Companies Bankruptcy Prediction.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Bankruptcy Prediction\n",
    "## Overview\n",
    "The dataset is about bankruptcy prediction of Polish companies. The data was collected from Emerging Markets Information Service (EMIS, ), which is a database containing information on emerging markets around the world. The bankrupt companies were analyzed in the period 2000-2012, while the still operating companies were evaluated from 2007 to 2013. The data contains 5910 instances (financial statements), 410 represents bankrupted companies, 5500 firms that did not bankrupt in the forecasting period. The dataset is downloaded from data.world (https://data.world/uci/polish-companies-bankruptcy-data)\n",
    "## Variable Description\n",
    "* **X1**:\tnet profit / total assets\n",
    "* **X2**:\ttotal liabilities / total assets\n",
    "* **X3**:\tworking capital / total assets\n",
    "* **X4**:\tcurrent assets / short-term liabilities\n",
    "* **X5**:\t[(cash + short-term securities + receivables - short-term liabilities) / (operating expenses - depreciation)] * 365\n",
    "* **X6**:\tretained earnings / total assets\n",
    "* **X7**:\tEBIT / total assets\n",
    "* **X8**:\tbook value of equity / total liabilities\n",
    "* **X9**:\tsales / total assets\n",
    "* **X10**:  equity / total assets\n",
    "* **X11**:\t(gross profit + extraordinary items + financial expenses) / total assets\n",
    "* **X12**:\tgross profit / short-term liabilities\n",
    "* **X13**:\t(gross profit + depreciation) / sales\n",
    "* **X14**:\t(gross profit + interest) / total assets\n",
    "* **X15**:\t(total liabilities * 365) / (gross profit + depreciation)\n",
    "* **X16**:\t(gross profit + depreciation) / total liabilities\n",
    "* **X17**:\ttotal assets / total liabilities\n",
    "* **X18**:\tgross profit / total assets\n",
    "* **X19**:\tgross profit / sales\n",
    "* **X20**:\t(inventory * 365) / sales\n",
    "* **X21**:\tsales (n) / sales (n-1)\n",
    "* **X22**:\tprofit on operating activities / total assets\n",
    "* **X23**:\tnet profit / sales\n",
    "* **X24**:\tgross profit (in 3 years) / total assets\n",
    "* **X25**:\t(equity - share capital) / total assets\n",
    "* **X26**:\t(net profit + depreciation) / total liabilities\n",
    "* **X27**:\tprofit on operating activities / financial expenses\n",
    "* **X28**:\tworking capital / fixed assets\n",
    "* **X29**:\tlogarithm of total assets\n",
    "* **X30**:\t(total liabilities - cash) / sales\n",
    "* **X31**:\t(gross profit + interest) / sales\n",
    "* **X32**:  (current liabilities * 365) / cost of products sold\n",
    "* **X33**:\toperating expenses / short-term liabilities\n",
    "* **X34**:\toperating expenses / total liabilities\n",
    "* **X35**:\tprofit on sales / total assets\n",
    "* **X36**:\ttotal sales / total assets\n",
    "* **X37**:\t(current assets - inventories) / long-term liabilities\n",
    "* **X38**:\tconstant capital / total assets\n",
    "* **X39**:\tprofit on sales / sales\n",
    "* **X40**:\t(current assets - inventory - receivables) / short-term liabilities\n",
    "* **X41**:\ttotal liabilities / ((profit on operating activities + depreciation) * (12/365))\n",
    "* **X42**:\tprofit on operating activities / sales\n",
    "* **X43**:\trotation receivables + inventory turnover in days\n",
    "* **X44**:  (receivables * 365) / sales\n",
    "* **X45**:\tnet profit / inventory\n",
    "* **X46**:\t(current assets - inventory) / short-term liabilities\n",
    "* **X47**:  (inventory * 365) / cost of products sold\n",
    "* **X48**:\tEBITDA (profit on operating activities - depreciation) / total assets\n",
    "* **X49**:\tEBITDA (profit on operating activities - depreciation) / sales\n",
    "* **X50**:\tcurrent assets / total liabilities\n",
    "* **X51**:  short-term liabilities / total assets\n",
    "* **X52**:\t(short-term liabilities * 365) / cost of products sold)\n",
    "* **X53**:\tequity / fixed assets\n",
    "* **X54**:\tconstant capital / fixed assets\n",
    "* **X55**:\tworking capital\n",
    "* **X56**:\t(sales - cost of products sold) / sales\n",
    "* **X57**:\t(current assets - inventory - short-term liabilities) / (sales - gross profit - depreciation)\n",
    "* **X58**:\ttotal costs /total sales\n",
    "* **X59**:\tlong-term liabilities / equity\n",
    "* **X60**:\tsales / inventory\n",
    "* **X61**:\tsales / receivables\n",
    "* **X62**:\t(short-term liabilities * 365) / sales\n",
    "* **X63**:  sales / short-term liabilities\n",
    "* **X64**:\tsales / fixed assets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import os"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "os.chdir(r'D:\\Bank\\Polish Companies Bankruptcy')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Files:  ['1year.arff', '2year.arff', '3year.arff', '4year.arff', '5year.arff']\n"
     ]
    }
   ],
   "source": [
    "files = []\n",
    "for i in range(1, 6):\n",
    "    counter = str(i)\n",
    "    name = counter + 'year.arff'\n",
    "    files.append(name)\n",
    "print(\"Files: \", files)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[(array([( 0.20055 , 0.37951,  0.39641, 2.0472 ,   32.351 , 0.38825 ,  0.24976 , 1.3305  , 1.1389 , 0.50494 , 0.24976 ,  0.6598  ,  0.1666  ,  0.24976 ,    497.42,  0.73378 , 2.6349,  0.24976 ,  0.14942 , 43.37  , 1.2479 , 0.21402 ,  0.11998 , 0.47706 , 0.50494 ,  0.60411 ,  1.4582 ,  1.7615 , 5.9443, 0.11788, 0.14942 ,  94.14 , 3.8772, 0.56393 ,  0.21402 , 1.741  , 5.9327e+02, 0.50591,  0.12804 , 0.66295 , 0.051402, 0.12804 , 114.42 ,  71.05 ,  1.0097  , 1.5225 ,  49.394 ,  0.1853  ,  0.11085 , 2.042  , 0.37854, 0.25792,  2.2437 ,  2.248  ,  3.4869e+05,  0.12196 ,  0.39718 , 0.87804, 1.9240e-03,  8.416 , 5.1372,  82.658, 4.4158,  7.4277 , b'0'),\n",
      "       ( 0.20912 , 0.49988,  0.47225, 1.9447 ,   14.786 , 0.      ,  0.25834 , 0.99601 , 1.6996 , 0.49788 , 0.26114 ,  0.5168  ,  0.15835 ,  0.25834 ,    677.96,  0.53838 , 2.0005,  0.25834 ,  0.152   , 87.981 , 1.4293 , 0.24806 ,  0.12304 ,      nan, 0.39542 ,  0.43992 , 88.444  , 16.946  , 3.6884, 0.26969, 0.152   , 122.17 , 2.9876, 2.9876  ,  0.20616 , 1.6996 ,        nan, 0.49788,  0.1213  , 0.086422, 0.064371, 0.14595 , 199.49 , 111.51 ,  0.51045 , 1.1252 , 100.13  ,  0.23727 ,  0.13961 , 1.9447 , 0.49988, 0.33472, 17.866  , 17.866  ,  2.3046e+03,  0.1213  ,  0.42002 , 0.853  , 0.0000e+00,  4.1486, 3.2732, 107.35 , 3.4   , 60.987  , b'0'),\n",
      "       ( 0.24866 , 0.69592,  0.26713, 1.5548 ,   -1.1523, 0.      ,  0.30906 , 0.43695 , 1.309  , 0.30408 , 0.31258 ,  0.64184 ,  0.24435 ,  0.30906 ,    794.16,  0.45961 , 1.4369,  0.30906 ,  0.2361  , 73.133 , 1.4283 , 0.3026  ,  0.18996 ,      nan, 0.28932 ,  0.37282 , 86.011  ,  1.0627 , 4.3749, 0.41929, 0.23815 , 176.93 , 2.063 , 1.4274  ,  0.31565 , 1.309  , 2.3019e+00, 0.51537,  0.24114 , 0.32202 , 0.07402 , 0.23117 , 165.51 ,  92.381,  0.94807 , 1.0101 ,  96.372 ,  0.29181 ,  0.22293 , 1.0758 , 0.48152, 0.48474,  1.2098 ,  2.0504 ,  6.3327e+03,  0.24114 ,  0.81774 , 0.76599, 6.9484e-01,  4.9909, 3.951 , 134.27 , 2.7185,  5.2078 , b'0'),\n",
      "       ...,\n",
