvABkiv/EyRtrNo0C/hF3oEVkesDZlZEY9UI/h74MfCXwNVV65+PiP5coyog1wfMrKjGSgQREb+SdEXtBkmHOxnsa/O2Ha4PmFkhjXdHcCFwF8nPR1W1LYATcoyrcCr1gZWuD5hZwYyaCCLiwvTv4saFU1zdPf2cMG8mR7o+YGYFk2WsoddJmpm+/x1JX5C0MP/QimNgcIh1j/T710JmVkhZfj76VeBFSacD/xl4FPi7XKMqmPufqDw/4ERgZsWTdfL6AC4CvhgRXyT5Camlhp8fWOxCsZkVT5bRR5+X9F+A9wJnS+oAJucbVrG4PmBmRZbljuBdJBPX/25EPAnMB67JNaoCcX3AzIouy1SVTwLfAuZIuhB4OSK+mXtkBbF5m+sDZlZsWX419E7gTuC3gXcCd0h6R96BFcXe8YVcHzCzYspSI/gkcGZEPAUgaR7wL8B38wysKLp7+lgybyZHznJ9wMyKKUuNYFIlCaT6Mn6u7Q0MDrHuV8+4W8jMCi3LHcFPJN1CMm8xJMXjm/MLqTg2b3uOna4PmFnBZZmz+OOSfgv4NZLxhlZHxPdzj6wAKvWBs1wfMLMCG2s+gqXA54ElwH3AxyLi8UYFVgSuD5hZOxirr/964EfA20lGIP2b/T24pPMkPShpi6Srx9jvTEmDRfo1kusDZtYuxuoamhURX0vfPyjp7v05cPoE8pdJprrsBdZJWhMR99fZ73PALftz/GbblNYHVi5xIjCzYhsrEUyT9Gr2zkMwvXo5IsZLDMuBLRHRAyDpRpLxiu6v2e9DwPeAM/cz9qYarg8sdiIws2IbKxE8AXyhavnJquUA3jDOsecDW6uWe4GzqneQNB94W3qsUROBpMuAywAWLmyNEbC7e/p4xZGHMG/W1GaHYmZ2UMaamObcgzy26qyLmuW/Bq6KiEGp3u7DsawGVgN0dXXVHqPhKuMLve2M+c0OxczsoGV5juBA9QLHVS0vALbV7NMF3JgmgbnABZIGIuIHOcZ10DZte44Xdg+6UGxmbSHPRLAOWCppMfA4cDHw7uodqqfBlHQD8KNWTwLg+oCZtZfcEkFEDEi6kuTXQB3A9RGxWdLl6fZVeX133tY+7PqAmbWPcROBkn6b9wAnRMSn0/mKj46IO8f7bETcTM1wFKMlgIh4f6aIm2zP4BDrf9XPb52xoNmhmJlNiCyDx30FWAlcki4/T/J8QCltenyH6wNm1laydA2dFRFnSLoHICKekTQl57haVndPP+DxhcysfWS5I9iTPv0bMDwfwVCuUbWw7p4+lh55CHMPcX3AzNpDlkRwLfB94EhJfwH8G/DZXKNqUZX6gLuFzKydZBmG+luS7gLeSPKQ2G9GxAO5R9aCXB8ws3aU5VdDC4EXgR9Wr4uIx/IMrBWt9fwDZtaGshSL/4mkPiBgGrAYeBA4Nce4WlJ3T7/rA2bWdrJ0Db2yelnSGcAf5BZRi6rUB97xGj8/YGbtZb8noU+Hny7UkNET4b7Hd/Ci6wNm1oay1Ag+WrU4CTgD2J5bRC2qMr7Q8sWuD5hZe8lSI5hV9X6ApGbwvXzCaV3dPf2ceJTrA2bWfsZMBOmDZIdExMcbFE9Lcn3AzNrZqDUCSZ0RMUjSFVRqG3tdHzCz9jXWHcGdJElgg6Q1wHeAFyobI+KmnGNrGXvnH3B9wMzaT5YaweFAH8m8wpXnCQIoVSI46ahZHOH6gJm1obESwZHpL4Y2sTcBVDR93uBGSeoDz/DOLtcHzKw9jZUIOoBDyDYJfdva2LuDl/a4PmBm7WusRPBERHy6YZG0KD8/YGbtbqwni+vdCZSO6wNm1u7GSgRvbFgULWr3QFIfWOHRRs2sjY2aCCKiv5GBtKL7Hn/W9QEza3v7Pehcmeydn9iJwMzalxPBGLp7+jj56FkcPnNKs0MxM8uNE8Eo9tYHfDdgZu3NiWAUe+sDLhSbWXtzIhhFpT6wfLHvCMysvTkRjML1ATMrCyeCOlwfMLMycSKoY2Ov6wNmVh5OBHXsnX/AdwRm1v5yTQSSzpP0oKQtkq6us/09kjamr9slnZ5nPFl19/Rz8tGzOMz1ATMrgdwSQTrf8ZeB84FlwCWSltXs9gjw6xFxGvAZYHVe8WS1e2CI9Y/2uz5gZqWR5x3BcmBLRPRExG7gRuCi6h0i4vaIeCZd7AaaPvvLxt5neXnPkBOBmZVGnolgPrC1ark3XTea3wN+XG+DpMskrZe0fvv27RMY4kien9jMyibPRJB5ZjNJ55IkgqvqbY+I1RHRFRFd8+bNm8AQR1qbPj/g+oCZlUWeiaAXOK5qeQGwrXYnSacB1wEXRURfjvGMa9fAIHc96ucHzKxc8kwE64ClkhZLmgJcDKyp3kHSQuAm4L0R8VCOsWSysXcHL+8ZYuUSJwIzK4+x5iw+KBExIOlK4BagA7g+IjZLujzdvgr4FHAE8BVJAAMR0ZVXTOPpfrgPyfUBMyuX3BIBQETcDNxcs25V1fvfB34/zxj2R/cjfZx89GwOneH6gJmVh58sTu2tD/huwMzKxYkgVakPuFBsZmXjRJBa6/qAmZWUE0EqmX/A9QEzKx8nAvbWB1a6W8jMSsiJALh36w52DQy5UGxmpeREQNItJMFy1wfMrIScCEgSwSmuD5hZSZU+EXh8ITMru9Ingg2PPev6gJmVWukTQXdPf/r8gO8IzKycnAh6+lh2zGzmzJjc7FDMzJqi1Ing5T2D3P2Y6wNmVm6lTgT3bq3UB5wIzKy8Sp0IKvWB5YtcKDaz8ip5InB9wMystIng5T2D3OX6gJlZeRPBhq3Pstv1ATOz8iaC4fGFXB8ws5IrdSI49VjXB8zMSpkIkucHnmWFnyY2MytnInB9wMxsr1Imgkp94EzPP2BmVs5EsPbhtD4w3fUBM7PSJYKX9wxyz1bXB8zMKkqXCO55zPUBM7NqpUsE3T19THJ9wMxsWCkTwanHznF9wMwsVapEMFwf8LSUZmbDSpUIXB8wMxupVIlgbVof6PL4QmZmw3JNBJLOk/SgpC2Srq6zXZKuTbdvlHRGnvG4PmBmNlJuiUBSB/Bl4HxgGXCJpGU1u50PLE1flwFfzSuel/cMsuEx1wfMzGrleUewHNgSET0RsRu4EbioZp+LgG9Gohs4VNIxeQRz92PPsHtwiJVLXB8wM6uWZyKYD2ytWu5N1+3vPki6TNJ6Seu3b99+QMFM7pjEuSfNc33AzKxGnolAddbFAexDRKyOiK6I6Jo3b94BBXPmosP520uXM3ua6wNmZtXyTAS9wHFVywuAbQewj5mZ5SjPRLAOWCppsaQpwMXAmpp91gDvS389tALYERFP5BiTmZnV6MzrwBExIOlK4BagA7g+IjZLujzdvgq4GbgA2AK8CFyaVzxmZlZfbokAICJuJmnsq9etqnofwBV5xmBmZmMr1ZPFZmY2khOBmVnJORGYmZWcE4GZWckpqdcWh6TtwKMH+PG5wNMTGE4R+JzLwedcDgdzzsdHRN0ncguXCA6GpPUR0dXsOBrJ51wOPudyyOuc3TVkZlZyTgRmZiVXtkSwutkBNIHPuRx8zuWQyzmXqkZgZmYjle2OwMzMajgRmJmVXFsmAknnSXpQ0hZJV9fZLknXpts3SjqjGXFOpAzn/J70XDdKul3S6c2IcyKNd85V+50paVDSOxoZXx6ynLOkcyRtkLRZ0s8aHeNEy/D/9hxJP5R0b3rOhR7FWNL1kp6StGmU7RPffkVEW71Ihrx+GDgBmALcCyyr2ecC4MckM6StAO5odtwNOOfXAoel788vwzlX7fd/SUbBfUez427Av/OhwP3AwnT5yGbH3YBz/gTwufT9PKAfmNLs2A/inF8PnAFsGmX7hLdf7XhHsBzYEhE9EbEbuBG4qGafi4BvRqIbOFTSMY0OdAKNe84RcXtEPJMudpPMBldkWf6dAT4EfA94qpHB5STLOb8buCkiHgOIiKKfd5ZzDmCWJAGHkCSCgcaGOXEi4jaScxjNhLdf7ZgI5gNbq5Z703X7u0+R7O/5/B7JFUWRjXvOkuYDbwNW0R6y/DufCBwm6VZJd0l6X8Oiy0eWc/4ScArJNLf3AR+OiKHGhNcUE95+5ToxTZOozrra38hm2adIMp+PpHNJEsGv5RpR/rKc818DV0XEYHKxWHhZzrkTeA3wRmA6sFZSd0Q8lHdwOclyzm8BNgBvAJYAP5X084h4LufYmmXC2692TAS9wHFVywtIrhT2d58iyXQ+kk4DrgPOj4i+BsWWlyzn3AXcmCaBucAFkgYi4gcNiXDiZf1/++mIeAF4QdJtwOlAURNBlnO+FPirSDrQt0h6BDgZuLMxITbchLdf7dg1tA5YKmmxpCnAxcCamn3WAO9Lq+8rgB0R8USjA51A456zpIXATcB7C3x1WG3cc46IxRGxKCIWAd8FPljgJADZ/t/+R+BsSZ2SZgBnAQ80OM6JlOWcHyO5A0LSUcBJQE9Do2ysCW+/2u6OICIGJF0J3ELyi4PrI2KzpMvT7atIfkFyAbAFeJHkiqKwMp7zp4AjgK+kV8gDUeCRGzOec1vJcs4R8YCknwAbgSHguoio+zPEIsj47/wZ4AZJ95F0m1wVEYUdnlrSt4FzgLmSeoE/ASZDfu2Xh5gwMyu5duwaMjOz/eBEYGZWck4EZmYl50RgZlZyTgRmZiXnRGAtKR0tdEPVa9EY++6cgO+7QdIj6XfdLWnlARzjOknL0vefqNl2+8HGmB6n8t9lUzri5qHj7P8qSRdMxHdb+/LPR60lSdoZEYdM9L5jHOMG4EcR8V1JbwY+HxGnHcTxDjqm8Y4r6RvAQxHxF2Ps/36gKyKunOhYrH34jsAKQdIhkv5PerV+n6QRI41KOkbSbVVXzGen698saW362e9IGq+Bvg14RfrZj6bH2iTpI+m6mZL+KR3/fpOkd6Xrb5XUJemvgOlpHN9Kt+1M//7v6iv09E7k7ZI6JF0jaZ2SMeb/IMN/lrWkg41JWq5knol70r8npU/ifhp4VxrLu9LYr0+/5556/x2thJo99rZfftV7AYMkA4ltAL5P8hT87HTbXJKnKit3tDvTv38EfDJ93wHMSve9DZiZrr8K+FSd77uBdL4C4LeBO0gGb7sPmEkyvPFm4NXA24GvVX12Tvr3VpKr7+GYqvapxPg24Bvp+ykko0hOBy4D/jhdPxVYDyyuE+fOqvP7DnBeujwb6Ezf/wbwvfT9+4EvVX3+s8DvpO8PJRmDaGaz/739au6r7YaYsLbxUkS8qrIgaTLwWUmvJxk6YT5wFPBk1WfWAden+/4gIjZI+nVgGfCLdGiNKSRX0vVcI+mPge0kI7S+Efh+JAO4Iekm4GzgJ8DnJX2OpDvp5/txXj8GrpU0FTgPuC0iXkq7o07T3lnU5gBLgUdqPj9d0gZgEXAX8NOq/b8haSnJSJSTR/n+NwP/SdLH0uVpwEKKPR6RHSQnAiuK95DMPvWaiNgj6VckjdiwiLgtTRRvBf5O0jXAM8BPI+KSDN/x8Yj4bmVB0m/U2ykiHpL0GpLxXv5S0j9HxKeznEREvCzpVpKhk98FfLvydcCHIuKWcQ7xUkS8StIc4EfAFcC1JOPt/GtEvC0trN86yucFvD0iHswSr5WDawRWFHOAp9IkcC5wfO0Oko5P9/ka8HWS6f66gddJqvT5z5B0YsbvvA34zfQzM0m6dX4u6VjgxYj4X8Dn0++ptSe9M6nnRpKBws4mGUyN9O8HKp+RdGL6nXVFxA7gD4GPpZ+ZAzyebn5/1a7Pk3SRVdwCfEjp7ZGkV4/2HVYeTgRWFN8CuiStJ7k7+GWdfc4BNki6h6Qf/4sRsZ2kYfy2pI0kieHkLF8YEXeT1A7uJKkZXBcR9wCvBO5Mu2g+Cfx5nY+vBjZWisU1/plkXtp/iWT6RUjmibgfuFvJpOX/k3Hu2NNY7iUZmvm/kdyd/IKkflDxr8CySrGY5M5hchrbpnTZSs4/HzUzKznfEZiZlZwTgZlZyTkRmJmVnBOBmVnJORGYmZWcE4GZWck5EZiZldz/B9jXpYSTLzwkAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fpr, tpr, _ = metrics.roc_curve(y2,  y_pred_class_TCN)\n",
    "#create ROC curve\n",
    "plt.plot(fpr,tpr)\n",
    "plt.ylabel('True Positive Rate')\n",
    "plt.xlabel('False Positive Rate')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "id": "3013c1b3",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy with TCN classifier model =  90.93912452023515\n",
      "Confusion Matrix = [[31200  1660]\n",
      " [ 5800 43672]]\n",
      "Recall = 0.9093912452023515\n",
      "Precision = 0.9154322918826371\n",
      "Classification Report =\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.84      0.95      0.89     32860\n",
      "           1       0.96      0.88      0.92     49472\n",
      "\n",
      "    accuracy                           0.91     82332\n",
      "   macro avg       0.90      0.92      0.91     82332\n",
      "weighted avg       0.92      0.91      0.91     82332\n",
      "\n",
      "F1 Score =  0.9072631692529096\n"
     ]
    }
   ],
   "source": [
    "from sklearn import metrics\n",
    "from sklearn.metrics import f1_score\n",
    "print('Accuracy with TCN classifier model = ', metrics.accuracy_score(y_pred_class_TCN, y2_TCN)*100)\n",
    "print(\"Confusion Matrix =\", metrics.confusion_matrix(y_pred_class_TCN, y2_TCN, labels=None, \n",
    "                                              sample_weight=None))\n",
    "print(\"Recall =\", metrics.recall_score(y_pred_class_TCN, y2_TCN, labels=None, \n",
    "                                             pos_label=1, average='weighted', \n",
    "                                             sample_weight=None))\n",
    "print(\"Precision =\", metrics.precision_score(y_pred_class_TCN, y2_TCN, labels=None, \n",
    "                                             pos_label=1, average='weighted', \n",
    "                                             sample_weight=None))\n",
    "print(\"Classification Report =\\n\", metrics.classification_report(y_pred_class_TCN, y2_TCN, \n",
    "                                                                 labels=None, \n",
    "                                                                 target_names=None, \n",
    "                                                                 sample_weight=None, \n",
    "                                                                 digits=2, \n",
    "                                                                 output_dict=False))\n",
    "print(\"F1 Score = \",f1_score(y_pred_class_TCN, y2_TCN, average='macro'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "id": "4b7d0788",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:>"
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "cm_TCN = metrics.confusion_matrix(y_pred_class_TCN, y2, labels=None, \n",
    "                                              sample_weight=None)\n",
    "import seaborn as sns\n",
    "sns.heatmap(cm_TCN, annot=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "id": "da1148da",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:>"