      "       ( 0.006338, 0.50276,  0.43923, 1.8736 ,    9.7417, 0.006338,  0.012022, 0.98356 , 1.0083 , 0.49449 , 0.012022,  0.023912,  0.014722,  0.012022,   3465.6 ,  0.10532 , 1.989 ,  0.012022,  0.003342, 34.862 ,     nan, 0.014559,  0.001762, 0.012022, 0.49449 ,  0.094015,      nan,  7.5711 , 4.041 , 0.13806, 0.003342,  51.443, 7.0952, 0.028958,  0.014559, 3.6054 ,        nan, 0.49449,  0.004048, 0.012816, 0.29788 , 0.004048,  94.936,  60.074,  0.018448, 1.1903 ,  35.153 , -0.02637 , -0.007331, 1.8736 , 0.50276, 0.14094,  8.5238 ,  8.5238 ,  4.8269e+03,  0.008258,  0.012817, 0.99174, 0.0000e+00, 10.47  , 6.0759,  51.019, 7.1542, 62.001  , b'1'),\n",
      "       (-0.041643, 0.8481 , -0.12852, 0.57485, -121.92  , 0.      , -0.036795, 0.17901 , 0.42138, 0.15182 , 0.006414, -0.12172 , -0.030657, -0.036795, -23963.  , -0.015232, 1.1791, -0.036795, -0.087322,  9.1932, 0.64405, 0.006414, -0.098825, 0.045607, 0.12978 , -0.020947,  0.14843, -0.15557, 4.104 , 2.0119 , 0.01522 , 212.42 , 1.7183, 0.61247 , -0.098064, 0.52581, 2.9911e-01, 0.69731, -0.23272 , 0.096108, 0.9333  , 0.01522 , 125.36 , 116.17 , -3.9237  , 0.13122,   7.4576, -0.017463, -0.041444, 0.2049 , 0.3023 , 0.58197,  0.18377,  0.84405, -1.6330e+03, -0.23272 , -0.27429 , 0.98788, 3.5931e+00, 39.703 , 3.142 , 261.85 , 1.3939,  0.51005, b'1'),\n",
      "       ( 0.014946, 0.94648,  0.03211, 1.0363 ,  -20.581 , 0.      ,  0.01526 , 0.056357, 2.9694 , 0.053341, 0.076621,  0.017266,  0.01542 ,  0.01526 ,   7545.  ,  0.048376, 1.0565,  0.01526 ,  0.005139, 24.275 ,     nan, 0.049167,  0.005034, 0.01526 , 0.015115,  0.048045,  0.80127,  0.37975, 3.1166, 0.31637, 0.014846, 110.37 , 3.3074, 3.0884  ,  0.046552, 2.9998 , 1.1422e+01, 0.11624,  0.015677, 0.007959, 0.39588 , 0.016558, 111.72 ,  87.443,  0.075684, 0.23141,  24.66  ,  0.018639,  0.006277, 0.96769, 0.88379, 0.30235,  0.63083,  1.3747 ,  4.2000e+01,  0.015705,  0.28021 , 0.97443, 1.1792e+00, 15.036 , 4.1741, 108.64 , 3.3599, 35.118  , b'1')],\n",
      "      dtype=[('Attr1', '<f8'), ('Attr2', '<f8'), ('Attr3', '<f8'), ('Attr4', '<f8'), ('Attr5', '<f8'), ('Attr6', '<f8'), ('Attr7', '<f8'), ('Attr8', '<f8'), ('Attr9', '<f8'), ('Attr10', '<f8'), ('Attr11', '<f8'), ('Attr12', '<f8'), ('Attr13', '<f8'), ('Attr14', '<f8'), ('Attr15', '<f8'), ('Attr16', '<f8'), ('Attr17', '<f8'), ('Attr18', '<f8'), ('Attr19', '<f8'), ('Attr20', '<f8'), ('Attr21', '<f8'), ('Attr22', '<f8'), ('Attr23', '<f8'), ('Attr24', '<f8'), ('Attr25', '<f8'), ('Attr26', '<f8'), ('Attr27', '<f8'), ('Attr28', '<f8'), ('Attr29', '<f8'), ('Attr30', '<f8'), ('Attr31', '<f8'), ('Attr32', '<f8'), ('Attr33', '<f8'), ('Attr34', '<f8'), ('Attr35', '<f8'), ('Attr36', '<f8'), ('Attr37', '<f8'), ('Attr38', '<f8'), ('Attr39', '<f8'), ('Attr40', '<f8'), ('Attr41', '<f8'), ('Attr42', '<f8'), ('Attr43', '<f8'), ('Attr44', '<f8'), ('Attr45', '<f8'), ('Attr46', '<f8'), ('Attr47', '<f8'), ('Attr48', '<f8'), ('Attr49', '<f8'), ('Attr50', '<f8'), ('Attr51', '<f8'), ('Attr52', '<f8'), ('Attr53', '<f8'), ('Attr54', '<f8'), ('Attr55', '<f8'), ('Attr56', '<f8'), ('Attr57', '<f8'), ('Attr58', '<f8'), ('Attr59', '<f8'), ('Attr60', '<f8'), ('Attr61', '<f8'), ('Attr62', '<f8'), ('Attr63', '<f8'), ('Attr64', '<f8'), ('class', 'S1')]), Dataset: '1year-weka.filters.unsupervised.instance.SubsetByExpression-Enot\n",
      "\tAttr1's type is numeric\n",
      "\tAttr2's type is numeric\n",
      "\tAttr3's type is numeric\n",
      "\tAttr4's type is numeric\n",
      "\tAttr5's type is numeric\n",
      "\tAttr6's type is numeric\n",
      "\tAttr7's type is numeric\n",
      "\tAttr8's type is numeric\n",
      "\tAttr9's type is numeric\n",
      "\tAttr10's type is numeric\n",
      "\tAttr11's type is numeric\n",
      "\tAttr12's type is numeric\n",
      "\tAttr13's type is numeric\n",
      "\tAttr14's type is numeric\n",
      "\tAttr15's type is numeric\n",
      "\tAttr16's type is numeric\n",
      "\tAttr17's type is numeric\n",
      "\tAttr18's type is numeric\n",
      "\tAttr19's type is numeric\n",
      "\tAttr20's type is numeric\n",
      "\tAttr21's type is numeric\n",
      "\tAttr22's type is numeric\n",
      "\tAttr23's type is numeric\n",
      "\tAttr24's type is numeric\n",
      "\tAttr25's type is numeric\n",
      "\tAttr26's type is numeric\n",
      "\tAttr27's type is numeric\n",
      "\tAttr28's type is numeric\n",
      "\tAttr29's type is numeric\n",
      "\tAttr30's type is numeric\n",
      "\tAttr31's type is numeric\n",
      "\tAttr32's type is numeric\n",
      "\tAttr33's type is numeric\n",
      "\tAttr34's type is numeric\n",
      "\tAttr35's type is numeric\n",
      "\tAttr36's type is numeric\n",
      "\tAttr37's type is numeric\n",
      "\tAttr38's type is numeric\n",
      "\tAttr39's type is numeric\n",
      "\tAttr40's type is numeric\n",
      "\tAttr41's type is numeric\n",
      "\tAttr42's type is numeric\n",
      "\tAttr43's type is numeric\n",
      "\tAttr44's type is numeric\n",
      "\tAttr45's type is numeric\n",
      "\tAttr46's type is numeric\n",
      "\tAttr47's type is numeric\n",
      "\tAttr48's type is numeric\n",
      "\tAttr49's type is numeric\n",
      "\tAttr50's type is numeric\n",
      "\tAttr51's type is numeric\n",
      "\tAttr52's type is numeric\n",
      "\tAttr53's type is numeric\n",
      "\tAttr54's type is numeric\n",
      "\tAttr55's type is numeric\n",
      "\tAttr56's type is numeric\n",
      "\tAttr57's type is numeric\n",
      "\tAttr58's type is numeric\n",
      "\tAttr59's type is numeric\n",
      "\tAttr60's type is numeric\n",
      "\tAttr61's type is numeric\n",
      "\tAttr62's type is numeric\n",
      "\tAttr63's type is numeric\n",
      "\tAttr64's type is numeric\n",
      "\tclass's type is nominal, range is ('0', '1')\n",
      "), (array([( 0.20235  , 0.465  ,  0.24038, 1.5171 ,  -14.547  ,  5.1069e-01, 0.25366  , 0.91816 , 1.1519 , 0.42695 , 0.25366  , 0.54561  , 0.17865 , 0.25366  ,   603.2 , 0.60511 , 2.1505, 0.25366  , 0.16105   ,  42.697, 1.0806 ,  0.22349 ,  0.12848   , 0.63105  , 0.42695 , 0.49478 ,  1.6345 ,  0.81559, 6.0038, 0.22803, 0.16105   , 124.1  , 2.9412 ,  0.48063 ,  0.22349 , 1.6258 , 5095.3  , 0.42705,  0.1419   , 0.46489  , 0.060856,  0.1419   , 113.36 ,  70.661,  1.0983   , 1.1207 ,  49.181,  0.19577 ,  0.1243  , 1.5167 , 0.4649 , 0.34   ,  1.4486 ,  1.449  ,  2.4250e+05,  0.13184  ,  0.47395  , 0.86816, 0.00023951,  8.5487,  5.1655 , 107.74 , 3.3879 ,  5.344 , b'0'),\n",
      "       ( 0.030073 , 0.59563,  0.18668, 1.3382 ,  -37.859  , -3.1864e-04, 0.04167  , 0.6789  , 0.32356, 0.40437 , 0.042199 , 0.075493 , 0.14563 , 0.04167  ,  4613.9 , 0.079109, 1.6789, 0.04167  , 0.12878   , 239.12 ,     nan,  0.042196,  0.092943  ,       nan, 0.029802, 0.059639, 79.752  ,  0.71432, 4.0672, 1.8052 , 0.12878   , 708.74 , 0.51499,  0.47725 ,  0.039298, 0.32356,      nan, 0.40437,  0.12146  , 0.028211 , 0.41671 ,  0.13041  , 815.69 , 576.57 ,  0.14187  , 0.95419, 272.17 ,  0.036746,  0.11357 , 1.2401 , 0.55197, 1.9418 ,  1.5473 ,  1.5473 ,  2.1795e+03,  0.12146  ,  0.074369 , 0.87235, 0.        ,  1.5264,  0.63305, 622.66 , 0.58619,  1.2381, b'0'),\n",
      "       ( 0.25786  , 0.29949,  0.66519, 3.2211 ,   71.799  ,  0.0000e+00, 0.31877  , 2.332   , 1.6762 , 0.69841 , 0.32453  , 1.0644   , 0.19745 , 0.31877  ,   330.29, 1.1051  , 3.339 , 0.31877  , 0.19017   ,  84.246, 1.1146 ,  0.30466 ,  0.15383   ,       nan, 0.60775 , 0.90169 , 52.886  , 18.836  , 3.7415, 0.14756, 0.19021   ,  78.101, 4.6734 ,  4.6734  ,  0.27656 , 1.6762 ,      nan, 0.69841,  0.16499  , 0.17942  , 0.031507,  0.18175  , 198.36 , 114.11 ,  0.66649  , 1.9292 , 100.89 ,  0.29247 ,  0.17448 , 3.2211 , 0.29949, 0.21398, 19.777  , 19.777  ,  3.6686e+03,  0.16499  ,  0.36921  , 0.81614, 0.        ,  4.3325,  3.1985 ,  65.215, 5.5969 , 47.466 , b'0'),\n",
      "       ...,\n",
      "       ( 0.015092 , 0.55759, -0.2846 , 0.48599,  -85.571  ,  1.5092e-02, 0.0098258, 0.69488 , 1.006  , 0.38746 , 0.0098258, 0.017746 , 0.04257 , 0.0098258,  2702.7 , 0.13505 , 1.7934, 0.0098258, 0.0055547 ,  24.248, 1.265  ,  0.018633,  0.0085317 , 0.0098258, 0.38746 , 0.14449 ,      nan, -0.38937, 4.9993, 0.31327, 0.0055547 , 114.93 , 3.1757 ,  0.033417,  0.018633, 1.7872 ,   38.766, 0.39137,  0.010534 , 0.0067107, 0.21795 ,  0.010534 ,  54.757,  30.508,  0.12842  , 0.27375,  24.394, -0.046844, -0.026482, 0.48258, 0.55368, 0.31489,  0.53011,  0.53546, -2.8417e+04,  0.0059714,  0.03895  , 0.99403, 0.010091  , 15.053 , 11.964  , 114.25 , 3.1948 ,  2.4201, b'1'),\n",
      "       (-0.0025542, 0.47076,  0.42401, 1.9007 ,    0.95483, -2.5542e-03, 0.0017845, 1.1144  , 0.99293, 0.52464 , 0.0017845, 0.0037906, 0.017902, 0.0017845,  3591.3 , 0.10163 , 2.1242, 0.0017845, 0.00066768,  56.775, 0.64299, -0.020174, -0.00095568, 0.0017845, 0.52464 , 0.092417,      nan,  4.0295 , 3.9782, 0.17274, 0.00066768,  63.837, 5.7177 , -0.042854, -0.020174, 2.7197 ,      nan, 0.52464, -0.0075483, 0.022033 , 0.59787 , -0.0075483, 120.78 ,  64.007, -0.006144 , 1.0176 ,  56.374, -0.066234, -0.024782, 1.9007 , 0.47076, 0.1749 ,  4.9857 ,  4.9857 ,  4.0323e+03, -0.0071218, -0.0048685, 1.0071 , 0.        ,  6.4289,  5.7025 ,  64.291, 5.6773 , 25.399 , b'1'),\n",
      "       ( 0.0020717, 0.94315, -0.13474, 0.85607, -119.92   ,  1.5226e-02, 0.0020717, 0.059818, 1.7749 , 0.056417, 0.067687 , 0.002213 , 0.011988, 0.0020717, 16179.  , 0.02256 , 1.0603, 0.0020717, 0.0011672 ,  87.703, 0.58677,  0.055467,  0.0011672 , 0.017617 , 0.017477, 0.02256 ,  0.84534, -0.67987, 3.1086, 0.52435, 0.02907   , 198.7  , 1.838  ,  1.8244  ,  0.055389, 1.7882 ,   50.042, 0.06391,  0.031207 , 0.019268 , 0.42101 ,  0.031251 , 161.09 ,  73.391,  0.0048576, 0.47484,  90.469,  0.036262,  0.02043 , 0.84971, 0.93614, 0.54406,  0.28468,  0.32249, -1.7300e+02,  0.031154 ,  0.03672  , 0.9622 , 0.1328    ,  4.1618,  4.9734 , 192.51 , 1.896  ,  8.9562, b'1')],\n",
      "      dtype=[('Attr1', '<f8'), ('Attr2', '<f8'), ('Attr3', '<f8'), ('Attr4', '<f8'), ('Attr5', '<f8'), ('Attr6', '<f8'), ('Attr7', '<f8'), ('Attr8', '<f8'), ('Attr9', '<f8'), ('Attr10', '<f8'), ('Attr11', '<f8'), ('Attr12', '<f8'), ('Attr13', '<f8'), ('Attr14', '<f8'), ('Attr15', '<f8'), ('Attr16', '<f8'), ('Attr17', '<f8'), ('Attr18', '<f8'), ('Attr19', '<f8'), ('Attr20', '<f8'), ('Attr21', '<f8'), ('Attr22', '<f8'), ('Attr23', '<f8'), ('Attr24', '<f8'), ('Attr25', '<f8'), ('Attr26', '<f8'), ('Attr27', '<f8'), ('Attr28', '<f8'), ('Attr29', '<f8'), ('Attr30', '<f8'), ('Attr31', '<f8'), ('Attr32', '<f8'), ('Attr33', '<f8'), ('Attr34', '<f8'), ('Attr35', '<f8'), ('Attr36', '<f8'), ('Attr37', '<f8'), ('Attr38', '<f8'), ('Attr39', '<f8'), ('Attr40', '<f8'), ('Attr41', '<f8'), ('Attr42', '<f8'), ('Attr43', '<f8'), ('Attr44', '<f8'), ('Attr45', '<f8'), ('Attr46', '<f8'), ('Attr47', '<f8'), ('Attr48', '<f8'), ('Attr49', '<f8'), ('Attr50', '<f8'), ('Attr51', '<f8'), ('Attr52', '<f8'), ('Attr53', '<f8'), ('Attr54', '<f8'), ('Attr55', '<f8'), ('Attr56', '<f8'), ('Attr57', '<f8'), ('Attr58', '<f8'), ('Attr59', '<f8'), ('Attr60', '<f8'), ('Attr61', '<f8'), ('Attr62', '<f8'), ('Attr63', '<f8'), ('Attr64', '<f8'), ('class', 'S1')]), Dataset: '1year-weka.filters.unsupervised.instance.SubsetByExpression-Enot\n",
      "\tAttr1's type is numeric\n",
      "\tAttr2's type is numeric\n",
      "\tAttr3's type is numeric\n",
      "\tAttr4's type is numeric\n",
      "\tAttr5's type is numeric\n",
      "\tAttr6's type is numeric\n",
      "\tAttr7's type is numeric\n",
      "\tAttr8's type is numeric\n",
      "\tAttr9's type is numeric\n",
      "\tAttr10's type is numeric\n",
      "\tAttr11's type is numeric\n",
      "\tAttr12's type is numeric\n",
      "\tAttr13's type is numeric\n",
      "\tAttr14's type is numeric\n",
      "\tAttr15's type is numeric\n",
      "\tAttr16's type is numeric\n",
      "\tAttr17's type is numeric\n",
      "\tAttr18's type is numeric\n",
      "\tAttr19's type is numeric\n",
      "\tAttr20's type is numeric\n",
      "\tAttr21's type is numeric\n",
      "\tAttr22's type is numeric\n",
      "\tAttr23's type is numeric\n",
      "\tAttr24's type is numeric\n",
      "\tAttr25's type is numeric\n",
      "\tAttr26's type is numeric\n",
      "\tAttr27's type is numeric\n",
      "\tAttr28's type is numeric\n",
      "\tAttr29's type is numeric\n",
      "\tAttr30's type is numeric\n",
      "\tAttr31's type is numeric\n",
      "\tAttr32's type is numeric\n",
      "\tAttr33's type is numeric\n",
      "\tAttr34's type is numeric\n",
      "\tAttr35's type is numeric\n",
      "\tAttr36's type is numeric\n",
      "\tAttr37's type is numeric\n",
      "\tAttr38's type is numeric\n",
      "\tAttr39's type is numeric\n",
      "\tAttr40's type is numeric\n",
      "\tAttr41's type is numeric\n",
      "\tAttr42's type is numeric\n",
      "\tAttr43's type is numeric\n",
      "\tAttr44's type is numeric\n",
      "\tAttr45's type is numeric\n",
      "\tAttr46's type is numeric\n",
      "\tAttr47's type is numeric\n",
      "\tAttr48's type is numeric\n",
      "\tAttr49's type is numeric\n",
      "\tAttr50's type is numeric\n",
      "\tAttr51's type is numeric\n",
      "\tAttr52's type is numeric\n",
      "\tAttr53's type is numeric\n",
      "\tAttr54's type is numeric\n",
      "\tAttr55's type is numeric\n",
      "\tAttr56's type is numeric\n",
      "\tAttr57's type is numeric\n",
      "\tAttr58's type is numeric\n",
      "\tAttr59's type is numeric\n",
      "\tAttr60's type is numeric\n",
      "\tAttr61's type is numeric\n",
      "\tAttr62's type is numeric\n",
      "\tAttr63's type is numeric\n",
      "\tAttr64's type is numeric\n",
      "\tclass's type is nominal, range is ('0', '1')\n",
      ")]\n"
     ]
    }
   ],
   "source": [
    "from scipy.io import arff\n",
    "file_data = []\n",
    "for file in files:\n",
    "    file_data.append(arff.loadarff(file))\n",
    "print(file_data[0:2])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "for i in range(0, len(file_data)):\n",
    "    if i==0:\n",
    "        cb_df = pd.DataFrame.from_records(data = file_data[i][0])\n",
    "        continue\n",
    "    if i!=0:\n",
    "        d = pd.DataFrame.from_records(data = file_data[i][0])\n",
    "        cb_df = cb_df.append(d)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(43405, 65)"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cb_df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Attr1</th>\n",
       "      <th>Attr2</th>\n",
       "      <th>Attr3</th>\n",
       "      <th>Attr4</th>\n",
       "      <th>Attr5</th>\n",
       "      <th>Attr6</th>\n",
       "      <th>Attr7</th>\n",
       "      <th>Attr8</th>\n",
       "      <th>Attr9</th>\n",
       "      <th>Attr10</th>\n",