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAV0AAAD4CAYAAABPLjVeAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuNCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8QVMy6AAAACXBIWXMAAAsTAAALEwEAmpwYAAAYgUlEQVR4nO3deXgV1f3H8ff3JkAQqAokoIAsCgIC7rhURFEUXECLAoJYFZoioqVaFapFbXGjLrUFpPlZpNVWXGppKFBq1RYoVEBRFBFMQSBsYZElCGQ7vz+SX343ZLtpbk4m4+f1PPM8d+bMnDnzlH785tyZO+acQ0RE/IjU9gBERL5JFLoiIh4pdEVEPFLoioh4pNAVEfEosaZP0GPi33V7hJSybOLltT0ECaCkRKy6fTQ8c2zMmXNo5ZRqn6+qVOmKiHhU45WuiIhXFuxaUqErIuESSajtEVRIoSsi4WLep2mrRKErIuGi6QUREY9U6YqIeKRKV0TEI1W6IiIeBfzuhWDX4SIiVWWR2JfKujLrZ2ZrzSzDzMaX0X6Jme0zs4+KlomV9alKV0TCJU7TC2aWAEwF+gKZwHIzS3fOfXbUroucc9fE2q8qXREJl/hVuj2BDOfceudcDjALGFjd4Sl0RSRcqhC6ZpZqZiuiltSonloBm6PWM4u2He0CM/vYzOab2WmVDU/TCyISLgmxf5HmnEsD0sppLmue4uhfMPsQaOucyzazq4DZQMeKzqlKV0TCxSz2pWKZQJuo9dbA1ugdnHP7nXPZRZ/nAfXMrHlFnSp0RSRc4jenuxzoaGbtzaw+MBRIL3Eqs5ZmheltZj0pzNTdFXWq6QURCZc43b3gnMszs7HAAiABmOGcW21mo4vapwM3AHeYWR5wCBjqnKvwR9QVuiISLnF8DLhoymDeUdumR32eAkypSp8KXREJFz0GLCLiUcAfA1boiki46FfGREQ80vSCiIhHqnRFRDxS6IqIeKQv0kREPNKcroiIR5peEBHxSJWuiIg/ptAVEfFHoSsi4pFFFLoiIt6o0hUR8UihKyLikUJXRMSnYGeuQldEwkWVroiIR5GInkgTEfFGla6IiE/BzlyFroiEiypdERGPFLoiIh7pMWAREY9U6YqIeKTQFRHxSKErIuKRQldExKdgZ65CV0TCRY8Bi4h4pOkFERGfgp25Ct1Y1E+M8NLtZ1M/MUJCxPj76iymvbeeyTd2o13zRgA0SUrkwOE8Br/wfqnjh5/fhkFntwKDtz7YwitLNwPwrYaJ/Hxwd048riFb9x7iR699woHDeZxx0rE8dE1ncvILeOCNT9m85xBNkhKZPLg7d/xupddrl9hs37aNByfcz+7duzCLcMONgxk+4rsl9nHO8dQTj7F44T9JapjEzx57ki5dT6vw2Oee+Tn/WryQUzt34bEnJgMwJ302+/ftK9W/FFKlGwI5eQWMmvkhh3LySYwYvx11Dou/2MX9b3xavM+9V3Yk+0heqWNPSWnEoLNbMSxtGbn5jhdGnMHCtbvYtOcQI3u14/31e5ixaCO392rLyF7t+MXbGdxyYVvumbWKE49vyOBzW/PMgi9I7d2eFxdu8HnZUgUJiQn86P7xdOl6GgcPZjP0xkGcf8G3OfmUU4r3WbxoIZs2fsmc+X/jk1UfM+mnj/D7WW+Ue2xKixZ8/NFK3vzTHCbcfy9frFtLm5Pakj77T0z79Yu1d7EBF/TQrXTG2cw6m9kDZvZLM3u+6HMXH4MLkkM5+QAkJhiJEcMd1X5ltxbMX7W91HHtkxuxKnMfh3MLyC9wrPhyL5d1TQHg0s7JpK/cBkD6ym306ZIMQF5+AQ3qJZBUL4G8Akfr4xuS8q0GfPDl3hq7Pqme5OQUunQ9DYBGjRrToUMHsrJ2lNjnvXff4doB12Fm9Dj9DA4c2M/OnVnlHhuJGLm5uTjnOHzkCImJicyc8SLDbh5BvXr1vF9jXWFmMS8x9NXPzNaaWYaZja9gv3PNLN/MbqiszwpD18weAGZROEuyDFhe9PnVigYQRhGD1+84j3/cfzFL/7OHTzL3F7ed3fY4dmfnsGnPoVLHZezI5qy2x3Fsw3ok1YvQq1MzWnwrCYCmjeqzKzsHgF3ZOTRtVB+A3yz6kocHdOHmC9ow6/3N3HX5yUx95z8erlLiYcuWTD5fs4buPU4vsT0rawctWrYsXm/RoiVZO3aUe2yjRo25vO8VDBl0Ha1ataZxkyas/vRTLu1zuZfrqKssYjEvFfZjlgBMBfoDXYGbzKxrOfs9BSyIZXyVTS+MBE5zzuUedZJngdXAk+UMNhVIBWh19TiannV1LGMJtAIHg194nyZJiTx3Uw9OSWlERtZBAPp3b8n8T0pXuQAbdn3NS4s3kvbdM/k6J5+127PJLzi6Ti5p7fZsbv6f5UBhoO88cAQzY/KN3cgrcDz91y/YczAnvhcocfH1wYPcO+5u7hv/Yxo3blyy0ZX+3z262irr2NtGfo/bRn4PgEcmPsiYu+7mrTffYOmSxXTsdCqpo8fU3MXUUXGcXugJZDjn1hf1OwsYCHx21H53AX8Ezo2l08qmFwqAE8vYfkJRW5mcc2nOuXOcc+eEIXCjHTicx4oNX/Htjs0ASIgYl3VNZsGnO8o95k8fbmXI9GXcNuMD9h/KZdPurwHYczCH5o0Lq9vmjeuXGaTf692eX/9jA6Mvac+099Yz9+PtDD+/TQ1cmVRXbm4u94y7m6uuvpbL+15Rqj2lRUt2bP///zjv2LGd5JSUmI5ds6bw/+dt27ZjTvpsfv7s82RkfMHGjV/WzMXUYVWZXjCzVDNbEbWkRnXVCtgctZ5ZtC36XK2A64HpsY6vstAdB7xjZvPNLK1o+SvwDvCDWE9S1x1/TD2aJBX+UdAgMcL5Jzdlw87C4Dy/Q1M27PqaHfuPlHt800aF828tj23AZV1SmFdUFf/j850MOPMEAAaceQLvfb6zxHEDzjiBRet2ceBwHkn1EnAOCpwjqV5C3K9Rqsc5xyMTH6RDhw7ccuttZe5zyaV9mJM+G+ccqz7+iMaNm5CcnBLTsVN/9Txjxt5NXl4eBfmF3y9ELMLhQ4dr7JrqKrPYl+gCsWhJi+6qjO6P/nPlF8ADzrn8WMdX4fSCc+6vZtaJwjK7VdEgMoHlVTlJXde8SQMmfec0EgwiZixYvYOF63YB0K976S/QkpvU55GBXbnzlY8AeHZoD45tWI+8Asfjcz/nwOHCuxx+s2gjTw/pzvVntWL7vsPc+9qq4j6S6kUYcOYJjP5t4S1iLy/dxLNDe5BbdBuZBMvKDz/gL+l/pmOnTgz+zkAA7hp3D9u2bQVg8JCb6HVxbxYv/CfX9O9LUlJDfjrp8QqP7XVxbwDefefvdOvWnZSUFgD0OONMBl13LZ06deLUzp19X2rgxXF6IROI/rOyNbD1qH3OAWYVnbM5cJWZ5TnnZpc7PlfGPFM89Zj495o9gdRJyybqyyApLSmx+o82nPrAgpgzZ+1TV5Z7PjNLBNYBlwFbKLyRYJhzbnU5+88E/uKce7Oic+o+XREJlXgVus65PDMbS+FdCQnADOfcajMbXdQe8zxuNIWuiIRKJI6v63HOzQPmHbWtzLB1zt0aS58KXREJlYA/kKbQFZFwCfpjwApdEQmVgGeuQldEwkU/Yi4i4pEqXRERjzSnKyLiUcAzV6ErIuGiSldExKOAZ65CV0TCJZ5PpNUEha6IhIqmF0REPAp45ip0RSRcVOmKiHgU8MxV6IpIuOiLNBERjzS9ICLikUJXRMSjgGeuQldEwkWVroiIRwHPXIWuiISL7l4QEfEoEvBSV6ErIqES8MxV6IpIuOiLNBERjwI+pavQFZFw0RdpIiIeGQpdERFvAl7oKnRFJFz0RZqIiEcBz1yFroiEix6OEBHxSHcviIh4FPBCl0htD0BEJJ4iZjEvlTGzfma21swyzGx8Ge0DzWyVmX1kZivM7KLK+lSlKyKhEq9C18wSgKlAXyATWG5m6c65z6J2ewdId845M+sBvA50rqhfVboiEipmFvNSiZ5AhnNuvXMuB5gFDIzewTmX7ZxzRauNAEclFLoiEioRi32pRCtgc9R6ZtG2EszsejP7HJgL3F7p+GK/FBGR4ItELObFzFKL5mL/b0mN6qqsWC5VyTrn/uSc6wxcB/yssvFpTldEQqUqT6Q559KAtHKaM4E2Ueutga0V9LXQzE42s+bOuV3l7adKV0RCJY7TC8uBjmbW3szqA0OB9OgdzOwUK0p5MzsLqA/srqhTVboiEirx+u0F51yemY0FFgAJwAzn3GozG13UPh0YBNxiZrnAIWBI1BdrZVLoikioxPPZCOfcPGDeUdumR31+CniqKn0qdEUkVBL0GLCIiD/6aUcREY8CnrkKXREJF/20o4iIRwHP3JoP3Xk/vLimTyF10PHnjq3tIUgAHVo5pdp9aE5XRMSjBIWuiIg/Ab9jTKErIuGi0BUR8UhzuiIiHqnSFRHxKOCFrkJXRMIlMeCpq9AVkVAJeOYqdEUkXPQYsIiIRwHPXIWuiISL7l4QEfFIP2IuIuJRwDNXoSsi4WJxfUta/Cl0RSRUVOmKiHik0BUR8Ug/eCMi4lFCpLZHUDGFroiEip5IExHxSHO6IiIeBbzQVeiKSLhEdJ+uiIg/qnRFRDxKDPikrkJXREJFla6IiEe6ZUxExKOAZ65CV0TCJeAPpAV+fCIiVRIxi3mpjJn1M7O1ZpZhZuPLaB9uZquKliVmdnplfarSFZFQidecrpklAFOBvkAmsNzM0p1zn0XttgHo7Zz7ysz6A2nAeRWOLy6jExEJCKvCUomeQIZzbr1zLgeYBQyM3sE5t8Q591XR6r+B1pV1qtAVkVAxq8piqWa2ImpJjeqqFbA5aj2zaFt5RgLzKxufphdEJFSq8nu6zrk0CqcEyuyqrEPKOeelFIbuRZWdU6ErIqESxz/fM4E2Ueutga1H72RmPYAXgf7Oud2VdarQFZFQiePDEcuBjmbWHtgCDAWGRe9gZicBbwEjnHPrYulUoSsioRKv1/U45/LMbCywAEgAZjjnVpvZ6KL26cBEoBkwrei8ec65cyrqV6ErIqESz7sDnHPzgHlHbZse9XkUMKoqfSp0RSRU9GJKERGPgh25Cl0RCZkEVboiIv4EPHMVuiISLhbwCQaFroiEiipdERGP9DZgERGPVOmKiHikd6SJiHgU8DewK3RFJFx094KIiEcBn11Q6FbV5o0b+NlD9xWvb9uSya2pdzJo6Ijibc45pj77JO8vXUSDBknc/5NJdOrctbg9Pz+fMbcNpVlyCo8/MxWAtCnPsmzpYk7p1JnxDz8OwNvz57B//z4GDbnZ09VJVX0+91EOHDxCfkEBefkFXDR8MhPHXM01vXtQ4Bw79xwg9eFX2LZzX6lj+17Yhafvu4GESISZs5fw9Etvl2gfN+Iynrjnelpf+gC79x7kgtM78PyPh5CTm8ctE15i/eZdHNu4IS8/dTsD7pzq65IDL+iVrl7XU0Vt2rYn7eU3SXv5TV6Y+RoNkpK4qPdlJfZZtnQRmZs38rs35nLPhId5fvKkEu1vvfYKJ7VrX7yenX2A1Z98zIu/f4uC/HzWZ6zjyOHDLJj7ZwYOGuLluuS/1y/1ec4f+iQXDZ8MwHO/fYeeQ57g/KFPMn/Rp0xI7V/qmEjE+MX4wQwcO40zB03ixn5n07lDy+L21i2Oo8/5ndm0bU/xth+M6MNN973IxF/NIfXGXgBMSO3H5BkLavgK65aIxb7Uyvhq57ThsHLF+5zYqg0tTjixxPZ/LXyPK64agJnRtdvpZGcfYPeunQDszNrO+0sWcdWAQcX7RyxCXm4uzjmOHDlCYmIir/3+Ja4fPIzExHper0mq78DBw8Wfj2nYAOdKv+Hl3G7t+M/mXXy5ZTe5efm8seBDrrmkR3H75B8N4sHnZ5c4Njcvn4YN6nFMw3rk5uXTvnVzTkw5jsUfZNTsBdUx8XwFe03Q9EI1vPf2fPpcUbqK2bUzi+SU/69aklNasGtnFs2aJzP1ucmkjv0hXx/8urj9mEaN6HXp5Xz/lhs585zzaNS4CWs/W80tI+/wch3y33POMWfaWJxz/OaP/2LGW/8C4JE7r2X4NT3Zl32Ifqm/LHXciSnHkrnjq+L1LTu+ome3dgBc3bs7W7P28sm6LSWO+fmMvzH1oZs4dCSXkQ/9jifuuZ5Hp/2l5i6ujgr25EI1Kl0zu62CtuI3bP5+5ov/7SkCLTc3lyWL/sHFfa4o3VhGZWMGSxf/k+OPb0qnzqeVah864nbSXn6TO35wHy/9egq3pt7J3D//kZ8+eC+vzPh1DVyBxEOf257jwmFPcd3YaXx/SC++fdbJADwydQ4d+/+EWfNXMHrIxaWOK2ve0QENk+rxwMgr+ekLc0u1r1q3hd7ffYZ+qb+kXetmbNu5D8N4+cnbmDHpFlKaNon79dVFQa90qzO98Gh5Dc65NOfcOc65c4bfWqUfVa8zli1dRMdTu9C0WfNSbc1TWrAza3vx+s6sHTRrnsLqVStZsug9hl13JZN+ch8frVjG4w+PL3HsF2vXAND6pLa8PT+diY89w4b1GWRu2lizFyT/lf/7gmznV9mkv7uKc09rV6L99fnLue6yM0odtyVrL61bHF+83qrF8WzduY8OrZNp26oZy16bwOdzH6VVynEs/cMDtGhWMlDHj+rHE2nzefD7/fnZ9Hm8Om85Y266JN6XVydZFZbaUOH0gpmtKq8JaBH/4dQd7/6t7KkFgAt7XcrsN/7ApX37s2b1Kho1bkyz5smMGjOOUWPGAfDRB8t5/Q8z+fGjT5Y49qW0Kdwz/mHy8/IoyC8ACr90OXLk8NGnkVp2TFJ9IhEj++sjHJNUn8sv6MzjafM5+aRk/rOpcA7/6t49WPfljlLHrli9kVNOSqbtic3YmrWXG688i1snzGTN+u20vWxC8X6fz32Ubw+fzO69B4u33Xztefx10Wr2HjjEMUn1KShwFBQ4jknS/D8Q+PmFyuZ0WwBXAl8dtd2AJTUyojrg8OFDfLBsKT8cP7F425y3Xgfg2u8M5rwLe/H+koWMuOEqkpKSuO+hSeV1VcLif75D5y7daJ6cAkDX7qczavj1dDi5Eyd3PDX+FyLVktKsCa89+z0AEhMSeG3+Ct5esoZXnx5Fx7YpFBQ4Nm3bw92PzQLghORjmTZxGNff9QL5+QX88KnXmTPtThIixm///G/WrN9e0emAwumHm689j2vGTAHgl6+8y6tPjyInN4/vTphZY9dalwT9MWAr65vV4kaz3wAvOecWl9H2B+fcsDIOKyHzq5zyTyDfWB373FPbQ5AAOrRySrUTc/n6fTFnzrkdjvWe0BVWus65kRW0VRq4IiLeBbvQ1S1jIhIuQX8iTaErIqES8Cldha6IhEvAM1ehKyLhYgEvdRW6IhIqAc9cha6IhEvAM1ehKyIhE/DUVeiKSKjoljEREY80pysi4pFCV0TEI00viIh4FPRKV+9IE5FQieePmJtZPzNba2YZZja+jPbOZrbUzI6Y2Y9iGZ8qXREJlzhVumaWAEwF+gKZwHIzS3fOfRa12x7gbuC6WPtVpSsioRLHd6T1BDKcc+udcznALGBg9A7OuSzn3HIgN+bxVfWCRESCrCrTC9Ev0S1aUqO6agVsjlrPLNpWLZpeEJFwqcL0gnMuDUirQk/VfhOOQldEQiWOt4xlAm2i1lsDW6vbqaYXRCRUzGJfKrEc6Ghm7c2sPjAUSK/u+FTpikioxKvOdc7lmdlYYAGQAMxwzq02s9FF7dPNrCWwAvgWUGBm44Cuzrn95fWr0BWRUInnj5g75+YB847aNj3q83YKpx1iptAVkVAJ+hNpCl0RCZWAZ65CV0RCJuCpq9AVkVDRr4yJiHikOV0REY8iCl0REZ+CnboKXREJFU0viIh4FPDMVeiKSLio0hUR8SiejwHXBIWuiIRKsCNXoSsiIRPwQlehKyLhoifSRER8CnbmKnRFJFwCnrkKXREJlxherV6rFLoiEioBz1y9mFJExCdVuiISKkGvdBW6IhIqumVMRMQjVboiIh4pdEVEPNL0goiIR6p0RUQ8CnjmKnRFJGQCnroKXREJlaA/BmzOudoewzeGmaU659JqexwSLPp38c2ix4D9Sq3tAUgg6d/FN4hCV0TEI4WuiIhHCl2/NG8nZdG/i28QfZEmIuKRKl0REY8UuiIiHil0PTGzfma21swyzGx8bY9Hap+ZzTCzLDP7tLbHIv4odD0wswRgKtAf6ArcZGZda3dUEgAzgX61PQjxS6HrR08gwzm33jmXA8wCBtbymKSWOecWAntqexzil0LXj1bA5qj1zKJtIvINo9D1o6xf4NC9eiLfQApdPzKBNlHrrYGttTQWEalFCl0/lgMdzay9mdUHhgLptTwmEakFCl0PnHN5wFhgAbAGeN05t7p2RyW1zcxeBZYCp5pZppmNrO0xSc3TY8AiIh6p0hUR8UihKyLikUJXRMQjha6IiEcKXRERjxS6IiIeKXRFRDz6X1yauX3dKK4wAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.heatmap(cm_TCN/np.sum(cm_TCN), annot=True, \n",
    "            fmt='.2%', cmap='Blues')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 191,
   "id": "d83b86d1",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "X train shape:  (238682, 42)\n",
      "X test shape:  (82332, 42)\n",
      "Y train shape:  (238682,)\n",
      "Y test shape:  (82332,)\n"
     ]
    }
   ],
   "source": [
    "# 5. TabNet CLASSIFIER MODEL\n",
    "\n",
    "print(\"X train shape: \", x1.shape)\n",
    "print(\"X test shape: \", x2.shape)\n",
    "print(\"Y train shape: \", y1.shape)\n",
    "print(\"Y test shape: \", y2.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 192,
   "id": "30fe631d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Requirement already satisfied: pytorch-tabnet in c:\\users\\admin\\anaconda3\\lib\\site-packages (3.1.1)\n",
      "Requirement already satisfied: torch<2.0,>=1.2 in c:\\users\\admin\\anaconda3\\lib\\site-packages (from pytorch-tabnet) (1.11.0)\n",
      "Requirement already satisfied: tqdm<5.0,>=4.36 in c:\\users\\admin\\anaconda3\\lib\\site-packages (from pytorch-tabnet) (4.59.0)\n",
      "Requirement already satisfied: scipy>1.4 in c:\\users\\admin\\anaconda3\\lib\\site-packages (from pytorch-tabnet) (1.6.2)\n",
      "Requirement already satisfied: scikit_learn>0.21 in c:\\users\\admin\\anaconda3\\lib\\site-packages (from pytorch-tabnet) (0.24.1)\n",
      "Requirement already satisfied: numpy<2.0,>=1.17 in c:\\users\\admin\\anaconda3\\lib\\site-packages (from pytorch-tabnet) (1.20.1)\n",
      "Requirement already satisfied: joblib>=0.11 in c:\\users\\admin\\anaconda3\\lib\\site-packages (from scikit_learn>0.21->pytorch-tabnet) (1.0.1)\n",
      "Requirement already satisfied: threadpoolctl>=2.0.0 in c:\\users\\admin\\anaconda3\\lib\\site-packages (from scikit_learn>0.21->pytorch-tabnet) (2.1.0)\n",
      "Requirement already satisfied: typing-extensions in c:\\users\\admin\\anaconda3\\lib\\site-packages (from torch<2.0,>=1.2->pytorch-tabnet) (3.7.4.3)\n"