       "      <th>...</th>\n",
       "      <th>Attr56</th>\n",
       "      <th>Attr57</th>\n",
       "      <th>Attr58</th>\n",
       "      <th>Attr59</th>\n",
       "      <th>Attr60</th>\n",
       "      <th>Attr61</th>\n",
       "      <th>Attr62</th>\n",
       "      <th>Attr63</th>\n",
       "      <th>Attr64</th>\n",
       "      <th>class</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.200550</td>\n",
       "      <td>0.37951</td>\n",
       "      <td>0.39641</td>\n",
       "      <td>2.0472</td>\n",
       "      <td>32.3510</td>\n",
       "      <td>0.38825</td>\n",
       "      <td>0.249760</td>\n",
       "      <td>1.33050</td>\n",
       "      <td>1.1389</td>\n",
       "      <td>0.50494</td>\n",
       "      <td>...</td>\n",
       "      <td>0.121960</td>\n",
       "      <td>0.39718</td>\n",
       "      <td>0.87804</td>\n",
       "      <td>0.001924</td>\n",
       "      <td>8.4160</td>\n",
       "      <td>5.1372</td>\n",
       "      <td>82.658</td>\n",
       "      <td>4.4158</td>\n",
       "      <td>7.4277</td>\n",
       "      <td>b'0'</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.209120</td>\n",
       "      <td>0.49988</td>\n",
       "      <td>0.47225</td>\n",
       "      <td>1.9447</td>\n",
       "      <td>14.7860</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.258340</td>\n",
       "      <td>0.99601</td>\n",
       "      <td>1.6996</td>\n",
       "      <td>0.49788</td>\n",
       "      <td>...</td>\n",
       "      <td>0.121300</td>\n",
       "      <td>0.42002</td>\n",
       "      <td>0.85300</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>4.1486</td>\n",
       "      <td>3.2732</td>\n",
       "      <td>107.350</td>\n",
       "      <td>3.4000</td>\n",
       "      <td>60.9870</td>\n",
       "      <td>b'0'</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.248660</td>\n",
       "      <td>0.69592</td>\n",
       "      <td>0.26713</td>\n",
       "      <td>1.5548</td>\n",
       "      <td>-1.1523</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.309060</td>\n",
       "      <td>0.43695</td>\n",
       "      <td>1.3090</td>\n",
       "      <td>0.30408</td>\n",
       "      <td>...</td>\n",
       "      <td>0.241140</td>\n",
       "      <td>0.81774</td>\n",
       "      <td>0.76599</td>\n",
       "      <td>0.694840</td>\n",
       "      <td>4.9909</td>\n",
       "      <td>3.9510</td>\n",
       "      <td>134.270</td>\n",
       "      <td>2.7185</td>\n",
       "      <td>5.2078</td>\n",
       "      <td>b'0'</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.081483</td>\n",
       "      <td>0.30734</td>\n",
       "      <td>0.45879</td>\n",
       "      <td>2.4928</td>\n",
       "      <td>51.9520</td>\n",
       "      <td>0.14988</td>\n",
       "      <td>0.092704</td>\n",
       "      <td>1.86610</td>\n",
       "      <td>1.0571</td>\n",
       "      <td>0.57353</td>\n",
       "      <td>...</td>\n",
       "      <td>0.054015</td>\n",
       "      <td>0.14207</td>\n",
       "      <td>0.94598</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>4.5746</td>\n",
       "      <td>3.6147</td>\n",
       "      <td>86.435</td>\n",
       "      <td>4.2228</td>\n",
       "      <td>5.5497</td>\n",
       "      <td>b'0'</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.187320</td>\n",
       "      <td>0.61323</td>\n",
       "      <td>0.22960</td>\n",
       "      <td>1.4063</td>\n",
       "      <td>-7.3128</td>\n",
       "      <td>0.18732</td>\n",
       "      <td>0.187320</td>\n",
       "      <td>0.63070</td>\n",
       "      <td>1.1559</td>\n",
       "      <td>0.38677</td>\n",
       "      <td>...</td>\n",
       "      <td>0.134850</td>\n",
       "      <td>0.48431</td>\n",
       "      <td>0.86515</td>\n",
       "      <td>0.124440</td>\n",
       "      <td>6.3985</td>\n",
       "      <td>4.3158</td>\n",
       "      <td>127.210</td>\n",
       "      <td>2.8692</td>\n",
       "      <td>7.8980</td>\n",
       "      <td>b'0'</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 65 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      Attr1    Attr2    Attr3   Attr4    Attr5    Attr6     Attr7    Attr8  \\\n",
       "0  0.200550  0.37951  0.39641  2.0472  32.3510  0.38825  0.249760  1.33050   \n",
       "1  0.209120  0.49988  0.47225  1.9447  14.7860  0.00000  0.258340  0.99601   \n",
       "2  0.248660  0.69592  0.26713  1.5548  -1.1523  0.00000  0.309060  0.43695   \n",
       "3  0.081483  0.30734  0.45879  2.4928  51.9520  0.14988  0.092704  1.86610   \n",
       "4  0.187320  0.61323  0.22960  1.4063  -7.3128  0.18732  0.187320  0.63070   \n",
       "\n",
       "    Attr9   Attr10  ...      Attr56   Attr57   Attr58    Attr59  Attr60  \\\n",
       "0  1.1389  0.50494  ...    0.121960  0.39718  0.87804  0.001924  8.4160   \n",
       "1  1.6996  0.49788  ...    0.121300  0.42002  0.85300  0.000000  4.1486   \n",
       "2  1.3090  0.30408  ...    0.241140  0.81774  0.76599  0.694840  4.9909   \n",
       "3  1.0571  0.57353  ...    0.054015  0.14207  0.94598  0.000000  4.5746   \n",
       "4  1.1559  0.38677  ...    0.134850  0.48431  0.86515  0.124440  6.3985   \n",
       "\n",
       "   Attr61   Attr62  Attr63   Attr64  class  \n",
       "0  5.1372   82.658  4.4158   7.4277   b'0'  \n",
       "1  3.2732  107.350  3.4000  60.9870   b'0'  \n",
       "2  3.9510  134.270  2.7185   5.2078   b'0'  \n",
       "3  3.6147   86.435  4.2228   5.5497   b'0'  \n",
       "4  4.3158  127.210  2.8692   7.8980   b'0'  \n",
       "\n",
       "[5 rows x 65 columns]"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cb_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>attr1</th>\n",
       "      <th>attr2</th>\n",
       "      <th>attr3</th>\n",
       "      <th>attr4</th>\n",
       "      <th>attr5</th>\n",
       "      <th>attr6</th>\n",
       "      <th>attr7</th>\n",
       "      <th>attr8</th>\n",
       "      <th>attr9</th>\n",
       "      <th>attr10</th>\n",
       "      <th>...</th>\n",
       "      <th>attr56</th>\n",
       "      <th>attr57</th>\n",
       "      <th>attr58</th>\n",
       "      <th>attr59</th>\n",
       "      <th>attr60</th>\n",
       "      <th>attr61</th>\n",
       "      <th>attr62</th>\n",
       "      <th>attr63</th>\n",
       "      <th>attr64</th>\n",
       "      <th>class</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.200550</td>\n",
       "      <td>0.37951</td>\n",
       "      <td>0.39641</td>\n",
       "      <td>2.0472</td>\n",
       "      <td>32.3510</td>\n",
       "      <td>0.38825</td>\n",
       "      <td>0.249760</td>\n",
       "      <td>1.33050</td>\n",
       "      <td>1.1389</td>\n",
       "      <td>0.50494</td>\n",
       "      <td>...</td>\n",
       "      <td>0.121960</td>\n",
       "      <td>0.39718</td>\n",
       "      <td>0.87804</td>\n",
       "      <td>0.001924</td>\n",
       "      <td>8.4160</td>\n",
       "      <td>5.1372</td>\n",
       "      <td>82.658</td>\n",
       "      <td>4.4158</td>\n",
       "      <td>7.4277</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.209120</td>\n",
       "      <td>0.49988</td>\n",
       "      <td>0.47225</td>\n",
       "      <td>1.9447</td>\n",
       "      <td>14.7860</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.258340</td>\n",
       "      <td>0.99601</td>\n",
       "      <td>1.6996</td>\n",
       "      <td>0.49788</td>\n",
       "      <td>...</td>\n",
       "      <td>0.121300</td>\n",
       "      <td>0.42002</td>\n",
       "      <td>0.85300</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>4.1486</td>\n",
       "      <td>3.2732</td>\n",
       "      <td>107.350</td>\n",
       "      <td>3.4000</td>\n",
       "      <td>60.9870</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.248660</td>\n",
       "      <td>0.69592</td>\n",
       "      <td>0.26713</td>\n",
       "      <td>1.5548</td>\n",
       "      <td>-1.1523</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.309060</td>\n",
       "      <td>0.43695</td>\n",
       "      <td>1.3090</td>\n",
       "      <td>0.30408</td>\n",
       "      <td>...</td>\n",
       "      <td>0.241140</td>\n",
       "      <td>0.81774</td>\n",
       "      <td>0.76599</td>\n",
       "      <td>0.694840</td>\n",
       "      <td>4.9909</td>\n",
       "      <td>3.9510</td>\n",
       "      <td>134.270</td>\n",
       "      <td>2.7185</td>\n",
       "      <td>5.2078</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.081483</td>\n",
       "      <td>0.30734</td>\n",
       "      <td>0.45879</td>\n",
       "      <td>2.4928</td>\n",
       "      <td>51.9520</td>\n",
       "      <td>0.14988</td>\n",
       "      <td>0.092704</td>\n",
       "      <td>1.86610</td>\n",
       "      <td>1.0571</td>\n",
       "      <td>0.57353</td>\n",
       "      <td>...</td>\n",
       "      <td>0.054015</td>\n",
       "      <td>0.14207</td>\n",
       "      <td>0.94598</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>4.5746</td>\n",
       "      <td>3.6147</td>\n",
       "      <td>86.435</td>\n",
       "      <td>4.2228</td>\n",
       "      <td>5.5497</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.187320</td>\n",
       "      <td>0.61323</td>\n",
       "      <td>0.22960</td>\n",
       "      <td>1.4063</td>\n",
       "      <td>-7.3128</td>\n",
       "      <td>0.18732</td>\n",
       "      <td>0.187320</td>\n",
       "      <td>0.63070</td>\n",
       "      <td>1.1559</td>\n",
       "      <td>0.38677</td>\n",
       "      <td>...</td>\n",
       "      <td>0.134850</td>\n",
       "      <td>0.48431</td>\n",
       "      <td>0.86515</td>\n",
       "      <td>0.124440</td>\n",
       "      <td>6.3985</td>\n",
       "      <td>4.3158</td>\n",
       "      <td>127.210</td>\n",
       "      <td>2.8692</td>\n",
       "      <td>7.8980</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 65 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      attr1    attr2    attr3   attr4    attr5    attr6     attr7    attr8  \\\n",
       "0  0.200550  0.37951  0.39641  2.0472  32.3510  0.38825  0.249760  1.33050   \n",
       "1  0.209120  0.49988  0.47225  1.9447  14.7860  0.00000  0.258340  0.99601   \n",
       "2  0.248660  0.69592  0.26713  1.5548  -1.1523  0.00000  0.309060  0.43695   \n",
       "3  0.081483  0.30734  0.45879  2.4928  51.9520  0.14988  0.092704  1.86610   \n",
       "4  0.187320  0.61323  0.22960  1.4063  -7.3128  0.18732  0.187320  0.63070   \n",
       "\n",
       "    attr9   attr10  ...      attr56   attr57   attr58    attr59  attr60  \\\n",
       "0  1.1389  0.50494  ...    0.121960  0.39718  0.87804  0.001924  8.4160   \n",
       "1  1.6996  0.49788  ...    0.121300  0.42002  0.85300  0.000000  4.1486   \n",
       "2  1.3090  0.30408  ...    0.241140  0.81774  0.76599  0.694840  4.9909   \n",
       "3  1.0571  0.57353  ...    0.054015  0.14207  0.94598  0.000000  4.5746   \n",
       "4  1.1559  0.38677  ...    0.134850  0.48431  0.86515  0.124440  6.3985   \n",
       "\n",
       "   attr61   attr62  attr63   attr64  class  \n",
       "0  5.1372   82.658  4.4158   7.4277      0  \n",
       "1  3.2732  107.350  3.4000  60.9870      0  \n",
       "2  3.9510  134.270  2.7185   5.2078      0  \n",
       "3  3.6147   86.435  4.2228   5.5497      0  \n",
       "4  4.3158  127.210  2.8692   7.8980      0  \n",
       "\n",
       "[5 rows x 65 columns]"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cb_df.columns = map(str.lower, cb_df.columns)\n",
    "cb_df['class'] = cb_df['class'].astype('int')\n",
    "cb_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>attr1</th>\n",
       "      <th>attr2</th>\n",
       "      <th>attr3</th>\n",
       "      <th>attr4</th>\n",
       "      <th>attr5</th>\n",
       "      <th>attr6</th>\n",
       "      <th>attr7</th>\n",
       "      <th>attr8</th>\n",
       "      <th>attr9</th>\n",
       "      <th>attr10</th>\n",
       "      <th>...</th>\n",
       "      <th>attr56</th>\n",
       "      <th>attr57</th>\n",
       "      <th>attr58</th>\n",
       "      <th>attr59</th>\n",
       "      <th>attr60</th>\n",
       "      <th>attr61</th>\n",
       "      <th>attr62</th>\n",
       "      <th>attr63</th>\n",
       "      <th>attr64</th>\n",
       "      <th>class</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.200550</td>\n",
       "      <td>0.37951</td>\n",
       "      <td>0.39641</td>\n",
       "      <td>2.0472</td>\n",
       "      <td>32.3510</td>\n",
       "      <td>0.38825</td>\n",
       "      <td>0.249760</td>\n",
       "      <td>1.33050</td>\n",
       "      <td>1.1389</td>\n",
       "      <td>0.50494</td>\n",
       "      <td>...</td>\n",
       "      <td>0.121960</td>\n",
       "      <td>0.39718</td>\n",
       "      <td>0.87804</td>\n",
       "      <td>0.001924</td>\n",
       "      <td>8.4160</td>\n",
       "      <td>5.1372</td>\n",
       "      <td>82.658</td>\n",
       "      <td>4.4158</td>\n",
       "      <td>7.4277</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.209120</td>\n",
       "      <td>0.49988</td>\n",
       "      <td>0.47225</td>\n",
       "      <td>1.9447</td>\n",
       "      <td>14.7860</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.258340</td>\n",
       "      <td>0.99601</td>\n",
       "      <td>1.6996</td>\n",
       "      <td>0.49788</td>\n",
       "      <td>...</td>\n",
       "      <td>0.121300</td>\n",
       "      <td>0.42002</td>\n",
       "      <td>0.85300</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>4.1486</td>\n",
       "      <td>3.2732</td>\n",
       "      <td>107.350</td>\n",
       "      <td>3.4000</td>\n",
       "      <td>60.9870</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.248660</td>\n",
       "      <td>0.69592</td>\n",
       "      <td>0.26713</td>\n",
       "      <td>1.5548</td>\n",
       "      <td>-1.1523</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.309060</td>\n",
       "      <td>0.43695</td>\n",
       "      <td>1.3090</td>\n",
       "      <td>0.30408</td>\n",
       "      <td>...</td>\n",
       "      <td>0.241140</td>\n",
       "      <td>0.81774</td>\n",
       "      <td>0.76599</td>\n",
       "      <td>0.694840</td>\n",
       "      <td>4.9909</td>\n",
       "      <td>3.9510</td>\n",
       "      <td>134.270</td>\n",
       "      <td>2.7185</td>\n",
       "      <td>5.2078</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.081483</td>\n",