     ]
    }
   ],
   "source": [
    "!pip install pytorch-tabnet\n",
    "from pytorch_tabnet.tab_model import TabNetClassifier\n",
    "import torch"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 228,
   "id": "1f7edea4",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Device used : cpu\n"
     ]
    }
   ],
   "source": [
    "# define the model\n",
    "model_tabnet = TabNetClassifier(optimizer_fn=torch.optim.Adam,\n",
    "                       optimizer_params=dict(lr=5e-2),\n",
    "                       scheduler_params={\"step_size\":10,\"gamma\":0.9},\n",
    "                       scheduler_fn=torch.optim.lr_scheduler.StepLR,\n",
    "                       mask_type='entmax')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 229,
   "id": "00e9dad4",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "epoch 0  | loss: 0.16036 | train_auc: 0.9888  | train_accuracy: 0.93356 | valid_auc: 0.9623  | valid_accuracy: 0.87897 |  0:02:58s\n",
      "epoch 1  | loss: 0.13573 | train_auc: 0.98917 | train_accuracy: 0.93264 | valid_auc: 0.96544 | valid_accuracy: 0.88091 |  0:04:51s\n",
      "epoch 2  | loss: 0.12792 | train_auc: 0.9899  | train_accuracy: 0.93757 | valid_auc: 0.96531 | valid_accuracy: 0.85935 |  0:06:42s\n",
      "epoch 3  | loss: 0.12862 | train_auc: 0.98928 | train_accuracy: 0.93625 | valid_auc: 0.96534 | valid_accuracy: 0.85933 |  0:08:33s\n",
      "epoch 4  | loss: 0.12427 | train_auc: 0.99069 | train_accuracy: 0.93735 | valid_auc: 0.96856 | valid_accuracy: 0.88695 |  0:10:22s\n",
      "epoch 5  | loss: 0.12147 | train_auc: 0.99058 | train_accuracy: 0.94044 | valid_auc: 0.97514 | valid_accuracy: 0.88801 |  0:12:14s\n",
      "epoch 6  | loss: 0.11984 | train_auc: 0.99145 | train_accuracy: 0.93995 | valid_auc: 0.97858 | valid_accuracy: 0.90799 |  0:14:05s\n",
      "epoch 7  | loss: 0.11932 | train_auc: 0.99144 | train_accuracy: 0.94424 | valid_auc: 0.97962 | valid_accuracy: 0.90368 |  0:15:57s\n",
      "epoch 8  | loss: 0.11659 | train_auc: 0.99142 | train_accuracy: 0.94384 | valid_auc: 0.9787  | valid_accuracy: 0.8979  |  0:17:48s\n",
      "epoch 9  | loss: 0.1155  | train_auc: 0.99183 | train_accuracy: 0.94397 | valid_auc: 0.98398 | valid_accuracy: 0.89134 |  0:19:40s\n",
      "epoch 10 | loss: 0.11377 | train_auc: 0.99213 | train_accuracy: 0.94671 | valid_auc: 0.98268 | valid_accuracy: 0.90317 |  0:21:31s\n",
      "epoch 11 | loss: 0.12047 | train_auc: 0.99197 | train_accuracy: 0.94593 | valid_auc: 0.98167 | valid_accuracy: 0.90652 |  0:27:37s\n",
      "epoch 12 | loss: 0.11336 | train_auc: 0.9922  | train_accuracy: 0.94738 | valid_auc: 0.98266 | valid_accuracy: 0.90088 |  0:29:42s\n",
      "epoch 13 | loss: 0.11337 | train_auc: 0.99208 | train_accuracy: 0.94621 | valid_auc: 0.98443 | valid_accuracy: 0.89174 |  0:31:40s\n",
      "epoch 14 | loss: 0.11307 | train_auc: 0.99225 | train_accuracy: 0.94651 | valid_auc: 0.98409 | valid_accuracy: 0.90662 |  0:33:39s\n",
      "epoch 15 | loss: 0.11229 | train_auc: 0.99221 | train_accuracy: 0.94597 | valid_auc: 0.9842  | valid_accuracy: 0.90881 |  0:35:32s\n",
      "epoch 16 | loss: 0.1155  | train_auc: 0.99234 | train_accuracy: 0.94735 | valid_auc: 0.98285 | valid_accuracy: 0.91067 |  0:37:26s\n",
      "epoch 17 | loss: 0.11032 | train_auc: 0.99225 | train_accuracy: 0.94624 | valid_auc: 0.98327 | valid_accuracy: 0.90785 |  0:39:20s\n",
      "epoch 18 | loss: 0.11228 | train_auc: 0.99223 | train_accuracy: 0.94476 | valid_auc: 0.9834  | valid_accuracy: 0.92092 |  0:41:11s\n",
      "epoch 19 | loss: 0.11049 | train_auc: 0.99243 | train_accuracy: 0.94593 | valid_auc: 0.9821  | valid_accuracy: 0.91761 |  0:43:02s\n",
      "Stop training because you reached max_epochs = 20 with best_epoch = 18 and best_valid_accuracy = 0.92092\n",
      "Best weights from best epoch are automatically used!\n"
     ]
    }
   ],
   "source": [
    "# fit the model \n",
    "model_tabnet.fit(x1,y1, eval_set=[(x1, y1), (x2, y2)],eval_name=['train', 'valid'],\n",
    "                   eval_metric=['auc','accuracy'],\n",
    "                   max_epochs=20 , patience=50, batch_size=256, virtual_batch_size=128, num_workers=0, weights=1,\n",
    "                   drop_last=False)   "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 256,
   "id": "de1d55a3",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, 'Feature index')"
      ]
     },
     "execution_count": 256,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# find and plot feature importance\n",
    "from IPython.core.pylabtools import figsize\n",
    "figsize(8, 8)\n",
    "model_tabnet.feature_importances_\n",
    "feat_importances = pd.Series(model_tabnet.feature_importances_)\n",
    "feat_importances.nlargest(10).plot(kind='barh')\n",
    "plt.title(\"Feature importance graph (UNSW NB15 Dataset) - TabNet\")\n",
    "plt.xlabel(\"Feature score\")\n",
    "plt.ylabel(\"Feature index\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 231,
   "id": "29c34133",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[1 1 0 ... 0 0 1]\n"
     ]
    }
   ],
   "source": [
    "y_pred_TabNet = model_tabnet.predict(x2)\n",
    "print(y_pred_TabNet)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 232,
   "id": "1fbfb9a4",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ROC AUC: 0.916144\n"
     ]
    }
   ],
   "source": [
    "# ROC AUC\n",
    "from sklearn.metrics import roc_auc_score\n",
    "auc = roc_auc_score(y2, y_pred_TabNet)\n",
    "print('ROC AUC: %f' % auc)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 270,
   "id": "c716e6d6",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "figsize(6, 5)\n",
    "fpr, tpr, _ = metrics.roc_curve(y2,  y_pred_TabNet)\n",
    "#create ROC curve\n",
    "plt.plot(fpr,tpr)\n",
    "plt.ylabel('True Positive Rate')\n",
    "plt.xlabel('False Positive Rate')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 234,
   "id": "497cb009",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy in TabNet classifier model: 92.09177476558324\n",
      "Confusion Matrix = [[32152  1663]\n",
      " [ 4848 43669]]\n",
      "Recall = 0.9209177476558325\n",
      "Precision = 0.924567357365638\n",
      "Classification Report =\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.87      0.95      0.91     33815\n",
      "           1       0.96      0.90      0.93     48517\n",
      "\n",
      "    accuracy                           0.92     82332\n",
      "   macro avg       0.92      0.93      0.92     82332\n",
      "weighted avg       0.92      0.92      0.92     82332\n",
      "\n",
      "F1 Score =  0.9193393993269945\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import accuracy_score\n",
    "from sklearn import metrics\n",
    "from sklearn.metrics import f1_score\n",
    "test_acc = accuracy_score(y_pred_TabNet, y2)\n",
    "print(\"Accuracy in TabNet classifier model:\",test_acc*100)\n",
    "print(\"Confusion Matrix =\", metrics.confusion_matrix(y_pred_TabNet, y2, labels=None, \n",
    "                                              sample_weight=None))\n",
    "print(\"Recall =\", metrics.recall_score(y_pred_TabNet, y2, labels=None, \n",
    "                                             pos_label=1, average='weighted', \n",
    "                                             sample_weight=None))\n",
    "print(\"Precision =\", metrics.precision_score(y_pred_TabNet, y2, labels=None, \n",
    "                                             pos_label=1, average='weighted', \n",
    "                                             sample_weight=None))\n",
    "print(\"Classification Report =\\n\", metrics.classification_report(y_pred_TabNet, y2, \n",
    "                                                                 labels=None, \n",
    "                                                                 target_names=None, \n",
    "                                                                 sample_weight=None, \n",
    "                                                                 digits=2, \n",
    "                                                                 output_dict=False))\n",
    "print(\"F1 Score = \",f1_score(y_pred_TabNet, y2, average='macro'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 251,
   "id": "915afa7c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:>"
      ]
     },
     "execution_count": 251,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x360 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "figsize(6, 5)\n",
    "cm_TabNet = metrics.confusion_matrix(y_pred_TabNet, y2, labels=None, \n",
    "                                              sample_weight=None)\n",
    "import seaborn as sns\n",
    "sns.heatmap(cm_TabNet, annot=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 252,
   "id": "49f744b1",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:>"
      ]
     },
     "execution_count": 252,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x360 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "figsize(6, 5)\n",
    "sns.heatmap(cm_TabNet/np.sum(cm_TabNet), annot=True, \n",
    "            fmt='.2%', cmap='Blues')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 237,
   "id": "bd0b5296",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of selected features=42\n",
      "Accuracy result of all models:\n",
      "Accuracy in XGBoost classifier model =  90.1229169703153\n",
      "Accuracy in DNN classifier model =  75.08016324150998\n",
      "Accuracy in CNN classifier model =  78.67657775834427\n",
      "Accuracy in TCN classifier model =  90.93912452023515\n",
      "Accuracy in TabNet classifier model: 92.09177476558324\n"
     ]
    }
   ],
   "source": [
    "print(\"Number of selected features=%d\" % (x1.shape[1]))\n",
    "print('Accuracy result of all models:')\n",
    "print('Accuracy in XGBoost classifier model = ', metrics.accuracy_score(y2, y_pred_xgb)*100)\n",
    "print('Accuracy in DNN classifier model = ', metrics.accuracy_score(y_predicted_class_DNN, y2)*100)\n",
    "print('Accuracy in CNN classifier model = ', metrics.accuracy_score(y_predicted_class_CNN, y2)*100)\n",
    "print('Accuracy in TCN classifier model = ', metrics.accuracy_score(y_pred_class_TCN, y2_TCN)*100)\n",
    "print(\"Accuracy in TabNet classifier model:\",test_acc*100)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 272,
   "id": "54804e19",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "data = {'XGBoost':90.1, 'DNN':75.1, 'CNN':78.7,\n",
    "        'TCN':90.9, 'TabNet':92.1}\n",
    "Classifier_model = list(data.keys())\n",
    "Classifier_accuracy = list(data.values())\n",
    "  \n",
    "fig = plt.figure(figsize = (5, 5))\n",
    " \n",
    "# creating the bar plot\n",
    "plt.bar(Classifier_model, Classifier_accuracy, color ='blue',\n",
    "        width = 0.3)\n",
    "plt.ylim(0, 100)\n",
    "plt.xlabel(\"Classifier\")\n",
    "plt.ylabel(\"Accuracy (%)\")\n",
    "plt.title(\"Comparative analysis before feature selection (UNSW NB15 Dataset)\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 274,
   "id": "2c0486e1",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x1c1210661f0>"
      ]
     },
     "execution_count": 274,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXYAAAF2CAYAAAB6XrNlAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuNCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8QVMy6AAAACXBIWXMAAAsTAAALEwEAmpwYAACAFElEQVR4nO2ddZwV1fvH37e3k2WLFhhSkO5GsfWH9bW+dqJiYouBIhZ2K6hfQMVWQCQFpDsHELaD7d5bM78/5u5lu7jb5/167Wt3Zs4585w7ez9z5jnPPEenqioCgUAgaD3om9oAgUAgEHgWIewCgUDQyhDCLhAIBK0MIewCgUDQyhDCLhAIBK0MIewCgUDQyjA2tQGCuiFJkgocAJyACvgAucDdsizvcJXxBZ4HLgGsrnK/AS/JslxUqq3/AncB3oAZ2Ag8JstydmP1p65IkvQqsFqW5ZWuzyJMluX0UsevAGbIsjxBkqSbgA+BobIsHyhV5ndgqSzLCyRJ8gfeBEYAiuvnfVmWP5Mk6RpglizL55SquxnoBHSQZVl17VsG/CbL8oflbFXRPvNnqrHvbeAkoANMwAngdlmWk0vVCQL+Bm4pdY37A5uB46VOebUsy3I5G9YBnYEc1y4T8AfwoizLedV81EiSNBS4VZblu6orVx8kSboNMMuy/IEkSYOBO2VZvsPT52mriBF7y2SiLMsDZVk+R5ZlCfgWeBdAkiQjsArt2g6UZbk/mmj5AX+6jiNJ0pPAbcBlsiwPBAYAdrQbQLNEkqQRQG9ZllfWoZoOWCxJklcVx+cC+cDZsiwPAC4EnpUk6VxgJdBXkqQQ1/nDgCjgFDDEtc8IjAGWVdH+w5IkjavGvg2uazlAluU+QBzwQslBSZIuALYCUrl6o4BFrrolPzKV82hJGWAg4A8sqsamEvoCHWpRrj6MQRuUIMvyTsAoSdJFDXSuNocYsbdwXMLSCch07boS0Muy/FBJGVmWCyVJmgnsBi53jTCfAAbJspzqKmOXJOlR13GzLMu2cue5CHgJ7YZRgDbSzwEOyLLs5yrTpWTbNRq9FfB1lbMAb8iy/IOr7Kuu886SJOlW4B5X2xloI9ojlXR3NvBeHT+i1WhPI68DMyo5Hgmkoo1kbbIsJ0mS9H9ApizLmZIk7QDGAT+jif6frvKXANuB4UCsLMuxVZz/KeAbSZIGyLKcVZ2hkiSZgAC0UXsJ9wPXA9+XKz4K6CZJ0i7AAcyVZfnH6toH93V+CEiRJKkXcBR4C+3m7492I7yN0zeYQEmSvkS7lhXKybK8SZKkMWhPPQa0p8NXZFn+QZIkM/AqMN51bLerP5PRPr+pkiQVybL8PvAJ2tPV7zX1QVAzYsTeMlkrSdI+SZKS0L6YADe7fo9Ce2wvg8ttsBptpNQLKJJl+Vi5MoWyLP+vElEPB74BbpZl+WzgNbSRbk30BSbIsjwR+LTERkmSDGhi9ZkkSeOB/wJjXS6PecBP5RtyuSPGoo2i64IK3AhcVcWIcDaa0KRLkrRCkqRngFxZlkvEdTkwwfX3xWjC8weaMOGq+0c15/8a7Xp8UsXxsZIk7ZEkaS+Q7DrX5yUHZVmeJsvy9krqFQCLgaFon99HkiQNqcYONy533FGgP9qNKQoY6XpiWAg8LstyPPAs2hPFzVWVczX5PPCmLMuDgVuASa79j6PddAa7noaS0G5APwG/Am+5RB1ZlrcAZ0mS1LU2fRBUjxD2lslEl8BehPY4u1aW5VOljpuqqGdBEzqFul370Wgj8d0Asiz/KMvy+bWot0+W5VzX398CIyVJigDOA466biwXAt2BfyRJ2oMm7MEl7o9SdAeSy910KsuHoUebf3Dj8lffCnzhOn/pY/vQ3BwT0W4ao4B9kiRd7CqyHJjgGn2ORXNzbQMiXDe8SVQv7KA9jQx2+ZXL43bFAO2Bj4EVkiTpqmtQluV7ZFn+QJZlpyzLh9E+34urq1MOFSiUZXkz8DRwpyRJrwNXoLntyp+vunLfAe9LkvQ/YDDwpGv/RcClwG7Xtb0M6FONTSep6HIS1AMh7C0YWZZ3AQ8CC1xuEIBNwDhJkspcW9f2OOAf4BBgkiSpR7kyXpIkLZMkKarcqRyUElFJknSSJJ3t2ldagMzl6uWXsrUQzZ1wLdrI/TPXIQPwdSkf8CA0/3V5t4VKxf/XdCC03L5wNHdOGWRZ/s11/q9KbJYkyShJ0sdAsCzLO2VZftN1w3oJuNNVdaerzcuAHa6nGgVYAUxFE6LN5c9X7ty5rn6/TjXC5Wr3XbQnqvZVlZMkySBJ0lOuid8SdGhzJDUiSZIP0Bs4KEnShZy+Mf0CfETZa1pSp8pysix/jDb6/wvtpr3PNadhAB4odW2Hod0QqsJOuZuyoH4IYW/hyLK8GG0E+ZZr11K0x/T5kiR5A7h+v4smtD/JsmxF831+7hp1IkmSxdWGryzLSeVOsxXoLUlSX9f2pWiumWzALElSySjsPzWY+yma22A08INr35/AfyRJinRt34XmMirPv0B4uUnQ5cD9JTcxSZKCXe1XOZGJ5k6YDCDLsgNNaJ9x+bdL5ix6A7tcZVQ0wXqKsv7f34FHgHWudqrF5Wp4A3imhqIXAjFAWjVtOdFcQXe4bO4MTOf0Z1olrv+F+cByWZZj0G5OJRE9O9BuYAZXcQenn/6qLCdJ0j/AObIsL3DZFAREoF3bGZIkmV3X6FPglUraxvWE0hmoagJYUAeEsLcOZgAXSJJ0nktkzkUT8Z2SJB1AE6l8YKosy3YAWZZfRhOCP12PyXvRRmCXlm/cNcF6HbDQVfYh4BpZlnOAx4DlkiRtB4rK1y3Xzk60EdlSWZaLXftWot1k/pIkaR/ayPb/SkIJS9XNBjaguUxKeADwAg646v6N5pJYWMX5i9FuPqXbvgIIBI5KknQQ2A/EUioyBe0GcjZlhX0lmluhqptIZcxBu0mWpsTHvtt1re4ALneN3qvjOuB8SZL2u+yb6XLJVMZrrnPscp0/H+0GCNrIe4KrnV1oN9CuLiHegjZB+2MN5R4DXpAkaTewDnjeddN4Ee0mtRvtKVGHdnPFZfNdkiQ94doeAvwry3JcDf0W1AKdSNsraClIkjQKeEqW5Qub2haBZ5EkaQHwvSzLNc1XCGqBGLELWgyyLP8DyJIkTWtqWwSew/WCkipE3XOIEbtAIBC0MsSIXSAQCFoZQtgFAoGgldHUKQUsaG/OJSPiVwUCgaC2GNDSYWxHS/RXhqYW9qFoIWwCgUAgqDtj0bKylqGphT0ZICurAEWp+yRuaKgfGRn5NRdsRYg+tw1En9sG9e2zXq8jONgXXBpanqYWdieAoqj1EvaSum0N0ee2gehz2+AM+1ypC1tMngoEAkErQwi7QCAQtDKEsAsEAkErQwi7QCAQtDKEsAsEAkErQwi7QCAQtDKEsAsEAkErQwi7QCAQtDJq9YKSJEkBaGtlXuRaGaX0sYFo61cGoK1gc1dtlgoTCAQCQcNQ44hdkqThaLkIelZR5BtghizLPdGWvrrdc+YJBAKBoK7UxhVzO3AvUH6B45JFdL1dC/UCLACu9Jh1AoFAIKgzNbpiZFm+DUCSpMoOR1E2CU0y0MEjlgkEDUihvZBlJ1dhU+yuPWXzdZRdWEyt5C9Qy9UpvamiouSkouSeqqqIhq7yY3qdDkVRKq9TXXtVHaviPJXWd3W+qrZr24fa261t6XQ6lGpWdKu1PVUcq67l2ratAjqnisGhYnCqGBwKBrv2t9G9z/XbrmJ0bRudpcq4tnUqRF97I13PHltND+rHmSYB01O27zqgptXVKxAa6ldvA8LC/Otdt6Ui+nzm/G/vKtYmbCTIK6DMfl1pldJVvr9MmSrKqYoTZ14mOr0B9Pryxcq3UPl+fdk9tapTzbHqy1V+tPb167a/7Dl11fayyjbUUr9K3RBUpZSY2lWMTkXbdgmwwalitCtuES4tyMZS+8qIsfP0Pn0dc3YpOnAYdTgMOpxGHU6D9mM36zFbvBrk+3ymwp6Aluy9hAgqcdnUREZGfr0ynIWF+ZOWllfnei0Z0eczJ99ewIpj6xjcfgC39LvOY+2WpujPt3EkxeN79avofQLrXL+lXWdFUbE5nNgcCna7ov1tV7A7FPd+m92pbdudOIqt2IutOIuLcRZbUaxWjDgozi1EtdvAZkW129DZbOgcNnR2O3qnHYND+21SHNqPWslvtW5r9qiAw2DCaTChGMwoRhOK0YRqMqN6m3GazChmMw6TGZ3Fgr7kx8sLo8WC0cuC0dsLo7cXZm9vTD5emH28sfh4Y/SyoDNWLbP1vc56va7aAfEZCbssy7GSJBVLkjRaluVNwA3A8jNpUyBoaNbFb8TqtDGty+QGad+RcBBH7G7Mw66sl6h7AkVVsTsUt5BqAqv9bXMo2F3CW16MtfIKdpsNh9WGYrWiFBej2KyoVivYbGC3o9qs6Bx29HYbeqcNg6NEVB3Viq63ogmvvloHSkWcemNZwbWYwOQFZgs6sxmd2YLeYkZvsYDFgsHLC4OXFyYvC0Yfb0xeFkw+3pi8vVxlLejNmlDrTCZ0uuqeLVoe9RJ2SZKWAc/KsrwDuA741BUSuQt4x4P2CQQepchRxLqETQwI60eUX4TH21cVBeuWxej8wzD3m3p6v6ricKqlRrIlAntabEvvN1tMZGUXavvL17ErWO0OFJfwqjYrqk0b5ersNnDY0dltmCsIq7PstmJ3j3D9XGJcUsdQR4+qqtOjmMyoJhOYzGAyozNZ0FkC3EJq8LJgcP02emkjXKOXF3ovC3qzheD2QeQWOdFbLNrI2CXWOrMFnV68clMXai3ssix3KfX3BaX+3gsM86xZAkHDsD7hH4ocxUzrMgmHs5ywlhvJaqPYim6FklFtecG12Z30LN7HZEcCP6hTOfTRFlSrDdVuRWezYXQJp7GUgFbqSij121dxEIQDczlRNip1e1VERQcloms2o/PWRrp6i79LdL0wemm/9Waz291QIrA61wi3wj6LWfu7GndDbQkK88fegtxPzZmmXkFJIHBT2k972h+rkJ5v51R63mnBrcKtYC+z77TwltSxOmwUnbUGtSCMFz44hqocKeM6KBFbo+rAXKnYOt3lfHESUiK4ruNmxY6fM5dkRccIZRWjqonwqBSjEcyagJp8vFENJvRevqd9uqUEtsyItkSEzWbXPstpX7DbVWFude4GQdUIYRdUSXk/ra2cv7YqP22Vk2YOBbvdibUK36+z/AS6qmJ0Cae5Gr+tWXHgpXNi0TkJwIkFJxa3v9eJSbFrI1y1AN3hIrwc6RgcR9ArdZtkQ68/7Z+1lLgJvF2Ca0bNS0HNzcXcYxiGoLCy4lpakMuNeHUuMdcZDO5TtbTJU0HzQgh7C0JVVWx2JwXF9tMuAHspkS3lIigtnNbyYly6XCl3g73ciNjuqMHPqqroUcqMdkvE1kun4KNz4qVXsOAk0DW6teB0jXLtbpeCUbFjcDowlEQ9OOzoHDb0Dnv1568EXXl3gdmM3hIAZhMH805g9mpHx6j+boHWlXY7lPLplt3nGvUajVWOepXcUxR89yTGkWPxnnBrne0WCDyJEPZmyndrjrP7WFoZt4LdrtQxlgB0qoJJdeKlc+KjV/B2/fbSKQTgxMs1wjXjwKI63W4Io+rA5LRrYusS3TKCa7dpE3VK3SbZdEbjabG0lLgOvCoIcunRbUBoAAU2ygjxafE97ZbQmUxVTrKti9/Eb8dSmXnOnbQPPquOn2LNWLd8C3oDlmHTPd62QFBXhLA3MaqqotrtqFYrik2L53UWF3N04046eeuICjRiNp727foYQS0uxqDYMTrtGBwO9E5NbHV2TXi1kDQbqs0KjjrmY9PpKp8gs/iV2mcu47+tzKdbYcRrqehuqC1n6pawKw7+ilvHWYFd6dEAou5IOowjZifmIf+H3ifI4+0LBHVFCHsTkfbdErLXr9XC1CqZZLuqmrolk2QlAqszW9B7mdEH+lUbxVB6lFxh8s3Vns7Y+mJ6tybvINuaw/W9PJ/GSFUUrJsXo/MLxXz2NI+3LxDUByHsTUTerh2Y2oXhd845pUbGmsAePVXI79uSuOnSAUREBJUR6vbRoaRnFDS1+S0Gp+JkZexaOgd0pFdID4+37zi6ESUjDq/Jd6Mzmj3evkBQH4SwNwFKcTGO9HRCL/s/Qi+6pMLxf9ccJ9FfR/TAPhgNZX3G4kWNurE9dTcZxVlc2fNSjz+JqLYirNuXog/vjrGbeJVD0HwQKtEE2JK1dDrmqOhKj8em5tEhzLeCqAvqhqIq/Bm7hmi/SPqF9vZ4+7Y9f6AW5eI18tpW574StGyEcjQB1iRN2C1RURWOqapKXGoencLbXgZHT7Pr1D5OFaYzrctkjwuvkpeGbf8KjD1GYWjfzaNtCwRnihD2JsCWlIjOaMQU1r7CsYzcYgqKHXSOEMJ+JiiqwoqY1UT4hjMwrJ/H27du/R7QYxl6hcfbFgjOFCHsTYAtKRFTRGSloX+xKfkAdAqvf456AexLP0RyQSrTOk9Cr/Psv7kj5SiOE9swD7wAvV+IR9sWCDyBEPYmwJqUiKUK/3pcah56nY6OYULY64uqqqw4uYow71AGtT/bw227wht9QzAPON+jbQsEnkIIeyOjFBfjyMjAXIl/HbSJ08hQH8ymur/II9A4mHGE+Pwkzus8CYPes5+j49hmlLSTWIZdgc5o8WjbAoGnEMLeyNQUESMmTs8MVVVZHrOaEK9ghkUM8mzb9mKs275HH9YNY/cRHm1bIPAkQtgbGWtSIkClrpicAhvZ+TY6C/96vZGzjhOTG8e5nSd4fLRu27sMtTAbr1HXovOw314g8CTiv7OROR0RE1bhWFyqlg9FRMTUnxUxqwk0BzAiYohH21XyM7DtXY7xrBEYwrt7tG2BwNMIYW9kbElJ1UTEaMLesb0Q9vpwPPskx7JPMLXzBEwGk0fbtm77HgDLcM/nmxEIPI0Q9kampoiY9kHe+HiJTA/1YUXMavxNfoyO8uzr/c7U4ziOb8F89jT0fqEebVsgaAiEsDcitYmIEfHr9SMmN47DmUeZ3GkcZoPnknGpqkrx5kXofIIwD7zQY+0KBA2JEPZGxJ1KILriiL2w2E5adrHwr9eT5SdX42v0YWy0Z6NVHP9uQTl1AsvQ6ehMXh5tWyBoKISwNyK2ZC0iprJQx7jUkjdOhbDXlfi8RA5kHGZixzF4GT0nvqrDinXr9+jbdcbYc7TH2hUIGhoh7I1IdTliSiJihLDXnRUxa/AyeDG+g2fF17ZvBWpBJpaRIrxR0LIQ/62NiDUxCXNkZKU51WNT8wjyMxPoKxZrqAtJ+SnsSdvPhA6j8DF5e6xdpSAL254/MHYdgjFS8li7AkFjIIS9EbElJVaTgz2fzmK0Xmf+jF2D2WBmYsexHm3Xum0pKAqW4Vd7tF2BoDEQwt5IKMVFODIzKhV2q91JckaBcMPUkVOFaexM3cvY6BH4mX091q7z1AkcxzZhPvs89AEVXyQTCJo7QtgbCWtSMlD54hoJp/JRVfHGaV35M3YtRr2ByR3He6xNVVW17I3eAZgHXuSxdgWCxkQIeyNhS6ouIqZk4lTEsNeWjKJMtqXsYnTUcAItnrshOk5sx5l6DPPQ6ejMnvPZCwSNiRD2RsKWlIjOZKo0IiY2NQ9fLyOhASJOurasjFuHHh1TOnlwtO6wYd36LfrQjph6etZnLxA0JkLYGwlrUiLmiKoiYvLpFO4vFkSuJdnWHLYkbWdE5BCCvYI81q5t/0rU/AwtvLGS6yQQtBTEf28jYUtKqjSVgMOpkJiWL/zrdWBV7HoUVKZ2nuixNpXCbGx7fsfYZRDGqN4ea1cgaAqEsDcC1UXEJKUX4HCqItSxluTa8tiYtIVh4YNo5+259UZt238Ap12ENwpaBULYGwF3jphKhD1WTJzWiTVxG3AoTs7t4rnRujM9Fru8EVO/qegDwz3WrkDQVAhhbwSqj4jJx2I2EB7i09hmtTjy7QWsT/yHweEDCPfxTHy5Ft64CJ2XH5ZBl3ikTYGgqRHC3gjYkpJcETEVxSg2NY+O7f3Qi4nTGlkXvxGb08Z5nSd5rE1HzE6cyTLmIZejM4ubq6B1IIS9EagqIkZRVeJFKoFaUeQoYl3CJgaG9SPKL8IjbapOO9Yt36IP7oCpl+fCJgWCpkYIeyNQVY6Y1MxCrHan8K/XgnXx/1DkKGZal8kea9N+4C/UvDQsI69B5+GFrwWCpkQIewPjLCrCkZlZ6eIaJTnYxYi9eortxayN30C/0F509K88iVpdUQpzsO76FUOnARg79PNImwJBc0EIewNjS9YiYsyRFWPYY1PzMBp0RLXzXAKr1sjKf/+mwFHo0dG6bcdP4LDjNeIaj7UpEDQXhLA3MDXliIkO88NoEJehKmxOO78dWUWv4B50DezskTadGfHY5fWY+k5GHxTpkTYFguaEUJQGxpaYWGlEjKqqxKbk0Vn416tlU9JWcqx5Hhutq6qKdctiMPtgGXypR9oUCJobQtgbGGtSIubIqAoRMZm5VgqKHcK/Xg12xcGquPX0DutOj+BuHmnTGbsHZ+IhLIMvR2cRLjBB60QIewNjS648R0ysWOO0RrYk7yDbmsP0Phd4pD3V6aB46xL0QVGY+kzwSJsCQXNECHsD4o6IqcK/rtNBh/bCFVMZTsXJX7Fr6RLQif7hvTzSpv3gatScVCwjrkGnN3qkTYGgOSKEvQGpbuI0NiWPyFBfLCYRP10Z21J3k1GcxbQukzySzlgpzsO662cMHftj7HS2BywUCJovQtgbkGojYk7li4nTKlBUhZUxa+jgF0W/UM+k0LXt+BnsViwivFHQBhDC3oDYkpLQmc2Y2rUrsz+3wEZWnlX416tgV+peThWlM63LZI+M1p2ZidgPr8XUeyKGYM+84CQQNGeEsDcgVeWIKVnjVETEVERRFVbEriHCN5wBYX3PuD13eKPJC/OQy87cQIGgBSCEvQGpatUkkYO9avalHSS5IJVpnSeh1535v6czfh/OhANYBl+K3kvcSAVtAyHsDYSzsBBHVuURMbGp+YQFeeHjZWoCy5ovqqqyImY17b3bMTh8wJm3pziwblmCLjACUx/PpSMQCJo7tYr5kiTpWuBpwATMl2X5/XLHBwEfA2YgHrheluVsz5rasnDniKls4jQlT/jXK+FgxhHi85O4vteVHhmt2w+tRclOxvu8B9AZRHijoO1Q47dHkqRoYA4wBhgI3CFJUp9yxd4GnpVleQAgA4942M4WR1URMYXFDk5lFwn/ejlUVWV5zGpCvYIZFjHozNsrzse682cM0X0xdBp45gYKBC2I2gyLpgBrZFnOlGW5AFgKXFGujAEIcP3tAxR5zsSWibWKiJj4U+KN08o4knWMmNw4pnaeiMEDudGtu34BW6GWa12sTiVoY9Tm+TQKSC61nQwMK1fmIWClJEnzgQJguEesa8HYqoiIiU1xRcRECGEvzYqY1QRZAhkROeSM23JmJ2E/uAZTr/EYQjp6wDqBoGVRG2HXA2qpbR2glGxIkuQNfA5MkWV5myRJDwFfARfW1ojQ0PpHh4SFNU+BjElJIrB//wr2peYUExJgoXuX0Hq33Vz7XF8OnTrG8eyT3HTOlUSFB1dapi59TlnzAzqzhajzbsTg23I/q9Z2nWuD6LNnqI2wJwBjS21HAEmltvsBRbIsb3Ntfwy8WBcjMjLyURS15oLlCAvzJy0tr871GhpnYSG2jEzUkLAK9h2NzaJDmF+97W6ufT4Tluz5DX+THwMCBlTat7r02ZFwgKLjO7EMv5rMQj0UtszPqjVe55oQfa49er2u2gFxbXzsq4DJkiSFSZLkA0wHVpQ6fhzoKEmS5Nq+FNheZ0tbEVVFxFjtTpIyCsTEaSlO5sRxJOsYkzuNw2wwn1FbquLEunkxuoD2mPpN8ZCFAkHLo0Zhl2U5EXgKWAvsARa5XC7LJEkaIstyFnAT8J0kSfuAW4CbG87k5o8t0RURU26d04S0fFRVTJyWZkXManyNPoyNHnHGbdmPrEfJSsQy/Gp0BvGOgKDtUqvgXlmWFwGLyu27oNTfy4HlnjWt5WJNStQiYkLLRsS4F6+OEG+cAsTnJXIg4zAXdT0PL6PXGbWlWguwbf8RQ2QvjF3OPFxSIGjJiDdPGwBbclKVETG+XkZCA85MxFoLK2LW4G30YkLHUWfclnX3b6jWAiwj/yPCGwVtHiHsDYAtKbGCGwa05F+dwv2F8ABJ+SnsSdvP+A6j8TZ6n1FbSk4K9gN/YZLGYmjnmQWvBYKWjBB2D+MsLMCRlVUhR4zDqZCQli8mTl38GbsGs8HMxI5jzrgt65ZvwWDCPPT/PGCZQNDyEcLuYWxJlUfEJGcU4nCqdBL+dVIL09iZupdx0SPxM53ZgtKOxEM4YndjPuci9D5BnjFQIGjhCGH3MKeFvWy6Xvcbp2LEzsqYtRj1BiZ3GndG7aiKooU3+rfD3O9cD1knELR8hLB7mKojYvKwmAyEB/s0kWXNg4yiTLal7mJ01HACzGd2k7PLf6NkxmMZfhU645nFwAsErQkh7B7GlpSIOTKqYkRMah4d2/uh17ftidOVsWvRo2NKp/Fn1I5qK8K240cMET0xdh3qIesEgtaBEHYPY01KrDBxqqiqa/Hqtu2GySrOZkvyDkZEDiHYK+iM2rLt/g21KFeENwoElSCE3YM4CwtwZmdXmDg9lVWE1eZs80vhrYpbj4LKuZ0nnlE7Su4pbPtXYuw5GkNYVw9ZJxC0HoSwe5CqJk7di1e34VS9OdY8NiVtZVjEIEK9Q86oLevW70CvxzK0/LIAAoEAhLB7FKtr1aTyrpjYlDwMeh1R7c4stK8lsyb+bxyKk/POcLTuSJZxnNyBeeBF6H0rT/ErELR1hLB7EJsrIsYYWjbXelxqHh3C/DAa2ubHnW8r4O/EzQwOH0B7n7B6t6MqCtZ/FqHzDcF89jQPWigQtC7aptI0ELakJMxR0WUiYlRVJTY1v03719cmbMTmtHFe50ln1I7j2CaUjFgR3igQ1IAQdg9iTUrEElnWv56VZyW/yN5m/euF9iLWxW9iYFh/ovwi6t2Oai/Gum0p+vZnYTyrza+8KBBUixB2D1FVREzJG6dtNQf7+oR/KHYWM63LmY3WbXv+QC3KwWvUtSK8USCoASHsHsKW6IqIiS6XSiA1D50OOoa1PVdMsaOYtfEb6Bfam47+FbNd1hZ7zils+5Zj7D4SQ/uzPGihQNA6EcLuIazJlUfExKXmExHig8VsaAqzmpQNiVsocBQyrcvkM2onc803gB7LMBHeKBDUBiHsHsIdERNSNiImNjWvTfrXbU4bq+P+pldwD7oGdqp3O86UYxQc2oR5wPno/UJrriAQCISwewpbYsWImNwCG1l5Vjq1b3vCvilpG3n2fM7vWv9FpVVVoXjzIgx+IZgHXFBzBYFAAAhh9xhajhjxximAXXHwV+w6ugd1pXtQ/V/5dxzfgpJ2kpBJ16EzWTxooUDQuhHC7gGcBQU4cyqJiEktiYhpWxOnW5K3k2PL5fwuZzBat1uxbvsefVhX/PqdWd52gaCtIYTdA1S5uEZqPu0CvfD1MjWFWU2CU3GyMnYdXQI6IQV3r3c7tr3LUAuysIy8Fp1O/JsKBHVBfGM8QFU5YuJS89pcqt5tKbvILM7i/C6T6x1vruRnYtu7HGO3YRgjenjYQoGg9SOE3QPYkhLRWSxlImIKix2cyiqiUxvyryuqwp+xa+joF0Xf0F71bse67XtAwTL8Ks8ZJxC0IYSwe4DKVk2KP9X21jjdmbqXtKIMpp3BaN156l8cxzdj7j8NvX+7misIBIIKCGH3ANakpAoRMbGp+QB0biMTp4qqsCJ2DZG+4Zwd1rdebaiqSvHmxei8AzEPvNDDFgoEbQch7GdIVRExcal5BPqZCfRrG2F6e9MOklKQyrTOk9DXc7LT8e9WlNTjWIZdgc7s7WELBYK2gxD2M8TmmjitLNSxrbhhVFVlRcxq2nu3Y1D4gPq14bBh3fod+tDOGHuO9rCFAkHbQgj7GeKOiIk+Lew2u5Pk9MI2k9HxQMZhEvKTOLdL/Ufrtn0rUAsyXYtTi39LgeBMEN+gM8SWlITO4lUmIiYhrQBFVduEf11VVZbHrCbUK5hh4efUqw2lIAvbnj8wdh2CMar+0TQCgUBDCPsZokXERJaJAnGnEmgDI/YjWceIzY1naueJGPT1y2Bp3f4DKE4R3igQeAgh7GeIliOmon/d18tIaKBXE1nVeCw/uZogSyAjIofUq74zLQbH0Y2Y+5+LPqC9h60TCNomQtjPAGd+Ps6cHMzRFSNiOoX7t/qVfo5l/cu/OSeZ0mk8Jr2xzvVVVcW6eRE67wDM51zcABYKBG0TIexngC1ZyxFTesTucCrEnypoE4m/VsSswd/sx+io+q1B6ji5A2fKUcxD/k+ENwoEHkQI+xlgdYc6nn45KSWjEIdTafX+9ZM5sRzJOsaUTuMxG+qe5EwLb/wWfUhHTJLI3igQeBIh7GeALTGxQkTM6VS9rVvYV8Ssxtfkw5ioEfWqbzuwEjUvXQtv1It/Q4HAk4hv1BlQsrhGaV96bGoeZpOeiBCfJrSsYYnLS+BAxhEmdRyLl7Hub9YqhdnYdv+OsfM5GKP7NICFAkHbRgj7GWBLTqqYSiAlj47t/dDrW+/E6Z8xa/A2ejG+w6h61bft+BGcdiwjrvawZQKBAISw1xt3REwp/7qiqsSdym/V/vWk/BT2pB1gQofReBvrPuHpTI/FfmQDpr5T0AdGNICFAoFACHs9qWxxjbSsIoptzlbtX/8zdg0Wg5kJHcfUua4W3rgYncUXy6BLGsA6gUAAQtjrTWXJv2Jb+RunqYVp7Ezdy7joUfiZfOtc3xG7C2fyEcxDLkdnqXt9gUBQO4Sw15PTOWJC3PtiU/Mw6HVEh7VO0fozZg1GvZHJneoenqg67Vi3fIs+OApT7wmeN04gELgRwl5PKouIiUvNJzrMF6Oh9X2s6UWZbE/dzZio4fib6/7ylf3AKtTcU9ri1PXMKSMQCGpH61OgRsKWlFjGDaOqKrEpea3Wv74ydi16dEzpPL7OdZWiXKy7fsXQaQDGDv0awDqBQFAaIez1wJmfjzM3t0xETFaelfwie6v0r2cVZ7MleQcjooYSZAmsc33bjp/AYRPhjQJBIyGEvR5UtrhGa544/StuPSoq53aaUOe6zsx47EfWYeo7CUNQVM0VBALBGSOEvR64I2IiSwl7Sh46oGP71pX8K8eaxz9JWxkWMYhQ75CaK5RCC29cAmYfLIMubSALBQJBeYSw1wNbUiJ6r7IRMXGp+USE+mAxt66JwdXx63EoTs7rPLHOdZ1xe3EmHsQy+DJ0Xq3rhicQNGeEsNcDa1IS5kpyxLQ2N0y+rYANiVsYHD6A9j5hdaqrOh0Ub1mCPjACU5+63xQEAkH9EcJeD8pHxOQW2sjKs7a6iJi18RuwO+1M6zK5znXth9ag5qS4sjfWfREOgUBQf4Sw1xF3REzk6YnA02ucth53Q6G9iHUJ/zAwrB+RvuF1qqsW52Pd+TOGDv0wdDy7gSwUCARVUauhlCRJ1wJPAyZgvizL75c7LgEfA8FACnCNLMtZHra1WVBZRExcaj4AnSJaz4h9fcImip3FnFeP0bp1509gL8Iy4j+tfnlAgaA5UuOIXZKkaGAOMAYYCNwhSVKfUsd1wK/AXFmWBwC7gccbxNpmgC2xkhwxKXm0C/TC16vuKwk1R4odxayN30j/dr3p6F+3EEVnViL2Q2sx9Z6IISS65goCgcDj1MYVMwVYI8typizLBcBS4IpSxwcBBbIsr3Btvwy8TyvFlpyI3tsbY3DpiJjWNXG6IXELBY7CevnWrVu+BZMF85DLG8AygUBQG2oj7FFAcqntZKBDqe3uQIokSZ9LkrQL+BDI95yJzQtrUhLmyNMRMUVWB6lZRa1m8Wqb08bquL/pHdKTLgGd6lTXEbcPZ/w+LIMuRe/Vem50AkFLozY+dj2gltrWAUq5NiYA42RZ3iFJ0ovAm8BNtTUiNLT+ohgW1rgCcjI5ieChg93nPXgiA4CzpfBGs6Uhz7Ps6Bry7PlcM/CiOp1HdTpI+PE7jMERRE24DF09Friujsa+zs0B0ee2QUP0uTbCngCMLbUdASSV2k4BjsmyvMO1vRjNXVNrMjLyURS15oLlCAvzJy0tr8716oszLw97Tg5qSHv3efceSQUg0MvQKLY0ZJ/tTjs/HfyTHkHdaEdEnc5jO7gKe3oC3uc+QHpmMVDsMbsa+zo3B0Sf2wb17bNer6t2QFwbV8wqYLIkSWGSJPkA04EVpY7/A4RJkjTAtX0xsLPOlrYArJUsrhGXmkegr5kgv7ov6tzc2Jy8gxxbbp1966q1AOuOnzBE9cbQeWDDGCcQCGpNjcIuy3Ii8BSwFtgDLJJleZskScskSRoiy3IRcDnwqSRJB4FJwMMNaHOTYUvSHlTKr5rUGl5McipOVsaupWtAJ6Tg7nWqa935C9gKtVzrIrxRIGhyahXHLsvyImBRuX0XlPp7KzDMs6Y1P6xJJRExwQDYHU6S0gsZ2KNdE1t25mxN2UWWNZtrpMvrJM5Kdgr2g6sxSeMxhHZsQAsFAkFtEW+e1gFbUmKZiJiEtAIUVaVT+5Y9YtdG62vo6B9N39BedapbvGUJGE0ivFEgaEYIYa8D5XPEuHOwt/A3Tnee2ktaUQbTukyu02jdkXAQZ9wezOdcgt6n7gtwCASChkEIey1x5OXizMvDUnriNCUPH4uRdoFeTWjZmaGoCn/GrCHKN4Kz2/WpuYILVXFi3bwYnX8Y5v5TG9BCgUBQV4Sw15LTE6enX7GPTc2nU7hfi54w3JN2gJTCU5zXZRJ6Xe3/HexH/kbJSsAy/CqPx6wLBIIzQwh7LbGVC3V0KgoJafktOiJGVVVWxKymvU87BrWvfRZG1VaIbcePGCIljF2HNKCFAoGgPghhryXlI2KSMwqxO5QW7V8/kHGYxPxkzutct9G6ddevqMX5Wq71Fvy0IhC0VoSw1xJbUhLmqGi3kMWmaBOnLXXErqoqy0+uJtQrhKHh59S6npKTiv3AX5ikMRjadWk4AwUCQb0Rwl5LSkIdS4hLzcds1BMZ4tOEVtWfI5nHiM2L59zOEzDoa79Oq3Xrd6A3Yh46vQGtEwgEZ4IQ9lpQWURMbGoeHdv7ode3PFeEqqosj1lFkCWQ4ZG195E7kg7jiNmJ+ZyL0PsENZyBAoHgjBDCXgvci2u4Vk1SVJX4U3ktdsWk49kn+DcnhqmdJmCq5XqkqqJo4Y1+oZj7n9fAFgoEgjNBCHstsCWXzRGTll1EkdXZYhfXWB6zGn+zH6Oiap8Fwn50A0pGnBbeaDQ3oHUCgeBMEcJeC9wRMUFBwOmJ05Yo7CdyYpGzjjOl03jMtYw/V21F2Lb/gD68O8ZurT4lkEDQ4hHCXgtsiYllImLiUvMx6HVEtfNtYsvqzoqY1fiafBgTNaLWdWx7/kAtysVLZG8UCFoEQthrgRbqWPqN0zyi2/liMrasjy8uN4GDGUeY1HEcXsba5Y9XctOw7V+BsccoDO27NbCFAoHAE7QsZWoCHHm5OPNPR8SoqkpcC83BviJ2Dd5Gb8Z3GFnrOtZt34FOj2XoFTUXFggEzQIh7DXgjohxCXtWnpW8QnuLe+M0MT+ZvWkHmNBhNN5G71rVcSTLOE5sxzzgQvR+IQ1soUAg8BRC2GugfI6YuNR8oOVNnP4ZswaLwczEjmNqVV5VXeGNviGYB0xrYOsEAoEnEcJeA9akpLIRMal56IAO7VvOxGlqwSl2ndrHuOhR+Jpq96as49g/KOkxWIZdga6W/niBQNA8EMJeAyWLa5yOiMkjItQHL3PtXuxpDvwZuxaj3sjkTuNqVV61F2PdthR9+24Yu9c+ekYgEDQPhLDXgC0pCUt0y128Or0og+2puxkTPRx/s1+t6tj2LkMtzHaFN4p/EYGgpSG+tdXgyNUiYkqSf+UV2sjMtbYo//rK2LXo0TGl0/halVfyM7DtXY7xrBEYwrs3sHUCgaAhEMJeDVVNnHYKr93It6nJKs5mS/JORkYNI8hSuzVJrVu/B8Ay/MqGNE0gEDQgQtiroUTYS1wxcaktKwf7X3HrUFGZ2mlCrco7U4/j+HcL5gHno/cLbVjjBAJBgyGEvRqsSUnofXwwBAYBmn89NMALP+/mv8ZnjjWPTUnbGB4xmFDv4BrLq6pC8eZF6HyCMA+4oBEsFAgEDYUQ9mooWVzDvWpSan6LeTFpddx6nIqTcztPrFV5x/EtKKdOaOGNJq8Gtk4gEDQkQtirQFVVrEmJbjdMkdVBamZhi/Cv59sK2JC4mSHhA2nv067G8qrDqoU3tuuCsceoRrBQIBA0JELYq8CZm4uSn++eOI0/1XLeOF0TvwG74uC8LpNqVd62dwVqQaZrcWrxLyEQtHTEt7gKyi+uEdtCJk4L7YWsT9jEwLB+RPqG11heyc/EtvcPjN2GYoyUGsFCgUDQ0AhhrwJrSUSMK11vXGoeAb5mgvya9+pB6xI2Uey0Mq3L5FqVt25fCqqCZdhVDWyZQCBoLISwV4EtMbFsRExKPp3C/Zr1QhNFjmLWxm+kf7vedPCPqrG889QJHMf+wdx/GvqAsEawUCAQNAZC2KugdI4Yu8NJckZBs/evb0jcTKGjiPO7TKmxrKqqWnijdwDmgRc2gnUCgaCxEMJeCe6IGJd/PSGtAKeiNmthtzptrI77m94hPekc0LHG8o4T21BSj2MeOh2duXb52QUCQctACHslOHNzUQoK3Mvhud84bcYx7JuStpJvL6iVb1112LBu/Q59aCdMPcc2gnUCgaAxEcJeCeVzxMSm5uNtMRIW2Dxf3LE77ayKXUePoG50D+paY3nb/j9R8zO08Ea9+BcQCFob4ltdCacjYk7niOncjCdONyfvIMeWV6vRulKYjW337xi7DMYY1bsRrBMIBI2NEPZKsCUloffxxRAYiFNRiD+V32zj1x2Kg5Wxa+ka0BkpuOY0u9ZtP4DiwDLi6kawTiAQNAVC2CtBi4jRcsSkZBRidyjNduJ0W8ousqzZTOsyqcYnCmd6DI6jGzH1m4o+oH0jWSgQCBobIezlKB8Rc/qN0+aXI8apOPkzdi2d/KPpG9qr2rKqqmqLU3v5YRl0SSNZKBAImgIh7OVw5ua4ImJcwp6Sj9moJyK0dotANyY7T+0lvSiDaV0m1zhad5zcgTNZxjzk/9CZm19fBAKB5xDCXg5bkpYjpvTiGh3a+2FoZtEjiqqwImYNUb4R9G/Xp9qyqtOuhTeGdMDUq3YLWgsEgpZL81KrZoDVHeoYhaKqxJ3Ka5b+9T1pB0gtPMW0LpPQ15CR0bb/L9S8NCwj/oNOb2gkCwUCQVMhhL0ctqRELSImIJD07CKKrM5m519XVZUVMasJ9wnjnPZnV1tWKczBtvtXDJ0GYuzQt5EsFAgETYkQ9nLYkpKwRGs5YmJdi1c3t1WT9qcfIjE/mfM612K0vuMncNjxGnFNI1knEAiaGiHspVBVFWtiYplUAga9juh2zWfEro3W1xDqFcKQ8IHVlnVmxGGX12PqOxl9UETjGCgQCJocIeylcObmoBSWjojJI6qdLyZj8/mY9qYcJjYvnvM6T8RQjb+8JLwRsw+WwZc2ooUCgaCpaT6K1QxwR8RERaOqKrGpzWviVFVVfji0jCBLIMMiB1db1hG7G2fSYSxDLkdn8W0kCwUCQXNACHsprImnI2Ky823kFdqb1cTpsewTyOn/MrXzBEx6Y5XlVKcD65Zv0QdFYeo9sREtFAgEzQEh7KWwJSei99UiYkreOG1OE6fLY1YT5BXAqMhh1ZazH1yFmpuKZeQ1IrxRIGiDCGEvhS0pCYtr1aS4lDx0QMf2zWPEfiInhqNZx7lYmorZYKqynFKUi3XXLxg6no2xY/WhkAKBoHUihN1F+YiY2NQ8wkN88DJX7fJoTJbHrMbX5MPU7tUvjGHb+TPYrVhEeKNA0GaplbBLknStJEmHJEk6JknSvdWUu1CSpJOeM6/xcOaUjYiJS81rNv71uNwEDmXITOo4Di+jpcpyzswE7IfXYuozEUNwzYtZCwSC1kmNwi5JUjQwBxgDDATukCSpQnISSZLCgdeB5rkaRQ3Ykk9HxOQX2cnItTYb//qKmNV4G70Z32FUlWVUVcW6ZQmYvLEMvrwRrRMIBM2N2ozYpwBrZFnOlGW5AFgKXFFJuc+A5z1pXGNyOiImulSq3qYX9sT8ZPamH2RCh9F4G6tems8Zvw9nwgEsgy9D59U8njQEAkHTUBthjwKSS20nAx1KF5Ak6X5gF7DFc6Y1LrakkoiYAPfi1c0hhv3PmDVYDGYmdhxTZRlVcWi51gMjMPWd1IjWCQSC5khtZgb1gFpqWwcoJRuSJPUDpgOTKSf4tSU0tP4jzLAwz4hvcloKfp070b59AKlZxYQFe9O1U4hH2q4vibkp7Dq1j0t7n0uXqHD3/vJ9ztn+B/k5KYRf9QS+4cGNbWaj4Knr3JIQfW4bNESfayPsCUDpUIwIIKnU9pVAJLADMANRkiRtkGW5+vCNUmRk5KMoas0FyxEW5k9aWl6d65VHVVUKYuPwHzqctLQ8jsZl0aGdr0faPhOWHPodo97IiNDhblvK91ktzid//bcYovtSENiTwia2uSHw1HVuSYg+tw3q22e9XlftgLg2wr4KmC1JUhhQgDY6v6PkoCzLzwHPAUiS1AVYVxdRbw5oETGFmKOjKbY5SM0sZESf8JorNiDpRRlsT93NhA6j8TdXfQGtu34BWyGWkf+pcRUlgUDQNqjRxy7LciLwFLAW2AMskmV5myRJyyRJGtLA9jUKJYtrWCKjiD+Vjwp0auKImJWxa9Hr9EzuVPWKR86sJOwHV2PqNQFDSL28YAKBoBVSq7dvZFleBCwqt++CSsrFAF08YVhjYksqFRFzLBdo2onTzOIstiTvZHTUMIIsgVWWs25ZAkYL5iEivFEgEJxGvHmKKyLGz88VEZNPgI+JID9zk9mzKm49KipTOk2osowjfj/O+H1YBl2C3jug8YwTCATNHiHsgLVUjpjY1Dw6hfs3mb86x5rLpqRtjIgYTKh35REuquLEumUxuoD2mPpNaWQLBQJBc6fNC7uqqtiSEjFHRmF3KCSlFzTpG6er4tbjVJxM7Vx1ul374XUoWUlYRlyNrpqEYAKBoG3S5oXdmZPtjohJTM/HqahN9sZpni2fjYlbGBJ+Du192lVaxllcgG3HTxiiemPsPKiRLRQIBC2BNi/sJakELFHRxJUsXt1Eyb/WxG/ArjiY1qXq0Xr2xu9RrQVYRlwjwhsFAkGltHlhL0n+ZY6KJjYlD2+LgXZB3o1uR6G9kL8T/mFg+/5E+FYeQ6/kpJCzfTmmXmMxtOvcyBYKBIKWghD2pEQMfv4YAwK0idP2/uibYCS8NmETxU4r53eZXGUZ65Zv0RlNmIdMb0TLBAJBS6PNC3vJ4hpORSHhVH6T+NeLHMWsi99I/3Z9iPaLrLSMI/EQjtjdBI/+P/Q+Vce2CwQCQZsWdndETFQ0KRmF2BwKnSMa37++IWEzhY6iKkfrqqJg3bwInX87AoZd1MjWCQSClkabFnZnTjZKURGWqCj3xGljj9itThur4/+md0hPOgd0rLSMXf4bJTMBy/Cr0Rub7sUpgUDQMmjTwl5+cQ2TUU9kqE+j2rApcQv59gLO71L5i0aqrRDb9h8wRPTE2LVVpOYRCAQNTJsW9tI5YuJS8+jY3g+DvvE+ErvTzqq49fQMOouzgrpUbuPu31GL87GMvFaENwoEglrRpoXd6oqIMfj7E5va+BOnm5O3k2PLY1oVvnUl9xS2/Ssx9hyNIaxLo9omEAhaLm1a2G1JSZijokjLKabI6mjUF5McioOVsevoFtiZnsFnVVrGuvU70BuwDBXhjQKBoPa0WWF3R8RERxOX0viLV29L2UWWNZtpXSZX6mJxJB3BcXIH5oEXovdtncvdCQSChqHNCrsj2xURExlFbGoeBr2ODmG+jXJup+Lkz5g1dPKPpk+IVOG4Ft64GJ1vCOazpzWKTQKBoPXQZoW9zOIaqXlEhvpiMhoa5dw7T+0lvTiz6tH6sU0oGbFYhl+FToQ3CgSCOtLmhd0UFUVcSl6jvZikqAorYtYQ5RtB/3Z9KhxXbUVYty1FH94d41nDG8UmgUDQumizwm5NSsTg70++zovcQnuj+dd3n9pPauEppnWZhF5X8eO37fkDtSgHL7E4tUAgqCdtVthtSUmYI6OIS9UmThtjjVNttL6acJ8wzml/dsXjeWnY9q/A2H0khvaVR8oIBAJBTbRJYS8dERObmocO6Ni+4V0x+9MPk1SQwnmdKx+tW7d+D+ixDLuywW0RCAStlzYp7O6IGNfiGu1DfPC2GBv0nKqqsiJmNe28QhgSPrCiTSnHcJzYhnnA+ej9QhrUFoFA0Lppk8JeJiImJa9RXkw6lHmUuLwEzu0yEYO+bPSNqrqyN/oGYx5wQYPbIhAIWjdtU9hdyb/swe3JyC1ucP+6NlpfRbAliOERgyscdxzbjJJ2EsuwK9GZLA1qi0AgaP20SWEviYhJLNS2Gzoi5lj2v5zIiWVq5wkY9WVdPqrdinXb9+jDumLsPqJB7RAIBG2DNinstuQkV0bHkhzsDeuKWX5yNQFmf0ZGDq1oy95lqIXZeI28Fl0lE6oCgUBQV9qckpReNSk2NY+QAAv+Pg33due/2TEczf6XKZ3GYzaYyhxT8jOw7V2O8azhGCJ6NJgNAoGgbdHmhN2RlVUqIiavwf3rK2JW42fyZUx0RTeLddtSQBXhjQKBwKO0OWEviYghLJyUjMIG9a/H5sZzKFNmUsexWAxlnwqcqcdxHN+M+exp6P3bNZgNAoGg7dFmhT3NFIRKw75xuiJmDd5Gb8Z1GFVmv6qqFG9ejM47EPPACxvs/AKBoG3S5oTdmpSEwT+AuHwVaLiJ08T8ZPalH2Rih9F4G73KHHP8uxXl1L9Yhl2BzuRVRQsCgUBQP9qcsLtTCaTk4e9jIti/YeLGV8SsxstgYULHMWX2qw4r1q3foQ/tjLHn6AY5t0AgaNu0KWFXVVULdXQl/+oU7t8gGRRTCk6x+9R+xnUYha/Jp8wx274VqAWZWEaJ8EaBQNAwtCllcWRlohQVYYyIIjG9oMH863/GrsGoNzKp49gy+5WCLGx7/sDYdQjGyIorJwkEAoEnaFPCbktKAiDXJwSnojaIfz2tMIMdqXsYGz0Cf3PZ9q3bl4KiYBl+lcfPKxAIBCW0MWHXImISdNpIvXOE50fsK2PXotfpmdJpfJn9zrSTOI5uwtz/XPQB7T1+XoFAICihTQm7liMmgJhcJ15mA2FB3h5tP7M4i60pOxkVOYxAS4B7v6qq2uLU3gGYz7nYo+cUCASC8rQpYS+JiIlL0SZO9R6eOP0rdj0AUzuXHa07Tm7HmXIU85D/Q2f27M1EIBAIytNmhF3LEaNFxMSfyvf4xGmONZd/krcxPGIwIV7Bp8/rsGnhjSEdMUnjPHpOgUAgqIw2I+yOrEyU4mKKAsOwORSPT5yuiluPoiqc23limf22AytR89K18EZ9m/m4BQJBE9JmlOZ0KgHN9+3JidM8Wz4bE7cwJHwgYT6h7v1KYTa23b9j7HwOxqjeHjufQCAQVEebEXara9WkGKcvJqOeyFCfGmrUnjXxG7ArDs7rPKnMftv2H8FpxzLiao+dSyAQCGqizQi7LTkJQ0AAJ7OddAjzw+Aht0iBvZD1CZs4p31/InxPhzE602Oxyxsw9ZuKPjDCI+cSCASC2tB2hN21uEZcar5HF69eF78Rq9PGtC6T3fu08MZF6Lz8sIjwRoFA0Mi0CWEviYhRQsMptDro5CH/epGjmLUJmzi7XV+i/SLd+x0xu3Amy5iHXI7O4uuRcwkEAkFtaRPC7sjUImKyvbUwRE+FOv6d8A9FjiKmdTntW1eddqxblqAPjsbUa3w1tQUCgaBhaBPCbkvWJk6TDQHodTo6hJ35KNrqtLEmfgN9QiQ6B3R077cfWIWal4Zl5H/Q6Q1nfB6BQCCoK8amNqAxKImIOW71JqqdDpPxzAV3Y+IW8u0FnN/1tG9dKcrFuutXDJ0GYOzQ74zPIRAIBPWhbYzYk7SImH+zHB5xw9iddlbFradn0Fl0C+xy+jw7fgSHTYQ3CgSCJqWNCHsi+vBIcgpsHlm8+p/k7eTa8sqM1p2Z8diPrMfUdxKGoKgzPodAIBDUl1q5YiRJuhZ4GjAB82VZfr/c8UuB5wEdcBK4WZblLA/bWi9UVcWalIS93xDIOfM3Th2Kg79i19EtsDM9gs46fY7Ni8Hsg2XQpZ4wWyAQCOpNjSN2SZKigTnAGGAgcIckSX1KHQ8APgQulGV5ALAPmN0QxtYHR2YmqrWYdHMgAB3bn1kM+9aUnWRZs5nWZYp7WT1n3B6ciYewDL4MnVfDLI4tEAgEtaU2rpgpwBpZljNlWS4AlgJXlDpuAu6VZTnRtb0P6ORZM+tPSY6YeNWP8GBvvC31ny92Kk5Wxqylk38H+oT0BEB1OijesgR9UCSmPhNraEEgEAgantoIexSQXGo7GehQsiHLcoYsyz8BSJLkDTwO/OxBG88Iq0vYjxRZzti/viN1D+nFmUzrMtk9WrcfWo2ak4plxDXo9G0iyEggEDRzaqNEekAtta0DlPKFJEkKBH4C9sqyvLAuRoSG1t99ERZWvVhnZ6ZhDAoisQCmnNWuxvJVoSgKq7avo3NgNJN6D0Ov0+MszCV+9694dxtIxKDRbrFvaOrbh5aM6HPbQPTZM9RG2BOAsaW2I4Ck0gUkSYoE/gTWAA/W1YiMjHwURa25YDnCwvxJS8urtkzuiRjsQWEAhPqZaixfFTtT95KUl8otfa8jI70AgOKN36BYi9ANupL09Px6tVtXatPn1oboc9tA9Ln26PW6agfEtRH2VcBsSZLCgAJgOnBHyUFJkgzAb8B3siy/VGcLG5CSiJjc7gPBSr1dMYqqsCJmNeE+7TmnfX8AnFmJ2A+vxdR7IoaQaA9aLRAIBGdGjcIuy3KiJElPAWsBM/CZLMvbJElaBjwLdAQGAUZJkkomVXfIsnxbQxldWxyZGajWYlIM/gT7WwjwMdernf3ph0gqSOHG3lej12nTEtYtS8BkwTzkMg9aLBAIBGdOrWb7ZFleBCwqt+8C1587aKYvOpWkEjjh8KFzVP1G66qqsiJmNe28QhgSPhAAR9w+nPH7sYz4D3qv1usTdDodZGWl4XDYmtSOU6f0KEqFaZ1Wjehz26CmPhuNZoKDwzAY6haY0arDOGzJ2lSAXOTFufXMwX4oUyYuL5Fre03HoDegKg6sW5agCwzH1HdyzQ20YLKy0vDy8sHXN6LRJoYrw2jU43C0rS+86HPboLo+q6pKQUEuWVlptGsXWWmZqmiWI21PYUtMBL8Aig2Wer1xqqoqy0+uJtgSxPCIwQDYD61DyU7Ca/g16Op4F21pOBw2fH0DmlTUBYK2ik6nw9c3oF5PzK1a2K1JiRQHtgPql4P9aNa/nMyN5dzOEzDqjajF+Vh3/oQhug+GzgM9bG3zRIi6QNB01Pf712qFXVUUbMlJZHoF4+dtItjfUuc2VsSsJtDsz8jIoQBYd/0CtkIsI/4jBE8gEDRbWq0vwZGViWq1Eo8fnSP86yzE/2bHcDT7X6Z3vwiTwYSSnYz94BpM0ngMoR1rbkDgUXbt2sHzzz/FggWLCQ4OAWDRoq84eHA/c+a8BsDOndv58stPychIR1EUevToyf33P0z79uHs2rWDWbMeJDq6I6qq4nDYufTS6Vx11X88ZmN+fj5z5szmlVder7Hsu+++xZ9//sGPPy7DbNaitZKTk7jvvjtZuvS3MmXHjBnCxo07AEhPT+eDD97m6FEZg8FAeHg4DzzwCNHRHSqcozTHjsnMnfsSBQUFDBx4Do888gRGY9mvf25uLi+88DRpaWmYzSYee+wpevSQyMzMYO7cl0hOTsTHx5f77nuQfv3OprCwkDlzniMhIR69Xs899zzA0KHDAbjyykvx9T29oM2rr75JeLi2qHtBQT533XUL8+bNJzJSy4S6fftW3nvvLaxWK5MmTeWOO+4BQJaP8NprL2O32wkPD+eZZ17E39+f5OQkXnrpOQoKCvDz8+Ppp58nIiKSkydP8OqrL1FUVERAQABPPTWbiIhI7HY7c+e+wJEjh7FYLDz33Bw6d+7Ca6+9zMGDB9x2njhxnOeff5mJE6ewePE3/PbbTyiKyt13z2D8+NMrpQG89958cnKyeeqp2QBs27aF999/G0Vx0rNnLx5//BliYk4wZ87z7jrZ2Vn4+/vz9dffkZycxJw5z5GfX7YPnqDVjthLImL+tfnQqR4TpytiVuNn8mV09AgAircsAaMJ89D/86idgtoxaNAQzj33Al59VXtV4sCBffz66088/vizAOzdu5sXXniGu+66j8WLf+Tbb3/mnHOG8OSTj7rbkKTeLFiwiIULF/PppwtZsuQbTp484TEb8/JyOXZMrrGcw+Fg7dpV9Ot3NuvWral1+0VFRcyYcQcDBpzD119/y8KFi5ky5TwefPBeHA5HtXVfeOEZHnzwMZYs+RFVVfntt58rlPn22//RrVt3Fi5czH//extvvjkP0ARMknrx9dff8eyzL/LCC89gtRazZMk3dOjQiYULlzB79su89NJzAOTkZGMyGVmwYJH7p0TUDx48wD333EZ8fJz7vFZrMa+88gKvvfYW33zzPUeOHGLz5k0AvP3269x6650sXLiYjh07s3jx1wB89tmHTJlyHgsWLGLChEl88skHALz55qvcdNNtLFy4mMmTp/LRR+8B8P33S/Dy8uZ//1vK/fc/zJw5swF49NEn3TZeccXVDBkynAkTJnP48EFWrlzGl18u4sMPP+P9998mNzfHbfOOHdtYseL3Mp/f3Lkv8vzzL/P1199htRazYsUf9Oghudv/6KMv8Pf355FHnnT3YerUaRX64Ala7Yi9JPlXqjGAC+roX4/NjedQpsyl3c7HYjDjSDiAM24vluFXofcOaAhzWwSb9iezcV9yzQXrwZizIxndv/rRyh133MPtt/+X779fwg8/fMvTTz+Pv792bRcs+Iz//vdW+vXr7y4/ffpVWK1WbLaKk09WqxW9Xo+fn3bTP3BgP2+//To2m42goCAeffRJOnToSFxcLPPmzSEvLxcvL29mznyE3r37snLlChYt+gq9Xk9UVBTPPPMi8+e/Rnp6Gk888Ui1o/bNmzcSFRXNtGkX8v33Szj33Gm1+oxWr/6T4OBgLr309ODi3HPPx2QyYbPZ+OWXHysI9qBBg7nqqmuxWq3uz+aCCy7m888/5vLLryhTVlEUCgsLASguLsJi0dyXR4/KXH/9TQBER3cgMDCQAwf2c8std7hvKElJifj7a9+Nw4cPoaoqd999C1arleuvv5lJk6YA8NtvP/HQQ7N48cVn3ec9dOggHTt2IioqGodD4dxzz2ft2lWMHDnaZVOB65oVExCgncPpVCgo0N72Lioqdtv61lvvYzQaURSFlJQUt02bN2/kttvuAmDgwEFkZ2eRkpJCRIR2w8nJyebzzz/mww8/R6fTsXnzJsaPn4TFYsFisXDOOYPZtGkD559/Ebm5OXzyyQfccMPNHD9+zN0Pp9NJQUEBTqcTm83mtqmEr7/+koEDBzNgwMBSfSio0AdP0KqF3enrj9VgqfPE6fKY1fgYvRnbYSSq4sS6eQk6/zBM/aY2kLWC2mAymXj22Re5+eZruf76m+jX72z3sYMHD3DffRWzWVx77Q3uv2X5MDfddC2qqpCQEM+kSVNp1y4Mu93O7NlP8uKLc+nduy9r1qxi9uyn+Oyzr3jxxWe4/vqbGD9+EgcO7Ofpp2exePGPfPrph3zyyZcEB4fw/vtvExcXw8yZj3LffXfW6IpZtuw3Jk2aysiRo3n55Rc4efIEXbt2q7H/R4/KSFKvCvsnTtREc/r0q5g+/aoKxw8c2EdoaDv3dmhoO06dOlWh3H/+cz133HEzl146jcLCAt56S1t2QZIkVq9eSbdud3PixHFOnjxBRkYGAEajkYcemsHOndt59FFtJGqz2Rg+fCR3330fmZmZ3Hvv7XTrdhZdunTl8cefqXDe9PS0CvalpWn2zZjxIA89NIN33nkDLy9vPvlkAQC33343d911C0uXfovDYeejj75025OXl8f111+J1VrMu+9+XM05Ut3C/u23i5g8+Vy3KyQ9PY3evftWatO8eS9zxx33cOpUapl+PPzw49x33534+voSGRnFhAmnw6Hz8/P59def+OqrJe59JX347rvFZfrgCVqtsFuTksj3C8XLbCAs2LvW9RLyktiffogLuk7F2+iF7dAalKwEvKbOQGcwNaDFzZ/R/WseVTc0+/fvJTAwiB07tnHzzbeX8xO7Mm7a7dx++38ByM3N4fnnXwY0V8x7730CaH7ehx++n2++WcDo0WPx9/d3f5EnTZrCvHlzyM/PJyEhwe1b7devPwEBAcTFxTJ69FjuvvtWxo2bwPjxk+jRQyI5uUwKpUrJyspk27YtPPbY01gsXowePZZffvmRmTMfQaer6BlVVdU9P6TX693++Mr44YfvKh2xT5gwucwck6qq6PUV55zefHMe06dfxZVXXsOBA/t49tkn+Oab75kx4yHmz5/HjTdeTd++/Rk0aAgmk7FUvfdISUnmrrtuoV+/sxk3bgLjxk0AIDIyinHjJrJt2xa6dOlaqd1aH8vsQafTY7UWM3fui8yf/z59+vRjyZJveOml53jttbd56aXneOyxJxk7dgLr1q3myScfYeHCJeh0Ovz9/fnllxVs2fIPs2Y9xPff/1rmcyx9DtCeVP7441c+/XRhlTZp23p+++1nwsPDGTJkGMuWnZ4LychI56OP3uXrr78lMjKKd999i3fffYuHH54FwJ9/LmPs2PHu+SGAl156jieeeJpRo8ZV6MOZ0ip97CURManGQDq190Nfhw9qRewavAwWJnYYjWotwLbjJwyREsYugxvQYkFtOHnyBF98oT0um0wmFi783H2sd+8+7N+/F9BG9iV+zaioaOx2e4W2fH39mDRpKvv3760iAZ1Kfn7FxG6qqj1yz5z5CC+9NA9//wBefPEZ/vxzWa368Oefy1BVuP32G7niiotdvto/XG4G/wrnzMrKdLsTJKk3R44crtDm3LkvcuLEv0yfflUZv/aCBYvck8cZGenu8pmZGbRrF1ahnY0b13PhhZcA0K/f2YSEhBIbe5Li4iJmzXqar776llmzniYpKYno6A7s3r2T9HSt3YiISPr1O5uTJ/9l06YNHDlyqMxnWX6itjRhYe1JT89wb2dkZNCuXTtOnPgXi8VCnz7awvCXXjqd3bt3kpWVRVxcDGPHTgBgwoTJZGZmkJ2dzerVf6Gq2vUcMWIUNpuVvLw81znSy51D+wwOHNhHx46daN8+vJxN5T+zdqxevZJt27Zw003X8tlnH7Fx49+8884b7N27h27dziI6ugN6vZ6LL76M3bt3uutv2LCOKVPOc2+X9KHkBli6D56gVQq7liPGSqzTh051eDEppSCVPaf2M67DKHxMPlh3/4ZanI9l5LUivLGJsVqtPPfcE9xzzwNER3fg6aef54cfvuPAgf0A3HLLnSxY8FmZCIfjx4+RlJSIwWCo0J7T6WT37p307NmLTp06k5OTw+HDBwFYvfovwsMjiYiIICoqmvXrtQnOAwf2k5mZQbduZ3HNNZcTFBTEDTfczLRpF7qjVJxOZ7X9WL78d5566jmWLv2NpUt/45dfVhAQEMDq1X/h4+NLx44dWbdutbv8r7/+xJAhwwDtSSI5OZnff//ZffyPP35l9+6ddOhQdaRWREQkZrOZffv2ALBixTJGjBhVoVz37j3ZsGEdAPHxcaSnp9GxY2eWLv2Wn3/+AYAtW/5BUZx0796TzZs38s03CwAtWufIkUP07t2XlJQkvvjiUxRFITMzg40b/2bUqDFV2tenTz/i42OJj4/D6XTy119/MmLEaKKjO3LqVCpxcTEAbNiwnl69+hAUFITZbGHv3t0A7Nu3B29vX4KDg1my5Gv+/nstoEVSBQYGERQUxMiRo1mx4g8A9u7dg9lscbthDhzYz9lnDyxj04gRo1i/fg3FxcVkZWWxc+d2hgwZxvz5H/D119+xYMEibrvtLsaMGcf99z9Mt25ncejQQTIztRvUxo3r6d1bW2hOVVVk+Qh9+56e/ynpw549uyr0wRPoSu5uTUQX4KSn0/bm79tD0jvz+Tp6GtOuGF9r98GCg0vYm7afF0Y9gW9RAQXfP4mpxyi8xt9aZ9saisZMbZqSEktEROdGOVd1GI165s6dQ15entutArBixR98+eWnfPnlInx8fNi3bw9ffPEJWVmZFBYWER4ezv/931VMmjSlTLijTqdFpnTv3pPHHnsKb29vDhzYx9tvv0FxcREBAYE89thTdO7chdjYGF577WVyc3MwmczMnPkI/fsP4K+/VrBw4RdYLBaCg4N56qnZ+PsHMGPGHZhMJt5992MeeeR+brvtLnr10r7gR44cYtasB/nhhz/KjGCXLl3CypUr+OSTBcTFxfLGG3PJycnGbrdz1lk9eOihWQQFBQGQmJjAu+++SUJCAjodREVFc//9D9ci3PEo8+Zp4Y49e/biySefw2w289lnH9GuXTsuu+wK4uPjeO21l8nKysRkMnP33fcxdOhwsrOzmT37STIzM/Dx8WXWrKfp2rUbhYUFvPLKi8TGnsRgMHDzzXcwbtwEHA4Hb7zxKvv370FRFG699U4mTz63jD1XXHEx7777sTvccceObe5wx5EjR3PffQ+5JzG1yBaVoKAQZs16iqioaA4dOsBbb72GzWbFx8eHBx98jJ49e3Hy5AnmzZtDUVERfn5+PPTQLLp1Owur1cprr73MkSOHMZtNzJr1jHu+4vXX59K9e3cuu6zsZPLixd/wxx+/4nQ6+O9/b2XatAvLHF+27Dd2797pDndcvvx3/ve/hRgMBqKjO/LYY08RFBREVlYm//3vf/j11z/L1D906ADz57+O1Vpcpg/lqex7WCptb1cgpnydVinsmSuWkb70O97qeg1P3T6GDrVY5zStMIPnt8xjYscxTO9xMUUr38GReAjfq+ei9wmqs20NRVsV9paYQ2TJkm8YNmwk3bqdVee6LbXPZ4Loc+XUR9hbpSvGlpSIzdsPp9mLiFCfWtVZGbsGg97AlE7jcSQdxhGzC/PAi5qVqAtaFsHBIbWKdhEIPE2rFHZrUhLZ3iF0bO+L0VBzFzOKstiSspNRkcMIMPlh3bwInV8o5v7n1lhXIKiK8867QMzNCJqEVifsqqJgS0okEb9ar5i0Km4dOnRM7Twe+9ENKBnxWIZfjc5Yv4U5BAKBoClpdcLuyMxAtdlIMQTU6sWkbGsO/yRvZ3jEYIL1Xti2/4AhvAfGbkMbwVqBQCDwPK1O2K2uVALp5qBajdhXx/2Noiqc12Uitj2/oxblYhkpsjcKBIKWS6sTdlui9vZfpiWYDmG+1ZbNs+WzIXELQ8PPIcSuYNv3J8YeozG0FxNeAoGg5dLqUgrYkhIptvgSEh6E2VTxxZTSrInfgENxcG7niVj/+Rb0eizDrqi2jqBpSE5O4j//+T+6dNFuulZrMf37D+Cuu2YQEhJKcnISV155CW+99R5Dh45w1yuJlwaqPV4ST10VTz31KAkJ8SxceDrXx65dO/jii0/caQpK7CydejcuLob333+blJRkVFWlW7fuzJz5qDsuvSq2bdvK22+/USGNbQlZWZk8+OAM93ZBQT7Z2Vn89dcGYmJO8tprL1NQUIDFYuGRRx6nRw+J9PR0Xn75eTIzM9Drddx770wGDx5aberaqtLv3nffnWRlZbnj8R999En69u3HgQP7eOedNyksLOSss7rz9NPPYzKZiIuLYd68l8nLyyM0NJTZs18mICCAffv28O67b2K3OwgKCuTxx58tk7r2999/Zu/ePe5Y8ar6UFRUxOuvv4wsH8FstnDLLXcwZsw4FEXh3XffYuvWfzCbLUyffhUXX3wZABs3/s2XX35KcXERQ4eOYObMR9i8eSMff3w6y2J6+in69OnHvHnzWb9+LV988TFOp0Lv3n147LGnMJmaZ5qRVifs1uQk0sxBNfrXC+yFrE/YxKD2Z9MuN4uikzswD7kcva9n3vwSeJ527cJYsEBbU11VVT7++H2efnoWH3zwGaAlgHr11Tl89dUSfHwqPq3VdLwqsrOzOXr0KKGhoezfv5f+/QfUql56ehr33XcXjz76JGPGjENVVb7++kuefPIRt82VYbUWM2fO87z77se0bx/OY4/NZPPmTYwcOdpdJjg4xP1ZKIrCAw/cze23a+L/6qsvccMNNzNq1Bh27tzOSy/NZuHCxXzwwduMHj2W6dOvIi4uhhkz7uSnn5a5k3cB/P77L6xZs4oJEyaXSb9bGlVViY+PY+nS38q8aFVQkM9TTz3KG2+8R/fuPXjuuSf5/fdfuOyy6cya9RAPPPAII0aM4sMP3+WbbxZwzz3388ILzzB37pt0796D5ct/Zf7815g7902sVitffPEJP/74PRMmnM6DXlUfvv76SwwGI1999S15ebnceectSFIvtmz5h5iYEyxcuARFcXLvvXfQo4eEv78/r7/+Cp98soCQkFDuv/8u12c8hpEjtbdkMzLSufvuW7nvvocoKirirbfm8cUX3xASEspzzz3BsmW/lcm02ZxoVcKuKgrWpERSvM+q0b++Ln4jVqdNG63/9TE63xDMZ9cufaqg6dHpdNx6651cfPG5HD9+DF9fX9q1C2Po0OG8++58Zs16qkKdmo5XxcqVyxk48By6devOzz//UGth/+mnpQwePJQxY8a5bb7uuv8SGRmFw+Hg888/ducdL+Hcc6fRu3dfOnbsSFRUtGvf6TS2lbFs2a94eXm50/9efPFlDB8+EoCzzupBamoKAOPGTWDQIC0oIDq6Izab1f2GJlRMXVtV+t24uFgAHnpoBjk5OVxyyWVMn34127dvpW/fs+nevQcADz74GE6nA1k+gre3tzuNwY033kxeXj42m43bb7/bXb5795589532RLR3725UVeGee+7n0KHTTxNV9eHYMZnLL78SvV5PYGAQPXr0ZOvWzRw9KjNmzHjXyNrEoEGD2bhxPb6+fkyePNWdH+aFF16pkGDtgw/e5rLLptOxYycA942sJM1ASQ6f5kirEnZHRgbYbKQFBjGkmsU1ihzFrE3YxIB2fQlP+pfi9Fi8Jt2Jzui5fMitEfvRTdjlvxukbZM0DlPPyoWryjomEx07diQ2NoY+fbTMjDNmzOTGG69h+/YtZVwuJdR0vDKWLfuNO++8h27duvPZZx/ywAMPExAQWGO9Y8dkBg0aUmafwWBg6lRNgO+8817uvPPeCvX++mtFlWlsy+N0Olm48Evmzn3Dve+CCy52//3ZZx+VSTRVwuLFX9Ojh+QWdaiYuraq9Lv5+XkMHjyUBx98DIfDwX333UGnTp1JSIjH29ub5557gtjYWPr3H8CMGTPZs2cXISGhvPLKCxw7JtO5c1cefPBRzGYz5513AaA9dXz22UfuxF7Dho1g2LARZTIoVteHnj17sXbtKoYNG0FWVib79++lR4+erpTDf3HhhZdgt9vZtm0rvXv3JTs7C6PRxKxZD5KamsqoUWO4/fa73W3Hx8exe/cuZs06nWbYaDSyefMmXnzxWcLCwhg2rHb/P01Bq5o8LYmIyTAHVjtiX5/wD0WOIs7rMBbr9h/Qtz8L41nN9yIJqkNXZoECX18/Zs16mldfneNeoKE0NR0vz7FjMmlpqQwdOoLw8Ah69pRYvlxbOUevryrNrrZfp9NhNlc9WPj44/e56aZry/wsWvRVtSlmy7N162Y6duzIWWd1r2DHe+/N59Ch/dx//8Nljn333SJ++eUnnn769JJtJalrSy8VOG7cBGbOfBSLxatM+t1+/c7mmWdewM/Pj6CgIC666FI2b96E0+lk27Yt3HnnDD7//GuKi4v45psF7oRrl19+BV988T+ioqJ599233Oex2+08//zTOJ1Obrzxlio/r+r6cMMNN2Mymbjllut4++03GDZsBEajiQsuuISePXtxxx3/Zfbspxg6dDgmkxGn08mOHdt4/PFn+fjjLzh06ID7uoKWfO3yy6+oMIofOXI0y5atZtSosbz++iu1srUpaFUjdluSFhGjax+Bt6XyrlmdNtbE/02fUImIf3djK8zGe+oMEd5YC0w9R9d5VN2Q2O124uNjK7y2P2zYCLfLpTJqOl6aP/74FZvNzjXXXA5AYWEhv/zyI1dffR3+/gGVpNnNcq/q1KtXn3LpazUBffrpWTzyyONVjthLp8OF02lsK2PDhnUVEmw5HA5eeuk50tPTeOedj8uMyj/44G3++WcT77//SZk0tZWlrt20aQOhoaHuJGYl6Xf37t2D3W5zZ51UVW1/aGgoffv2c7uQJk2awg8/fMeAAefQoUMndztTppzHM8/Mcn+ejz/+EAEBgcyb9xZ6fc2SVFkf8vPzufPOewkMDALg0UcfYPToceTl5XLllddw9933AfDGG6/SoUMHcnNzGTJkmDub4rhxEzl06KD7aWfDhnW8+eZ77nPm5uZw5Mhh9yh96tRpPPfcEzXa2lS0qhG7LSmRApMvEdGVfwkANiRupsBeyHlhQ7DtW4Gx+wgM4d2rLC9oniiKwueff0yfPv0rzWw4Y8ZMtm3bXCYPeV2Og3bj+OuvFcyf/4E7ze733/9CRkY6u3btoEuXruTm5rgjShRF4ffff3YL3iWXXM7mzZvYvHkjoAngggWfkZWVSUhIaJXn7dOnH3FxsSQkxJdJY1sZBw7sY8CAc8rse//9tyksLODNN98rI+rffbeIXbt28uGHn5cRcK2diqlrq0q/m5+fxwcfvI3VaqWwsIDly/9g3LiJDB06Alk+4vbpb9q0EUnqTf/+Z5OdncWxY0dd+/92Z1Z88cVniI7uWKmPuzKq6sO6dav47DMt+un48WMcPXqEoUOHsX//PubNm4Oqqpw6lcqGDesYN24io0aNZdu2zeTl5eF0Otmy5R969dJsys7Oxmq1um9QoF27F154hpQUrW9r166q8Hk1J1rViL0oIYFUYwCdq/Cv25x2Vsf9Tc/g7kQf/gcHOizDrmxkKwX1JT09jZtuuhYARXHSo4fE7NlzKi1b4nJ56KEZtT5+003X8vrrb7sXYNi4cT3h4ZH07duvTL2LLrqMX375gUGDhvDCC3N55x0tLNFqLWbw4KHccssdgOYbf/31d/jgg7f58MN3URSFnj178corb1AdFouFZ56ZzVNPPYbNpqWxnThR8y3PnfsiY8aMY8yY8YC21mhYWHt33aysLH788TsiI6O4446b3Pu//PJ/fPHFp/j6+nLffXe695f0Nykpke7dyw5wLr10OsePH+fGG69GURTuvvs+IiIiiYiI5NChA9xyy3U4nQr/939XupcpfPTRJ5k16yFsNis9evRkxoyZWCxevPzy68yb9xJFRcW0b9+eZ555gaNHj7Bhw3q6dOnGLbdcj053+jOrDFVVq+zDxRdfzosvPsONN16N0WjkhRfm4uPjy+jRY9myZRM33HA1oE3oRkZGERkZxbXX3sg999yKw+Fg6NDhXHDBJa7PNKHMZwoQGBjEY489xaxZMwEdXbp0LRNN1NxoNWl7VUXh6L13scPnLAbcexv9ulUcEa1L2MT3R3/hvs4XEr36S8yDLsEypHmGK1WFSNvbcLzzzhvceuud+PrWnOa5oREpbNsGDZW2t9WM2O0Z6ejsNtKqSCXgUBz8FbuOboGd6bBvHfgEYR5wQeMbKmiWqKpKr159moWoCwRnSqvxsZdMnNqCwwjwreir25q8k2xrDlNNUahpJ7EMuxKdyauxzRQ0U3Q6Heeee35TmyEQeIRWJOxaqKNfx4prPzoVJ3/GrqWTXzRd9/+NPqwrxh4jG9tEgUAgaBRajbAXxSeQZ/QhqkNF3/qO1D1kFGcyRfGHgixX9sZW03WBQCAoQ6tRt4L4eNJMgRVyxCiqworY1UR5h9Hj4BaM3YZhjOjZRFYKBAJBw9MqhF1VFJS0VNLNQXSOKCvsu0/t41RhOpOLDOhQsAwX4Y0CgaB10yqE3Z6Rjt5hJ88vlGD/069wK6rCipg1hFuC6X18H+b+09D7hzWhpQKBQNDwtIpwR1uiNnFqiYwqkxpgf/ohkgpSuKbIC4N3AOaBFzaViQIPUFCQz0cfvc+ePTsxGIz4+/szY8aDSFIvkY+9VD72goJ8Xn99LjExJwB4/PFnkaReVeYyBy2Gf9u2reh0cOONtzB16jScTidvvTXPlWkRLrnkMq66SntB7M8/l/H11wsAGDFiFDNmzKw2l7ksH+G1117GbrcTHh7OM8+8iL+/P+np6bz66oukp6fh5eXNs8++SGRkVJU55TMzM5g79yWSkxPx8fHlvvsepF+/s6vMx/7ll5+yfv1at01xcbHcdttdXHvtDVXmlH/55efZt28PXl7eANx88+2MHz+RzZs38dFH7wLQrVt3Hn30SXx8fKq9jk1FqxD24sQEAAK7dnLvU1WV5TGraWf0pX/iSSzjbkFn9m4qEwVniKIoPPLIAwwaNIQvv1yE0Whk164dPPLI/XzzzXeAyMdeko/93Xffon37cJ577iW2bPmH119/hU8/XVhlLvPdu3dy6NBBFi5cTE5ONtdeewVjx07gr79WkJOTw8KFS7Bardx++40MGDCIzp27MH/+6yxe/CN+fn7cffetbN++tcpc5gBvv/06t956JyNHjubdd99i8eKvueOOe3jxxWeZOHESl112Bb/++iMffvguL7zwSpU55d97bz6S1It5894iMTGBBx+8l6+//rbKfOw333w7N998OwA7dmzj/ffnM336VVXmlAc4cuQQ7733aZn8PHl5ecyZM5t33/2Yrl278b//LeSTT95n5sxHa/W/0Ni0CmHPPhlHrtGHjp1Ou1kOZcrE5yUyPUfBFNoJY88xTWhh62Br8k42J29vkLZHRg5leOTgKo/v2rWD1NQUbr31TndWxUGDhvDkk8+iKNqbeyIf+zRUVWXdujV8//2vgDaaLsmpUlUuc0VRsNmsOBwOiouL3TlbunY9i759+6HX6/H29iYqKppTp1Lo2LEjqqpQXFyEt7cXTqejTIZNqJjLXFEUdzZNq7WYgIAAsrOz+fffo8yf/z4AF110Ceeco6U5riqn/NGjMtdff5OrDx0IDAzkwIH9VeZjv+iiSwEt78/rr8/lmWdewGKxEBsbA1TMKV9cXExqagqvvPIC6emnGDduIjfffDsJCXFERES6E86NGjWWhx++r9kKe6vwsVsTE8ssXq2qKstPriZYZ+GctHQtvLGSFKuClsPRozI9evSskCp35MgxBAeHuLdLkntt376l0nZqOl4Zy5b9xqRJU5g0aQrr1q0mNzenVvWOHZPp2VMqs68kH7vRaOTOO+9lwYJFZX6uvfZG0tPT6pyP/a67tOyFWVmZmM1mfvrpe+6882buv/8unE4noOUyDwjQFoconct82LARREVFc/nlF3D99Vdy/fU34eXlRb9+/enWTcsfs3//Xg4dOsiAAYPw8fF1uTOu4PLLLyQiIqrMza4kl/kVV1zj3jdjxoO8+uocLr30PLZv38pll00nMTGB9u0jeO+9t7jttht54olH3UvNXXDBxRgM2tKWpXPKa/nVVwLaEn4nT54gIyPDnY/d4XCQlnaK/fv3kpmZ4T7/ihV/0L17d3fen7y8XAYPHsrLL7/O229/yM8//8D27VvIyEhn0KChPPHEs3z88QL27t3N77//QocOnTh1KtWdyGzNmr/KtN/caPEjdlVR0GeeIitQon2w5mqRs45zMjeWSzOK8OoyGGNU7ya2snUwPHJwtaPqhkSvrz63eQml861/9dWSOh8vT+l87Eaj0Z2P/eqrr/NIPvbKRuzt2oXVOx+70+kkMzMDX18/Pv74S7Zv38KTTz7K99//4q5Tksv8vfe0uYdffvkRg8HAr7/+SU5ONvfffxd9+vSjX7/+AOzZs4vnnnuC5557kYCAAI4fP8Yff/zKDz/8hq+vHy+88AyLF3/NtdfeCFTMZW61FjN37ovMn/8+ffr0Y8mSb3jppee44YabOXZM5tZb7+S++x7ijz9+4aWXnnPPWaiqyvvvv82hQ/t55x3N1hkzHmL+/HnceOPV9O3bn0GDhmAyGbnhhpt55503uOWW6+jUqYs7H3sJv/zyIzNnPuLe7tfvbHfiMsCdU/7++x/mlVded++fPv1qVqz4g0suuZynn37enSny4osva7brnUIrEHZ7ejoGpwPCwtG7vgwrYlYTgIEhuUVYzru6iS0UeIJevfrw009LKyxC8fHH7zN06PAyk59tOR97YGBQmVWahg4dQVFRIVlZmQQHh1Say3zjxvVcdtkVrpzq7Rg1aiz79u2mX7/+rF+/htdfn8vzz7/sXg1q27bNDB48zP2kdMEFF/PTT9+7hX3DhrK5zE+c+BeLxUKfPtpo+dJLp/PZZx8xc+aj+Pj4MHr0WADOO28ab745D6g6p3xxcRGzZj3tniO5/vqriI7uUGU+doC0tFNkZ2eVEfKqcsr/++9x4uNjS63UpO13Op2EhbXn008XAnD48EGioiqmi24utHj/hNUVEePTQUslcDz7JMeyTzA2LQefvlPRB7SvrrqghTBgwDkEB4fwxRefuF0LW7duZtmyX+nSpWuF8m01H7vZbGbo0OFud8WBA/vx9vYmMDCoylzm3bv3ZMOGdQAUFRWxc+c2evXqw+HDB3n99bm89db7ZZb46969Jzt2bKOoqAhVVdm06W/3IhqV5TKPju7IqVOpxMXFALBhw3p69epDdHQHwsLC3U8tGzb8jSRpT9dV5ZRfuvRbfv75BwC2bPkHRXHSvXvPKvOxl3xG5XOnV5VTXlVV3nnnTXJzc3E4HPzyy0+MGzcBnU7HQw/NIC3tFKqqsmTJ/5g8eWqV17CpafEj9ox/YwBo170LoI3WfRUdI2wGLIMurrqioEWh0+mYO/dN3n33DXfO7cDAIF577W1CQkJJTk4qU76t5mMHLbzxtdfm8OOP32E0Gpk9+2V0Ol2VucxvvPEW3nhjLtdddwV6vZ6LLrqUQYOG8PjjD+F0Onnppefc5W+77U7GjBnP0aNHuPXW6zEajfTu3dc9oVlZLvOAgACefPI5nnnmCUAlKCiEJ5/U2nz55XnMm/cyH3zwNn5+fjz55Owqc8ovWLCIG264mdmzn2TFij/w8fFlzpzX0Ov1VeZjr+ozGj16bJU55a+//iZ3nvbx4ye5n34effRJHn74Pmw2O0OGDHM/oTRHWnw+9jWPv0TxMZmIF+bh9MrktR3vMS09n/P6X4W59wRP29vkiHzsDYfIx960iD5XTn3ysbd4V4wzJZlMcxCRoT6sOLkKb0VllD4YkzSuqU0TtCBEPnZBa6JFC7vqdGLJScMaFEZKYQr7M44wOquAwJHXifBGQZ0Q+dgFrYkWrX7FqacwKE6MkVEs/3cFFkVlXEAPjNF9aq4sEAgErZQWLezJh/8FwNApiL2ZRxiVU0Tw8P80sVUCgUDQtLRoYU85fByAk77HMCoqEyKHoQ+KaGKrBAKBoGlp0cKeFxtHrsmbA844RuQ7CB08valNEggEgianRcexKynJ5AfpMagqU7pNQWepfcY+QcvijTdeZf/+vTgcdhIS4unSRUvGdOWV13DhhZdUmx532bLfeP/9+Xz99XfuF4PKp9etCofDwfTpFzJhwmQefPAx9/7PP9dehrn11tMx4cuW/cbu3Tt56qnZAPzzz0a+/voLCguLUBQn48ZNLJPErCoWL/6G33//GadT4e67ZzB+/KQKZZYt+41Fi75Cr9czaNBQZsyYidFo5NgxmblzX6KgoICBA8/hkUeewGg0snfvbt5++w0cDjuRkVE89dTzBAQEcPLkCV599SWKiooICAjgqadmExER6T7P9u1b+Oabhbz99oeA9ubtK6+8QFxcDKqqcuONtzBlynnVpvitKj1u+T7MnPkQoOeLLz7hjz9+xd9fy2tz8cWXMX36VSQnJ/HSS89RUFCAn58fTz/9fBlbf//9Z/bu3eP+/FNSkrnhhquJjtbeEA0JCeHNN99DVVUWLvyc9evXUFxczH//eyvTpp1O6Z2WdorbbruBX375071v48a/+fLLTykuLmLo0BFl0hM0R2ol7JIkXQs8DZiA+bIsv1/u+EDgMyAA+Bu4S5Zlh2dNLYuqKPjkZXA8yswwq57QftMa8nSCJubhh2cBpwW5JGUt1C49bmFhIa+99kqZPCC1YcuWTfTu3Zc1a1Zx99334+XlVct6//DWW/N444136dSpM1ZrMc8++wSff/4xt99+d5X1Dh8+yMqVy/jqq8Xk5uZx5503c845gwkICHSXiYuL4dNPP+TTT7+iXbt2vP76XJYuXcI111zPCy88w6xZz9CvX39eeeUFfvvtZy6//Apefvl55s59k65du/Hhh++yePHX3Hnnvbz55qvcdNNtjBgxip9/XspHH73H7NlzUBSFb79dxNdff0m3bme5z/3NNwsID4/gxRfnkpWVyU03XcugQUPYtGlDpSl+e/aUKk2PW1kfvvtuCVdddS1Hjhzm+edfLpMCAOCzzz5kypTzuPzyK1i6dAmffPIBzz77IlarlS+++IQff/yeCRNO3wSPHDnE1Knn8dhjZTN5rly5nO3bt/LJJwvJzc3hppuuZfTocfj7+7N580befvtNMjJOJ/hKTEzg9ddf4ZNPFhASEsr9999VIY1yc6NGV4wkSdHAHGAMMBC4Q5Kk8mEn3wAzZFnuCeiA2z1sZwUyYxMxqk6yAgyc2/tydHpDQ59S0EypKj3u5ZdfgcOhjS/Gj59EQkIcK1euqFPbf/zxG+PGTaR3776sWvVnzRVcfPXVF9x44y106qS9WGKxePHww48zcOAgQEsXe9NN15b52bVrB5s3b2L8+ElYLBaCg0M455zBbNq0oUzbx48fp2/f/u78MaNHj2HDhvWkpCRjtVrdybsuuOBi1q5dBcD//reUrl274XA4SE8/5c5n89Zb7zNixCgURSElJcU9So6JOUls7MkK6Y3POWeQO2tjcHAIAQEBZGRk0LXrWdx8820VUvzGxcW6+/vf//6HH374tso+/P23tiCGLB/iq6++5L//vYY333wVq9UKgNOpUFCg5eYpKip2pwrWnhIU7rnn/jK2Hj58iJMn/+Wmm67l/vvv4t9/tTm51av/4j//uR6TyURoaDs++OAzd1u///4LL788r0w7f/+9jsmTp9K+fbjrrdZXyryN3BypzYh9CrBGluVMAEmSlgJXAC+4tjsD3rIsl+RBXQA8D3zocWtLEbtvF35AqK8PYd1G1FhecObk/rOJnI1/N0jbgWPGETCqfiOgY8fkMrlMgDKJsABMJhNPPTWbxx57kCFDhtaq3aysLHbs2MoTTzyLwWBg6dJv3fm9a2PTAw+UfVxv3z7cnaOldJKs0qxa9Se9e/d1b1eWsrd79x68995bpKam0K5dGGvXriYjI6PSdL+nTml1SxJczZx5D0ajkTvuuNe9Py8vj+uvvxKrtdi92lS3bmfx+OPPsGvXjjLnLr361OrVf2G32+natVuZ0XhJit+nn9ZcNoMHD+XBBx/D4XBw33130KlT5yr7UFhYSI8eEvfe+wDR0R14+eXnWbDgM+68815uv/1u7rrrFpYu/RaHw85HH30JaEndhg0bwbJlZd1qZrOZc8+9gEsv/T+2bv2HJ554mP/9bymJifHExJzku+8Wk5+fx3XX3eTOGz9nzmsVrkliYjxGo4lZsx4kNTWVUaPGVPvU1RyojbBHAcmltpOBYTUcr1PaM9ersXWiIOcYZqOOqVNuIizMv+YKrYjG6u+pU3qMxtMPdXqDrlw6Wc+hN+jKnKs8JccMBn2ZbQC9Xo+Xl1eV9fV6ze5+/fpx8cWX8MYbc3nggYcrtFOeVauWM2TIMEJCgpgwYSLz5s3h33+PIkm9MBoNKIpSzg7NPqNRj06nx8enapseeOAeMjMzy+ybOfPhCn3U6cBoNJRpp1u3rtxzz3088cTDWCwWJk+eypEjh9DptL6WlNXrdW57ACSpJ8uXr+Knn5Yye/aTfPrpAgCCgwP544+VbN68iccff5gff/zNnQvdYNCj01W8NqtX/8U777zB/Pnv4eVldu/fvXsnTz/9BM8/P4eQkCBCQgYycOBA9/FLLrmMrVv/YebMRyrtQ0CAH/Pnn77pXXfdDcyZ8zz33nsfc+bM5oknnmbcuAmsWbOap556lG+++db9P1lynUtsvfPO0+I7duw4Pv74fRISYnE6nZw48S/vvPM+GRkZ3HHHLfTu3ZtOnU6vwFby+YOW3G3nzm188MGneHt78+ijD/Lnn39w0UWXVHpt60p1/4Nav/R1/s7XRtj1QOlELjpAqcPxGqlPrpjhV99H8flJBIR1bbTcKc2BxswVoyhKmTwWfsNH4Td8VIOdr6qcGaXzaTidSoWyktSbQ4cOltlXOj2uoqioqorDoXDTTXdwyy3XsWLFsmrPCfD777+RkZHGZZdpE2s6nY4ff1zKo48+ia+vH4mJiWXqp6dn4Ovrj8Oh0KtXbw4cOEDHjl3cx+PiYlm48HOeeeYF3nij8hH7zp073aNsh0MhPT2dDh06lTmP1WpFkvrwxRf/A2DNmlVERUUTGhpGenq6u2xamjaCLygoYuvWze7FKqZMOZ933nkLh0Nh9eq/mDRpCjqdjqFDR2K1FpOVleNek9XpVNyfXQlLly5h0aKvefPN9+jatbv7WPkUvw6HUiE9rtOpoNcbKCgoqtCH6OgOJCQksWPHVveTkcOhYDAYSEvLIDb2JKNGjcPhUBg3TrvRpqdnEhwc7Lrmahlbly5dwtSp09ypfLW8WHpCQkJdE9IGQkPb06dPPw4fPlwhDW9JO8HBIQwePAx/f22eY+zYCRw4cIBp0y6q9BrWhdrkilEUpcJ3vlSumEqpTbhjAhBZajsCSKrD8QbBbLFwVp+zay4oaPXUJT1uiUvmq6++qLbNI0cOc+pUKj/88Ls7Ze+8efNZuXIFhYUFDBo0hH/+2UBWVhYA+fn5rF690i1g1157I19++Snx8XGANnn73ntvER5e/XsWI0aMckVrFJGVlcXOndvdbZZQXFzEAw/cTWFhAXa7nR9++JZJk84lIiISs9nMvn17AFixYhkjRozCaDTy5puvcuTIYUBb/ad//4EALFnytdu3vWvXDgIDg6pdaPvvv9fx7beL+fDDz92LewBVpvitKj1uZX2YMuVcLBYLH374DklJiaiqyo8/fse4cRMJCgrCbLawd+9uAPbt24O3t69b1Ctjz55d/P67tsDI7t07cToVOnfuwujRY1mz5i9UVSUnJ5tDhw7Qo0fPKtsZNWos27ZtJi8vD6fTyZYt/9CrV68qyzcHajNiXwXMliQpDCgApgN3lByUZTlWkqRiSZJGy7K8CbgBWN4g1goElVDX9Li9evXhyiv/w19/aROpR44c4rPPPuL1199xl1m27FcuuOBiLJbTUTCDBg2hY8dOrFy5nMsuu4IbbriZmTO1BaSdTieXXHKZO1JixIhR3HHHPTz33BM4nQpOp4OJE6e4F1auij59+nHuuRdw88034HA4uO22u9wpZ0unFb7lltu5446bcTgcTJ16Hueeq80nPPvsS8ybp4U79uzZiyuuuAaDwcDzz7/CvHlatEtYWBiPP/40AE8+OZt58+bw5Zef4efnx4svvlqtfZ9//jE2WzGzZj3k3vf440+zcOHnVab4rSo9bvk+nHfe+TgcCo8++iSzZj2I3e7g7LMHcM0116PT6ZgzZx5vvfUaNpsVHx8f5syp3tYHHniEOXNms2LFH1gsXsyePQe9Xs/VV1/HBx+8zQ03XI2iOLn55tvck9yV0bdvP6699kZ3Kt+hQ4dzwQWeccM0FLVK2+sKd3wSMAOfybI8T5KkZcCzsizvkCRpAPApWrjjLuBmWZattTh/F84wbW9bcsOASNvbEKiqyquvvsTjjz/TYOeoKyKFbdugodL21iqOXZblRcCicvsuKPX3XspOqAoELYZTp1I577wLai4oELQQWvSbpwKBJwgPj6jR9y0QtCRadK4YgUAgEFRECLugWpp46USBoE1T3++fEHZBlRiNZgoKcoW4CwRNgKqqFBTkYjSaay5cDuFjF1RJcHAYWVlp5OdnN6kder0eRWlb0RKiz22DmvpsNJoJDg6rc7tC2AVVYjAYadcusuaCDYwIa20biD57DuGKEQgEglaGEHaBQCBoZTS1K8YA2ltU9eVM6rZURJ/bBqLPbYP69LlUnUoXoqhVSoEGZAywocZSAoFAIKiMscDG8jubWtgtwFC0HO7OpjREIBAIWhAGtKy624EKebmaWtgFAoFA4GHE5KlAIBC0MoSwCwQCQStDCLtAIBC0MoSwCwQCQStDCLtAIBC0MoSwCwQCQStDCLtAIBC0Mpo6pUCtcC2m/TRgAubLsvx+ueMDgc/QFtP+G7hLlmVHY9vpSWrR50uB5wEdcBJtAfGsRjfUg9TU51LlLgTek2W5a2Pa1xDU4jpLwMdAMJACXNPar7MkSYPQ+mwG4oHrZVnObmw7PYkkSQHAP8BFsizHlDs2EA/rV7MfsUuSFA3MQUs/MBC4Q5KkPuWKfQPMkGW5J5rQ3d6oRnqYmvrs+if5ELhQluUBwD5gduNb6jlqeZ2RJCkceB3tOrdoanGddcCvwFzXdd4NPN4EpnqMWl7nt4FnXX2WgUca1UgPI0nScLTX/ntWUcTj+tXshR2YAqyRZTlTluUCYClwRclBSZI6A96yLG9x7VoAXNnoVnqWavuMNtK5V5blRNf2PqBTI9voaWrqcwmfoT2ptAZq6vMgoECW5RWu7ZeBSp9iWhC1uc4GtNErgA9Q1Ij2NQS3A/cCSeUPNJR+tQRXTBRaLpkSkoFhNRzv0Ah2NSTV9lmW5QzgJwBJkrzRRnHvNqaBDUBN1xlJku4HdgFbaB3U1OfuQIokSZ8D5wCHgfsaz7wGocbrDDwErJQkaT5QAAxvHNMaBlmWbwPQvGoVaBD9agkjdj1QOqGNDlDqcLwlUqs+SZIUCPwB7JVleWEj2dZQVNtnSZL6AdOBFxvZroakputsBCYAH8qyPAg4AbzZaNY1DDVdZ2/gc2CKLMuRwAfAV41qYePSIPrVEoQ9AS2LWQkRlH2kqel4S6TGPkmSFImW8ngfcFvjmdZg1NTnK13HdwDLgChJklp6yuea+pwCHJNleYdrezEVR7ctjZr63A8okmV5m2v7Y7SbW2ulQfSrJQj7KmCyJElhkiT5oI3aSnyOyLIcCxRLkjTatesGYHnjm+lRqu2zJEkG4DfgO1mWZ8qy3BpSdNZ0nZ+TZbmnLMsDgQuAJFmWxzaNqR6j2j6jRVGESZI0wLV9MbCzkW30NDX1+TjQUTrtt7gULTVtq6Sh9KvZC7trgvApYC2wB1gky/I2SZKWSZI0xFXsOuAtSZKOAH7AO01irIeoRZ8vQZtYu0KSpD2un8+azuIzp5bXuVVRU59lWS4CLgc+lSTpIDAJeLjJDPYAtehzFnAT8J0kSfuAW4Cbm8rehqKh9UvkYxcIBIJWRrMfsQsEAoGgbghhFwgEglaGEHaBQCBoZQhhFwgEglaGEHaBQCBoZQhhFwgEglaGEHaBQCBoZQhhFwgEglbG/wMHF36HXQzKhgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "figsize(6, 6)\n",
    "#set up plotting area\n",
    "plt.figure(0).clf()\n",
    "\n",
    "fpr, tpr, _ = metrics.roc_curve(y2,  y_pred_xgb)\n",
    "auc = roc_auc_score(y2, y_pred_xgb)\n",
    "plt.plot(fpr,tpr,label=\"XGBoost, AUC=\"+str(auc))\n",
    "\n",
    "fpr, tpr, _ = metrics.roc_curve(y2,  y_predicted_class_DNN)\n",
    "auc = roc_auc_score(y2, y_predicted_class_DNN)\n",
    "plt.plot(fpr,tpr,label=\"DNN, AUC=\"+str(auc))\n",
    "\n",
    "fpr, tpr, _ = metrics.roc_curve(y2,  y_predicted_class_CNN)\n",
    "auc = roc_auc_score(y2, y_predicted_class_CNN)\n",
    "plt.plot(fpr,tpr,label=\"CNN, AUC=\"+str(auc))\n",
    "\n",
    "fpr, tpr, _ = metrics.roc_curve(y2,  y_pred_class_TCN)\n",
    "auc = roc_auc_score(y2, y_pred_class_TCN)\n",
    "plt.plot(fpr,tpr,label=\"TCN, AUC=\"+str(auc))\n",
    "\n",
    "#add legend\n",
    "plt.title(\"ROC curve (UNSW NB15 Dataset)\")\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 266,
   "id": "0f5976e7",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x1c120b1c640>"
      ]
     },
     "execution_count": 266,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fpr, tpr, _ = metrics.roc_curve(y2,  y_pred_TabNet)\n",
    "auc = roc_auc_score(y2, y_pred_TabNet)\n",
    "plt.plot(fpr,tpr,label=\"TabNet, AUC=\"+str(auc))\n",
    "#add legend\n",
    "plt.title(\"ROC curve of proposed model (UNSW NB15 Dataset)\")\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 275,
   "id": "6cab62db",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x1c120f23df0>"
      ]
     },
     "execution_count": 275,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "figsize(6, 5)\n",
    "#set up plotting area\n",
    "plt.figure(0).clf()\n",
    "\n",
    "fpr, tpr, _ = metrics.roc_curve(y2,  y_pred_xgb)\n",
    "auc = roc_auc_score(y2, y_pred_xgb)\n",
    "plt.plot(fpr,tpr,label=\"XGBoost, AUC=\"+str(auc))\n",
    "\n",
    "fpr, tpr, _ = metrics.roc_curve(y2,  y_predicted_class_DNN)\n",
    "auc = roc_auc_score(y2, y_predicted_class_DNN)\n",
    "plt.plot(fpr,tpr,label=\"DNN, AUC=\"+str(auc))\n",
    "\n",
    "fpr, tpr, _ = metrics.roc_curve(y2,  y_predicted_class_CNN)\n",
    "auc = roc_auc_score(y2, y_predicted_class_CNN)\n",
    "plt.plot(fpr,tpr,label=\"CNN, AUC=\"+str(auc))\n",
    "\n",
    "fpr, tpr, _ = metrics.roc_curve(y2,  y_pred_class_TCN)\n",
    "auc = roc_auc_score(y2, y_pred_class_TCN)\n",
    "plt.plot(fpr,tpr,label=\"TCN, AUC=\"+str(auc))\n",
    "\n",
    "fpr, tpr, _ = metrics.roc_curve(y2,  y_pred_TabNet)\n",
    "auc = roc_auc_score(y2, y_pred_TabNet)\n",
    "plt.plot(fpr,tpr,label=\"TabNet, AUC=\"+str(auc))\n",
    "\n",
    "#add legend\n",
    "plt.title(\"ROC curve (UNSW NB15 Dataset)\")\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 273,
   "id": "ffe7ecd1",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from IPython.core.pylabtools import figsize\n",
    "figsize(10, 8)\n",
    "# Import Library\n",
    "import numpy as np \n",
    "import matplotlib.pyplot as plt \n",
    "\n",
    "# Define Data\n",
    "data3 = ['Before feature extraction','XGBoost feature extraction','Autoencoder feature extraction']\n",
    "xvals = [90.1,90.5,85.32]\n",
    "yvals = [75.1,78.7,81.03]\n",
    "zvals = [78.7,80.9,85.10]\n",
    "avals = [90.9,89.7,89.00]\n",
    "bvals = [92.1,93.2,89.30]\n",
    "\n",
    "x_axis = np.arange(len(data3))\n",
    "\n",
    "# Multi bar Chart\n",
    "plt.bar(x_axis +0.1, xvals, width=0.1, label = 'XGBoost')\n",
    "plt.bar(x_axis +0.1*2, yvals, width=0.1, label = 'DNN')\n",
    "plt.bar(x_axis +0.1*3, zvals, width=0.1, label = 'CNN')\n",
    "plt.bar(x_axis +0.1*4, avals, width=0.1, label = 'TCN')\n",
    "plt.bar(x_axis +0.1*5, bvals, width=0.1, label = 'TabNet')\n",
    "\n",
    "# Xticks\n",
    "plt.xticks(x_axis,data3)\n",
    "plt.ylabel(\"Accuracy (%)\")\n",
    "plt.xlabel(\"Experiments on classification\")\n",
    "plt.title(\"Comparitive analysis of different experiments (UNSW NB15 Dataset)\")\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a0cad762",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.8.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}