       "      <td>0.30734</td>\n",
       "      <td>0.45879</td>\n",
       "      <td>2.4928</td>\n",
       "      <td>51.9520</td>\n",
       "      <td>0.14988</td>\n",
       "      <td>0.092704</td>\n",
       "      <td>1.86610</td>\n",
       "      <td>1.0571</td>\n",
       "      <td>0.57353</td>\n",
       "      <td>...</td>\n",
       "      <td>0.054015</td>\n",
       "      <td>0.14207</td>\n",
       "      <td>0.94598</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>4.5746</td>\n",
       "      <td>3.6147</td>\n",
       "      <td>86.435</td>\n",
       "      <td>4.2228</td>\n",
       "      <td>5.5497</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.187320</td>\n",
       "      <td>0.61323</td>\n",
       "      <td>0.22960</td>\n",
       "      <td>1.4063</td>\n",
       "      <td>-7.3128</td>\n",
       "      <td>0.18732</td>\n",
       "      <td>0.187320</td>\n",
       "      <td>0.63070</td>\n",
       "      <td>1.1559</td>\n",
       "      <td>0.38677</td>\n",
       "      <td>...</td>\n",
       "      <td>0.134850</td>\n",
       "      <td>0.48431</td>\n",
       "      <td>0.86515</td>\n",
       "      <td>0.124440</td>\n",
       "      <td>6.3985</td>\n",
       "      <td>4.3158</td>\n",
       "      <td>127.210</td>\n",
       "      <td>2.8692</td>\n",
       "      <td>7.8980</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 65 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      attr1    attr2    attr3   attr4    attr5    attr6     attr7    attr8  \\\n",
       "0  0.200550  0.37951  0.39641  2.0472  32.3510  0.38825  0.249760  1.33050   \n",
       "1  0.209120  0.49988  0.47225  1.9447  14.7860  0.00000  0.258340  0.99601   \n",
       "2  0.248660  0.69592  0.26713  1.5548  -1.1523  0.00000  0.309060  0.43695   \n",
       "3  0.081483  0.30734  0.45879  2.4928  51.9520  0.14988  0.092704  1.86610   \n",
       "4  0.187320  0.61323  0.22960  1.4063  -7.3128  0.18732  0.187320  0.63070   \n",
       "\n",
       "    attr9   attr10  ...      attr56   attr57   attr58    attr59  attr60  \\\n",
       "0  1.1389  0.50494  ...    0.121960  0.39718  0.87804  0.001924  8.4160   \n",
       "1  1.6996  0.49788  ...    0.121300  0.42002  0.85300  0.000000  4.1486   \n",
       "2  1.3090  0.30408  ...    0.241140  0.81774  0.76599  0.694840  4.9909   \n",
       "3  1.0571  0.57353  ...    0.054015  0.14207  0.94598  0.000000  4.5746   \n",
       "4  1.1559  0.38677  ...    0.134850  0.48431  0.86515  0.124440  6.3985   \n",
       "\n",
       "   attr61   attr62  attr63   attr64  class  \n",
       "0  5.1372   82.658  4.4158   7.4277      0  \n",
       "1  3.2732  107.350  3.4000  60.9870      0  \n",
       "2  3.9510  134.270  2.7185   5.2078      0  \n",
       "3  3.6147   86.435  4.2228   5.5497      0  \n",
       "4  4.3158  127.210  2.8692   7.8980      0  \n",
       "\n",
       "[5 rows x 65 columns]"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cb_df.reset_index()\n",
    "cb_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1584x720 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "X = cb_df.columns\n",
    "y = cb_df.isnull().sum()\n",
    "sns.set()\n",
    "sns.set(rc={'figure.figsize':(22,10)})\n",
    "sns.barplot(X,y)\n",
    "ax = plt.gca()\n",
    "for p in ax.patches:\n",
    "    height = p.get_height()\n",
    "    ax.text(p.get_x() + p.get_width()/2.,\n",
    "            height + 2,\n",
    "            int(height),\n",
    "            fontsize=15, ha='center', va='bottom')\n",
    "sns.set(font_scale=2)\n",
    "ax.set_xlabel(\"Features\")\n",
    "ax.set_ylabel(\"Count of missing values for each feature\")\n",
    "plt.xticks(rotation=90)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**X37**: (current assets - inventories) / long-term liabilities<br>\n",
    "**X21**: sales (n) / sales (n-1)<br>\n",
    "These two features have maximum missing values."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1584x720 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "total_len = len(cb_df['class'])\n",
    "percentage_labels = (cb_df['class'].value_counts()/total_len)*100\n",
    "percentage_labels\n",
    "sns.set()\n",
    "sns.countplot(cb_df['class']).set_title('Data Distribution for Target Variable')\n",
    "ax = plt.gca()\n",
    "for p in ax.patches:\n",
    "    height = p.get_height()\n",
    "    ax.text(p.get_x() + p.get_width()/2.,\n",
    "            height + 2,\n",
    "            '{:.2f}%'.format(100*(height/total_len)),\n",
    "            fontsize=14, ha='center', va='bottom')\n",
    "sns.set(font_scale=3)\n",
    "ax.set_xlabel(\"Classes\")\n",
    "ax.set_ylabel(\"Frequency\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA70AAANSCAYAAACpz6n7AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvhp/UCwAAIABJREFUeJzs3X94W/V99/+X7Cg/iAUWt9JwGwNBcopqYDI/8q1J41JxO5qx2y0O6zAxpt16Ffpdfe97tbjAMFAXCjRU9cYG6y5uauK6DVCw3GHWuoKSMs+QTukoULd26xgMFGOb4TZSiOMIn+8fviPqOkbnnFhCcZ6P6+K6aun9Oe/355yPpLx7jo4chmEYAgAAAABgCcp7vwsAAAAAACBTaHoBAAAAAEsWTS8AAAAAYMmi6QUAAAAALFk0vQAAAACAJYumFwAAAACwZC17vwswI/7ELtOxrs1BvTq5z3T8ae4TJUmH3hgzPcZ5ylrF43HzNblcluMl6/O2HG+jpp+9/FvTYy5Yd6piw6+Zjt/gLZZkft6uzUFJ0oEX+03nWHXu2dk5FhZzTP3iV6bjV57zodmarK7BLKyPXHpdHF4fmd5PkvTIT18wFf/JD/+JpCytWYtjkhNvmo5ftsaTu+vD5JjD8YdeM/+e5iw+1db6+I/Bl03FbzprnSTr+8nW+2yGj8VSWR85WVMOvpdPJA6Yjl9TsCon10cu1mQlRy7WZDeHnZrG42+bHvMB1wm2cuDYxpleAAAAAMCSRdMLAAAAAFiyFq3pHRwcVCwWkyTFYjENDAwsGLtjxw6Fw+HFSg0AAAAAwBEtWtMbjUY1NDQkSers7NT4+Pi8mKmpKTU1NWnnzp2LlRYAAAAAgAXZupFVIpFQc3Oz4vG4JicnFQqF1NXVJafTqdLSUvX29qq/v18lJSWqr6+X1+uV1+tVY2OjtmzZoo0bN2p4eHix5wIAAAAAwBy2mt6RkRHV1NQoFAppbGxMDQ0Nqq2tlcfjUSAQUEVFhaqrq1VUVKTR0VFFIhG53W5J0qZNmxSJRBZ1EgAAAAAAHImtptfj8ai9vV3RaFQFBQVKJpMLxrrd7lTDCwAAAABANtn6Tm9bW5vKysoUDodVVVUlwzDkcDg0MzMjSXI4HDIMYzZBHjeIBgAAAAC8P2yd6Q0Gg2ppaVF3d7cKCwuVn58vv9+v1tZW+Xw+BQIBhcNhFRcXL3a9AAAAAACYZqvpLS8vV09Pz7zHKysrU8/X1dVJkvr6+ubFbd261U5aAAAAAAAscRiHr0MGAAAAAGCJsXWmN9tendxnOvY094mKP7HLdLxrc1CS9PHt/8f0mMev/6zi8bj5HC6X5XhJqv/HDtNjvvu3DZp++RXT8cvXnW6rJqtjHvnpC6bjP/nhP7GUw25Nduad6X2bjWMxPfKq6fjlZ5yWlZoymcNuTVbfbyTpwWeeMxV/xcbzbNWUjWPx3/unTMf/j9Urj9v1kYs1Pfzsz03HX35Rma0cubSfspGDmizmsPjvrlxcH7lYk5UcuVjT4THJsQnT8cvWrsnZY4FjG3eZAgAAAAAsWTS9AAAAAIAla9Eubx4cHNS+ffu0YcMGxWIxuVwu+f3+OTGvv/66brzxRr3zzjsyDEO33nqrvF7vYpUAAAAAAMAci3amNxqNamhoSJLU2dmp8fHxeTF33323rrzySnV0dOiaa65Ra2vrYqUHAAAAAGAeW2d6E4mEmpubFY/HNTk5qVAopK6uLjmdTpWWlqq3t1f9/f0qKSlRfX29vF6vvF6vrr/++tSXwd955x2tWLFiUScDAAAAAMAfstX0joyMqKamRqFQSGNjY2poaFBtba08Ho8CgYAqKipUXV2toqIijY6OKhKJyO12p8YPDw9r+/btuvfeexdtIgAAAAAA/DFbTa/H41F7e7ui0agKCgqUTCYXjHW73XMa3t27d+srX/mK7rrrLr7PCwAAAADIKFvf6W1ra1NZWZnC4bCqqqpkGIYcDodmZmYkSQ6HQ4ZhzCbIezfF7t27dfvtt+v+++/XueeeuwjlAwAAAACwMFtneoPBoFpaWtTd3a3CwkLl5+fL7/ertbVVPp9PgUBA4XBYxcXFc8bdcccdOnTokG644QZJ0plnnqlbb7316GcBAAAAAMAR2Gp6y8vL1dPTM+/xysrK1PN1dXWSpL6+vtTzjz32mJ10AAAAAADY4jAOX4cMAAAAAMASs2i/0wsAAAAAQK6xdXlzth16Y8x0rPOUtfr49v9jOv7x6z8rSYo/scv0GNfmoOLxuPl4l0vJiTdNxy9b48lKTVbjJVnPYXEOVnLYrikb86am9z2H3ZqmX3nNdPzy02fvW2B2nVtd44drsrOfrL5vJscmTMcvW7vmuF0fOVlTBt9nUzlyaD9lIwc1WRvzsZZ7TMf/pKUxJ9dHLtZkJUcu1nR4zFtvT5mOP/mElTl7LHBs40wvAAAAAGDJoukFAAAAACxZi3Z58+DgoPbt26cNGzYoFovJ5XLJ7/fPiZmYmFBTU5MOHTqkNWvW6Gtf+5pWrVq1WCUAAAAAADDHop3pjUajGhoakiR1dnZqfHx8Xsx9992n2tpa7dy5UyUlJXr44YcXKz0AAAAAAPPYOtObSCTU3NyseDyuyclJhUIhdXV1yel0qrS0VL29verv71dJSYnq6+vl9Xrl9Xp14403yjAMzczMaHR0VOvWrVvk6QAAAAAA8C5bTe/IyIhqamoUCoU0NjamhoYG1dbWyuPxKBAIqKKiQtXV1SoqKtLo6KgikYjcbrckKZlM6s///M918OBBff7zn1/UyQAAAAAA8IdsNb0ej0ft7e2KRqMqKChQMplcMNbtdqcaXklyOp36wQ9+oGeeeUbXX3+9vvOd79gpAQAAAACAtGx9p7etrU1lZWUKh8OqqqqSYRhyOByamZmRJDkcDhmGMZsg790ULS0t2r17tyRp9erVcjgcR1s/AAAAAAALsnWmNxgMqqWlRd3d3SosLFR+fr78fr9aW1vl8/kUCAQUDodVXFw8Z1xDQ4NaWlp07733Ki8vTy0tLYsxBwAAAAAAjshW01teXq6enp55j1dWVqaer6urkyT19fWlnvf5fOro6LCTEgAAAAAAyxzG4euQAQAAAABYYmyd6c22eDxuOtblclmOt5XjiV3m4zcHs1NThueQjZqs5DgcP/WrQdM5Vn7orJw8FlbnICnjazAX95OVHLlc08Hf7DWdY8V6X04ei1ysSZIO7n3JVPwK35lZq+lYXrOpHEvgM89Kjmztp6zUZGc/HYefL7lYk5UcuViT3RzJiTdNxy9b48lKTTj22bqRFQAAAAAAxwKaXgAAAADAkrVoTe/g4KBisZgkKRaLaWBgYMHYWCymiy++eLFSAwAAAABwRIvW9EajUQ0NDUmSOjs7NT4+fsS40dFRtbW1KZlMLlZqAAAAAACOyNaNrBKJhJqbmxWPxzU5OalQKKSuri45nU6Vlpaqt7dX/f39KikpUX19vbxer7xer5qamvTlL39Zt912m7Zu3brYcwEAAAAAYA5bTe/IyIhqamoUCoU0NjamhoYG1dbWyuPxKBAIqKKiQtXV1SoqKtLo6KgikYjcbream5v113/911q7du1izwMAAAAAgHlsNb0ej0ft7e2KRqMqKCh4z0uV3W633G63xsbGtGfPHr3yyiu699579fvf/15f+MIX9Pd///e2iwcAAAAA4L3Yanrb2tpUVlambdu2affu3Xr66aflcDg0MzMjSXI4HDIMQ5KUlzf7teG1a9fqRz/6UWobH/nIR2h4AQAAAAAZZavpDQaDamlpUXd3twoLC5Wfny+/36/W1lb5fD4FAgGFw2EVFxcvdr0AAAAAAJhmq+ktLy9XT0/PvMcrKytTz9fV1UmS+vr6jriNhR4HAAAAAGCxOIzD1yEDAAAAALDE2DrTm23xeNx0rMvlshwvScmJN02PWbbGY72mJ3aZj98clJT5edvZT7mUIxdrspuDmhY/BzWxPqjp/cuRizVZyUFNx9/6yMWarOTIxZrs5sjFmnDsy3u/CwAAAAAAIFNoegEAAAAAS9aiXd48ODioffv2acOGDYrFYnK5XPL7/XNifve73+lP//RP9cEPflDS7I2vPvWpTy1WCQAAAAAAzLFoTW80GpXH49GGDRvU2dmp6urqeU3vL3/5S3384x/XzTffvFhpAQAAAABYkK2mN5FIqLm5WfF4XJOTkwqFQurq6pLT6VRpaal6e3vV39+vkpIS1dfXy+v1yuv1as2aNerv79eVV16pk08+WTfddJM+8IEPLPacAAAAAACQZLPpHRkZUU1NjUKhkMbGxtTQ0KDa2lp5PB4FAgFVVFSourpaRUVFGh0dVSQSkdvt1pNPPqlzzjlHGzdu1GOPPaavfvWr+sd//MfFnhMAAAAAAJJsNr0ej0ft7e2KRqMqKChQMplcMNbtdsvtdkuSysvLtWrVKknS5s2baXgBAAAAABll6+7NbW1tKisrUzgcVlVVlQzDkMPh0MzMjCTJ4XDIMIzZBHnvprjpppv0ox/9SJL07LPP6uyzzz7a+gEAAAAAWJCtM73BYFAtLS3q7u5WYWGh8vPz5ff71draKp/Pp0AgoHA4rOLi4jnjrr32Wt1444168MEHtWrVKn31q19dlEkAAAAAAHAktpre8vJy9fT0zHu8srIy9XxdXZ0kqa+vL/X8aaedpo6ODjspAQAAAACwzGEcvg4ZAAAAAIAlxtZ3egEAAAAAOBbYurw52+JP7DId69ocVP0/mr+E+rt/22ArRzweNx/vclmOz9marOawOAdJmvrVoKn4lR86y15N2Zg3NWUsx8Hf7DUVv2K9z15NNtas2TGp+Bw8FofeGDMd7zxlbc6uD7Nj7NY0NfBr0/Er/R/MSk221uxx+v6R8WORhZqsrsFsHIu/+dajpuP/+TN/kZPrIxdrspIjF2uymyMXa8KxjzO9AAAAAIAli6YXAAAAALBkLVrTOzg4qFgsJkmKxWIaGBiYF/P222/ruuuu07Zt2/TJT35SL7zwwmKlBwAAAABgnkVreqPRqIaGhiRJnZ2dGh8fnxfzrW99S+vXr9fOnTt12223aXh4eLHSAwAAAAAwj60bWSUSCTU3Nysej2tyclKhUEhdXV1yOp0qLS1Vb2+v+vv7VVJSovr6enm9Xnm9Xr3wwgu69NJL9ZnPfEarV6/Wl7/85cWeDwAAAAAAKbaa3pGREdXU1CgUCmlsbEwNDQ2qra2Vx+NRIBBQRUWFqqurVVRUpNHRUUUiEbndboVCIe3bt0/f+ta39P3vf1/bt2/XXXfdtdhzAgAAAABAks2m1+PxqL29XdFoVAUFBUomkwvGut1uud1uSVJhYaEuueQSSVIwGNR9991nJz0AAAAAAKbY+k5vW1ubysrKFA6HVVVVJcMw5HA4NDMzI0lyOBwyDGM2Qd67KS644AI9/fTTkmZvdlVSUnK09QMAAAAAsCBbZ3qDwaBaWlrU3d2twsJC5efny+/3q7W1VT6fT4FAQOFwWMXFxXPGXXPNNbrpppt0+eWXa9myZdq+ffuiTAIAAAAAgCOx1fSWl5erp6dn3uOVlZWp5+vq6iRJfX19qecLCwt1zz332EkJAAAAAIBlDuPwdcgAAAAAACwxts70Zlv8iV2mY12bg5p++RXT8cvXnT6bIx43n8Plyni8rZos7idbNVnM8fae50zHn3DheZZyuDYHJUkHf7PXdI4V631ZmbfVHNPDL5uOX+5dN1tTLq6PbLwuLK6PTO8nSZpIHDAVv6ZglSQpOfGm6RzL1nhy8ljkYk2SjfVh9XVhY31kes0e+PmLpuNXlZ1rK8eSWR8mx+R0TUvgvdxOTdOvvGZ6zPLTiy3H5+J+spLDbk2H3hgzHe88Za0k6dBrvzU/pvjUrHwOT/UPmB6z8my/rWOBY5utG1kBAAAAAHAsoOkFAAAAACxZi3Z58+DgoPbt26cNGzYoFovJ5XLJ7/fPibn99ts1MDB7+cHExIROPPFEfe9731usEgAAAAAAmGPRzvRGo1ENDQ1Jkjo7OzU+Pj4vprm5WR0dHWpra5PL5dJtt922WOkBAAAAAJjH1pneRCKh5uZmxeNxTU5OKhQKqaurS06nU6Wlpert7VV/f79KSkpUX18vr9crr9er5uZmSdJ3vvMdfeQjH9FZZ521qJMBAAAAAOAP2Wp6R0ZGVFNTo1AopLGxMTU0NKi2tlYej0eBQEAVFRWqrq5WUVGRRkdHFYlE5Ha7JUnT09N66KGH9Oijjy7qRAAAAAAA+GO2ml6Px6P29nZFo1EVFBQomUwuGOt2u1MNryQ9++yz2rBhA7f/BgAAAABknK3v9La1tamsrEzhcFhVVVUyDEMOh0MzMzOSJIfDIcMwZhPkzU3xzDPP6KMf/ehRlg0AAAAAQHq2zvQGg0G1tLSou7tbhYWFys/Pl9/vV2trq3w+nwKBgMLhsIqLi+eNfemll7Rly5ajLhwAAAAAgHRsNb3l5eXq6emZ93hlZWXq+bq6OklSX1/fnJj77rvPTkoAAAAAACxzGIevQwYAAAAAYIlZtN/pBQAAAAAg19i6vDnb4vG46ViXy2U53laOJ3aZj98czE5NGZ5DNmqykiMVf7weiwzPOxf3k5UcOV3TUlizOViTJE2PvGoqfvkZp2WtJqvrIznxpukcy9Z4+MzLQI5s7aes1GRnPx2Hny+5WJOVHLlYk90cVt8Ds1ETjn2c6QUAAAAALFmLdqZ3cHBQ+/bt04YNGxSLxeRyueT3++fEvP7667ruuutkGIZOOukkfeMb39CqVasWqwQAAAAAAOZYtDO90WhUQ0NDkqTOzk6Nj4/Pi9mxY4cuvfRSffe739X69ev16KOPLlZ6AAAAAADmsXWmN5FIqLm5WfF4XJOTkwqFQurq6pLT6VRpaal6e3vV39+vkpIS1dfXy+v1yuv1qrS0VG+88UZqG6eccsqiTgYAAAAAgD9kq+kdGRlRTU2NQqGQxsbG1NDQoNraWnk8HgUCAVVUVKi6ulpFRUUaHR1VJBKR2+3Ws88+q2984xt6/PHHNT09rcbGxsWeDwAAAAAAKbaaXo/Ho/b2dkWjURUUFCiZTC4Y63a75Xa7JUl33XWX7rzzTlVUVOgnP/mJrr/+et133332KgcAAAAAIA1b3+lta2tTWVmZwuGwqqqqZBiGHA6HZmZmJEkOh0OGYcwmyHs3xYknnpi67fcHPvAB7du372jrBwAAAABgQbbO9AaDQbW0tKi7u1uFhYXKz8+X3+9Xa2urfD6fAoGAwuGwiouL54y7+eabdeutt2pmZkaGYeiWW25ZlEkAAAAAAHAktpre8vJy9fT0zHu8srIy9XxdXZ0kqa+vL/V8SUmJvv3tb9tJCQAAAACAZQ7j8HXIAAAAAAAsMYv2O70AAAAAAOQaW5c3Z9vPXv6t6dgL1p2qeDxuOv7wjbWsjsl0vK2anthlPn5z0F5NVnNYjLeS43D8wd/sNZ1jxXpfduZtMYfVOUg5uj6y8bqwuD4yvZ+yUlMWjsXYvv2m49eeuHrprI8cfE/LyppdCp95GcyR0zUtgfdyOzUd3PuS6TErfGdajs/F/WQlh92aDr1m/t/YzuJTJUnTwy+bHrPcuy4r72mJ3mdMjymo2GjrWODYxpleAAAAAMCSRdMLAAAAAFiyFq3pHRwcVCwWkyTFYjENDAzMi3n11VdVX1+vbdu2qampSQcOHFis9AAAAAAAzLNoTW80GtXQ0JAkqbOzU+Pj4/Nivv71r6uurk47d+7Uhz/8YT3wwAOLlR4AAAAAgHls3cgqkUioublZ8Xhck5OTCoVC6urqktPpVGlpqXp7e9Xf36+SkhLV19fL6/XK6/VqaGhIt912myTp/PPP1x133LGokwEAAAAA4A/ZanpHRkZUU1OjUCiksbExNTQ0qLa2Vh6PR4FAQBUVFaqurlZRUZFGR0cViUTkdrt17bXX6qmnnlJtba1+/OMfc3kzAAAAACCjbDW9Ho9H7e3tikajKigoUDKZXDDW7XbL7XZLkq6//nrddtttevzxx3XRRRelHgcAAAAAIBNsfae3ra1NZWVlCofDqqqqkmEYcjgcmpmZkSQ5HA4ZhjGbIO/dFM8884w+//nP61vf+pby8vK0cePGRZgCAAAAAABHZutMbzAYVEtLi7q7u1VYWKj8/Hz5/X61trbK5/MpEAgoHA6ruLh4zrgzzzxTN954o5YvX67169frlltuWZRJAAAAAABwJLaa3vLycvX09Mx7vLKyMvV8XV2dJKmvry/1fCAQUCQSsZMSAAAAAADLHMbh65ABAAAAAFhibJ3pzbbY8GumYzd4i/XIT18wHf/JD/+JJCn+xC7TY1ybg4rH4+bjXS7L8XZqenvPc6bjT7jwvKzUZDVekhI/+Q9T8QUf2zRbUxaORab3ra1jYXFMVtZHNo5F7L9MxZ+w4Xx7NdlYs2bHpOJzcH289faU6fiTT1iZs+vD7Bi7NZl9f5L+4D0q0+vDzprN8LE4XtdHNmo6+Ju9pnOsWO/LyrF47Ge/NB3/ZxeU5uS/P1izuZPDVk0/ftr8mP91sa0cOLbZupEVAAAAAADHAppeAAAAAMCSZavpHRwcVCwWkyTFYjENDAwsGLtjxw6Fw+HU30899ZQuu+wyXX755fre975nJz0AAAAAAKbYanqj0aiGhoYkSZ2dnRofH58XMzU1paamJu3cuTP12KFDh3TnnXeqra1NHR0devjhhzUxMWGzdAAAAAAA3lvaG1klEgk1NzcrHo9rcnJSoVBIXV1dcjqdKi0tVW9vr/r7+1VSUqL6+np5vV55vV41NjZqy5Yt2rhxo4aHhyVJe/fu1emnn66TTjpJknTBBRdoz549uvTSSzM7SwAAAADAcSlt0zsyMqKamhqFQiGNjY2poaFBtbW18ng8CgQCqqioUHV1tYqKijQ6OqpIJCK32y1J2rRp05zf5U0kEnPugLZ69WolEokMTAsAAAAAABNNr8fjUXt7u6LRqAoKCpRMJheMdbvdqYb3SAoKCrR///7U3/v37+c24AAAAACAjEn7nd62tjaVlZUpHA6rqqpKhmHI4XBoZmZGkuRwOGQYxuzG8t57cz6fTyMjI/rd736n6elp7dmzR+edd94iTAMAAAAAgPnSnukNBoNqaWlRd3e3CgsLlZ+fL7/fr9bWVvl8PgUCAYXDYRUXF6dN5nQ6dcMNN+gzn/mMDMPQZZddprVr1y7KRAAAAAAA+GNpm97y8nL19PTMe7yysjL1fF1dnSSpr69vXtzWrVvn/H3JJZfokksusVUsAAAAAABWOIzD1yYDAAAAALDE2PqdXgAAAAAAjgVpL2/OBfEndpmOdW0OKh6Pm4//v3ePtjom0/GSNPWrQdNjVn7orOzsJ4s5Ej/5D9PxBR/bZCmHa3NQko39lIV5W81x8Dd7TcevWO+brcnqGlwCryPJ+vrI9H6SpLfenjIVf/IJKyVJyYk3TedYtsaTk8ciF2uSbKwPq68LG+sj02vW1vvs8bo+TI7J6ZqWwHu5nZqmX37F9Jjl6063HJ+L+8lKDrs1HXpjzHS885TZ+/BMD79sesxy77qsfA4feP4XpsesCpxj61jg2MaZXgAAAADAkmWr6R0cHFQsFpMkxWIxDQwMLBi7Y8cOhcPhOY8dOHBAdXV12rvX/NktAAAAAACsstX0RqNRDQ0NSZI6Ozs1Pj4+L2ZqakpNTU3auXPnnMdffPFF1dfX69VXX7WTGgAAAAAA09J+pzeRSKi5uVnxeFyTk5MKhULq6uqS0+lUaWmpent71d/fr5KSEtXX18vr9crr9aqxsVFbtmzRxo0bNTw8nNre9PS07r33Xl133XUZnRgAAAAAAGmb3pGREdXU1CgUCmlsbEwNDQ2qra2Vx+NRIBBQRUWFqqurVVRUpNHRUUUiEbndbknSpk2bFIlE5mzvggsuyMxMAAAAAAD4I2mbXo/Ho/b2dkWjURUUFCiZTC4Y63a7Uw0vAAAAAADvt7Tf6W1ra1NZWZnC4bCqqqpkGIYcDodmZmYkSQ6HQ4ZhzG4sj5tBAwAAAAByR9ozvcFgUC0tLeru7lZhYaHy8/Pl9/vV2toqn8+nQCCgcDis4uLibNQLAAAAAIBpaZve8vJy9fT0zHu8srIy9XxdXZ0kqa+vb17c1q1bj7jdjo4OS4UCAAAAAGAV1yMDAAAAAJYsh3H4C7kAAAAAACwxaS9vzgUHXuw3Hbvq3LMVj8dNx7tcLkmyPGbqV4Om41d+6Kys1HTwN3tNx69Y71P8iV3mt785KEnWc9iYt9l9u/JDZ0mS5XnYqSnT87ZzLKyOObj3JdPxK3xnKjk2YTp+2do1szVZnbed9THwa1PxK/0flCTL85h+5TXT8ctPn72Xgdl5pF7bWVizmT4Wdmqyeizs5Djw/C9Mxa8KnCPJ+rGw+t4vSdPDL5uKX+5dN1tTFo5Fxj9fcnDNWsmRrX8b2KkpG//+yHSOXFwfuViTlRx2azr02m9NxzuLT7WVw2q82fdx6d338uTEm6bHLFvjsXUscGzj8mYAAAAAwJJF0wsAAAAAWLJsNb2Dg4OKxWKSpFgspoGBgQVjd+zYoXA4nPr78ccf1yc/+UnV1dXplltuSf3eLwAAAAAAi81W0xuNRjU0NCRJ6uzs1Pj4+LyYqakpNTU1aefOnXMe+4d/+Ad9+9vf1kMPPaREIqFdu8x/nwoAAAAAACvS3sgqkUioublZ8Xhck5OTCoVC6urqktPpVGlpqXp7e9Xf36+SkhLV19fL6/XK6/WqsbFRW7Zs0caNGzU8PCxJWr58uR566CGtWrVKkpRMJrVixYrMzhAAAAAAcNxK2/SOjIyopqZGoVBIY2NjamhoUG1trTwejwKBgCoqKlRdXa2ioiKNjo4qEonI7XZLkjZt2qRIJJLaVl5enjwejySpo6NDb7/9tj7ykY9kaGoAAAAAgONd2qbX4/Govb1d0WhUBQUFSiaTC8a63e5Uw7uQmZkZff3rX9dLL72kf/qnf5LD4bBeNQAAAAAAJqT9Tm9bW5vKysoUDodVVVUlwzDkcDhSN6ByOBwyDGN2Y3npvyJ8yy236ODBg/qnxzi6AAAgAElEQVTnf/7n1GXOAAAAAABkQtozvcFgUC0tLeru7lZhYaHy8/Pl9/vV2toqn8+nQCCgcDis4uLitMn6+/v16KOP6sILL9SnPvUpSdJVV12lzZs3H/1MAAAAAAD4I2mb3vLycvX09Mx7vLKyMvV8XV2dJKmvr29e3NatW1P/++yzz37PnzcCAAAAAGAxOYzD1yYDAAAAALDEpD3Tmwvi8bjpWJfLZTk+GzmoafFz2K7pCfO/De3aHLSXI4f20+ExyYk3TccvW+Oxt58sjsnK+sjFmnJwfWTldcSxyJmapvrNX3W18mx/dvaT1fWRwTWYrfd+Kznsro+/e/DfTMffeUWNrRy5tJ+yWdP/t6PL9Ji7P12b0ddFLq7ZVI4ceu9P1WQjB45t6e88BQAAAADAMYqmFwAAAACwZNlqegcHBxWLxSRJsVjsPW9OtWPHDoXD4dTfP/rRj3TZZZfpL/7iL/TII4/YSQ8AAAAAgCm2mt5oNKqhoSFJUmdnp8bHx+fFTE1NqampSTt37kw99s477+gb3/iGduzYoYcfflj333+/3nrrLZulAwAAAADw3tLeyCqRSKi5uVnxeFyTk5MKhULq6uqS0+lUaWmpent71d/fr5KSEtXX18vr9crr9aqxsVFbtmzRxo0bNTw8LEnKz8/XD37wAy1btkz//d//LUlavXp1ZmcIAAAAADhupW16R0ZGVFNTo1AopLGxMTU0NKi2tlYej0eBQEAVFRWqrq5WUVGRRkdHFYlE5Ha7JUmbNm1SJBKZm3DZMkWjUd166626+OKLtWzZMXEDaQAAAADAMSjt5c0ej0dPPvmkmpqa9M1vflPJZHLBWLfbnWp430soFNK///u/69ChQ/r+979vrWIAAAAAAExKe5q1ra1NZWVl2rZtm3bv3q2nn35aDodDMzMzkiSHwyHDMCRJeXnv3UMnEgl97nOfU1tbm5YvX65Vq1alHQMAAAAAODpWfjPZjsO/F52L0ja9wWBQLS0t6u7uVmFhofLz8+X3+9Xa2iqfz6dAIKBwOKzi4uK0yQoKCvSJT3xC9fX1WrZsmc466yz92Z/92aJMBAAAAACAP5a26S0vL1dPT8+8xysrK1PP19XVSZL6+vrmxW3dunXO35dffrkuv/xyW8UCAAAAAGCFwzh8bTIAAAAAYEmK//jpjG7f9b8uzuj2jwZfqAUAAAAALFnHxO8FWfnStWtzUNMvv2I6fvm602dzxOPmc7hcGY+3VZPF/WSrJos53t7znOn4Ey48z1KOw1+WP/ibvaZzrFjvy8q8reaYHn7ZdPxy77rZmnJxfWTjdWFxfWR6P0nSROKAqfg1BaskScmJN03nWLbGk5PHIhdrkmysD6uvCxvrI9Nr9sDPXzQdv6rsXFs5lsz6MDkmp2taAu/ldmqafuU102OWn15sOT4X95OVHHZrOvTGmOl45ylrJUmHXvut+THFp2blc3iqf8D0mJVn+20dCxzbONMLAAAAAFiyjokzvQAAAAAA+xwOx/tdwvvG1pnewcFBxWIxSVIsFtPAwMKXFOzYsUPhcHje4zfffPMRHwcAAAAAYLHYanqj0aiGhoYkSZ2dnRofH58XMzU1paamJu3cuXPecw899JB+/etf20kNAAAAALAqLy+z/+WwtJc3JxIJNTc3Kx6Pa3JyUqFQSF1dXXI6nSotLVVvb6/6+/tVUlKi+vp6eb1eeb1eNTY2asuWLdq4caOGh4dT23vuuef0/PPP6/LLL5/zOAAAAAAAiy1t0zsyMqKamhqFQiGNjY2poaFBtbW18ng8CgQCqqioUHV1tYqKijQ6OqpIJCK32y1J2rRpkyKRSGpb4+Pjuueee3TPPffohz/8YeZmBQAAAACATDS9Ho9H7e3tikajKigoUDKZXDDW7XanGt4j6enp0eTkpK6++mpNTExoampKXq9XW7dutVc9AAAAACC94/hGVmmb3ra2NpWVlWnbtm3avXu3nn76aTkcDs3MzEiavQuYYRiSpLw013JfddVVuuqqqyRJkUhEw8PDNLwAAAAAgIxJ2/QGg0G1tLSou7tbhYWFys/Pl9/vV2trq3w+nwKBgMLhsIqLi7NRLwAAAADAqjzO9C6ovLxcPT098x6vrKxMPV9XVydJ6uvrmxe30JlczvACAAAAADLNYRy+NhkAAAAAsCQlep/J6PYLKjZmdPtHI+2Z3lwQj8dNx7pcLsvx2chBTYufIxdrSuV4Ypf5+M3BnDwWuViTlRzUxPqgpvcvRy7WZCUHNR1/6yMXa7KSIxdrspsjF2vC4piZmVFLS4sGBwe1fPlyffWrX9UZZ5yRen7Hjh36t3/7N0nSxRdfrMbGRhmGoY9+9KNat26dJKmsrEzXXnut5dzHRNMLAAAAALDP4Xjvmw5n2pNPPqnp6Wk9/PDD+vnPf66vfe1r+uY3vylJevXVV/XYY4/pkUcekcPh0LZt21RZWalVq1bp7LPP1r/8y78cVe73d+YAAAAAgCXvZz/7mSoqKiTNnrH9xS9+kXrulFNO0f3336/8/Hzl5eUpmUxqxYoV6u/v19jYmBoaGvTZz35Ww8PDtnLbanoHBwcVi8UkSbFYTAMDAwvG7tixQ+FwOPX3Aw88oJqaGjU0NKihocF24QAAAAAAkxyOzP6XRiKRUEFBQerv/Px8JZNJSZLT6dTJJ58swzC0fft2lZaW6swzz9SaNWt09dVXq6OjQ9dcc42+9KUv2Zq6rcubo9GoPB6PNmzYoM7OTlVXV8vv98+JmZqa0k033aQXXnhBoVAo9Xh/f7+2b9+uc845x1bBAAAAAIBjS0FBgfbv35/6e2ZmRsuWvduOHjx4UDfeeKNWr16tL3/5y5Kkc845R/n5+ZKkCy+8UGNjYzIMQw4TTfYfStv0JhIJNTc3Kx6Pa3JyUqFQSF1dXXI6nSotLVVvb6/6+/tVUlKi+vp6eb1eeb1eNTY2asuWLdq4ceOcs7n9/f267777NDExoY997GO65pprLBUMAAAAADi2nH/++dq1a5eqq6v185//XB/84AdTzxmGob/5m7/Rhz/8YV199dWpx++55x4VFhbqs5/9rAYGBlRUVGS54ZVMNL0jIyOqqalRKBRKXU9dW1srj8ejQCCgiooKVVdXq6ioSKOjo4pEInK73ZKkTZs2KRKJzNleTU2Ntm3bpoKCAjU2NmrXrl0KBoOWCwcAAAAAHBs2b96svr4+1dXVyTAM3XHHHXrggQd0+umna2ZmRv/5n/+p6elp9fb2SpK++MUv6uqrr9aXvvQlPf3008rPz9edd95pK3faptfj8ai9vV3RaFQFBQWp666PxO12pxreIzEMQ5/61KdSt/6++OKL9ctf/pKmFwAAAAAyKc/6GdJFTZ+Xp1tvvXXOYz6fL/W/X3zxxSOOu++++44+d7qAtrY2lZWVKRwOq6qqKnUN9czMjCTJ4XDIMIzZjeW99+YSiYQ+/vGPa//+/TIMQz/96U/5bi8AAAAAIGPSnukNBoNqaWlRd3e3CgsLlZ+fL7/fr9bWVvl8PgUCAYXDYRUXF6dN5nK59IUvfEFXXXWVli9frosuukgXX3zxokwEAAAAALCA9/l3et9PaZve8vJy9fT0zHu8srIy9XxdXZ0kqa+vb17c1q1b5/y9ZcsWbdmyxVaxAAAAAABY4TAOX5sMAAAAAFiS9u/ek9Htry6/MKPbPxq2fqc326Z+8SvTsSvP+ZDi8bjp+MM31bI6ZupXg+Zr+tBZWalpevhl0/HLvesUf2KX+e1vnr3Z2MHf7DU9ZsV6n615m82xYv3sF9+tzsNOTZb3rcUcdo6F1TFW55CceNN0/LI1ntmarM47C+vD6jwO/XbUdLzz1P8pyfy8U6/tLKzZTB8LOzVZPRZ2crz9nz8zFX/C/3OBJOvHwup7oGT+/WO5d52k7Owny595A782Hb/S/8GcXLNWcmTr3wZ2asrG5/BU/4DpMSvP9ufUsbObIxdrspLDbk2HXvut6Xhn8am2cliNf3vPc6bjT7jwPElScmzC9Jhla9fYOhY4th0TTS8AAAAAwD7H+3z35vcTTS8AAAAALHWO47fptXULr8HBQcViMUlSLBbTwMDCl8Hs2LFD4XA49fcLL7ygbdu26YorrtDf/u3f6uDBg3ZKAAAAAAAgLVtNbzQa1dDQkCSps7NT4+Pj82KmpqbU1NSknTt3ph4zDEM333yz7rzzTj344IOqqKjQb39r/rsEAAAAAABYkfby5kQioebmZsXjcU1OTioUCqmrq0tOp1OlpaXq7e1Vf3+/SkpKVF9fL6/XK6/Xq8bGRm3ZskUbN27U8PCwJOmll15SYWGh2tvb9etf/1oXX3yxvF5vxicJAAAAADg+pW16R0ZGVFNTo1AopLGxMTU0NKi2tlYej0eBQEAVFRWqrq5WUVGRRkdHFYlE5Ha7JUmbNm1SJBJJbWtyclLPPfecbr75Zp1xxhn63Oc+p3POOUcXXXRR5mYIAAAAAMe74/g7vWmbXo/Ho/b2dkWjURUUFCiZTC4Y63a7Uw3vkRQWFuqMM85QSUmJJKmiokK/+MUvaHoBAAAAABmR9ju9bW1tKisrUzgcVlVVlQzDkMPh0MzMjCTJ4XDIMIzZjeW99+ZOO+007d+/XyMjI5KkPXv2aP369Uc7BwAAAAAAjijtmd5gMKiWlhZ1d3ersLBQ+fn58vv9am1tlc/nUyAQUDgcVnFxcdpky5cv1+23365rr71WhmHovPPO08c+9rHFmAcAAAAAAPOkbXrLy8vV09Mz7/HKysrU83V1dZKkvr6+eXFbt26d8/dFF12kRx991FaxAAAAAADrHGmuyl3KHMbha5MBAAAAAEvSgf96PqPbX3V+IKPbPxppz/QCAAAAAI5xx/GZ3mOi6Y3H46ZjXS6X5XhbOZ7YZT5+czA7NWV4DtmoyUqOVPzxeiwyPO9c3E9WcuR0TUthzeZgTZI0PfKqqfjlZ5yWtZqsro/kxJumcyxb4+EzLwM5srWfslKTnf10HH6+5GJNVnLkYk12c1h9D8xGTTj2Hb/tPgAAAABgyaPpBQAAAAAsWbYubx4cHNS+ffu0YcMGxWIxuVwu+f3+I8bu2LFDb775ppqamjQxMaEvfvGLqed+9atf6dprr9UVV1xhr3oAAAAAQHoOx/tdwfvG1pneaDSqoaEhSVJnZ6fGx8fnxUxNTampqUk7d+5MPbZmzRp1dHSoo6NDX/ziF1VaWqq//Mu/tFk6AAAAAADvLe2Z3kQioebmZsXjcU1OTioUCqmrq0tOp1OlpaXq7e1Vf3+/SkpKVF9fL6/XK6/Xq8bGRm3ZskUbN27U8PDwnG0ahqHbbrtN4XBY+fn5GZscAAAAAEBS3vF7pjdt0zsyMqKamhqFQiGNjY2poaFBtbW18ng8CgQCqqioUHV1tYqKijQ6OqpIJCK32y1J2rRpkyKRyLxtPvXUU1q/fr28Xu/izwgAAAAAgP8rbdPr8XjU3t6uaDSqgoICJZPJBWPdbneq4X0vjz32mK666iprlQIAAAAAYFHa7/S2tbWprKxM4XBYVVVVMgxDDodDMzMzkiSHwyHDMGY3ZvIHj/v7+3X++ecfRdkAAAAAAKSX9kxvMBhUS0uLuru7VVhYqPz8fPn9frW2tsrn8ykQCCgcDqu4uNhUwrfeekurV6+W4zi+exgAAAAAZJPDcfz+Wm3apre8vFw9PT3zHq+srEw9X1dXJ0nq6+ubF7d169Y5f5988sn613/9V1vFAgAAAABghcM4fG0yAAAAAGBJmuofyOj2V57tz+j2j0baM725IP7ELtOxrs1BTY+8ajp++RmnzeaIx83ncLkyHm+rJov7yVZNFnO8vec50/EnXHiepRyuzUFJ0sG9L5nOscJ3ppITb5qOX7bGY6mmw3VZ3bfTwy+bjl/uXTdbUy6uj2y8Liyuj0zvJ0maSBwwFb+mYJUkWV6DuXgscrEmycb6sPq6sLE+Mr1mD/z8RdPxq8rOtZVjyawPk2NyuqYl8F5up6bpV14zPWb56cWW43NxP1nJYbemQ2+MmY53nrJWknTotd+aH1N8alY+h600cyvP9ts6Fji2Hb8XdgMAAAAAlrxj4kwvAAAAAOAo5B2/NxK2daZ3cHBQsVhMkhSLxTQwsPAlBTt27FA4HE79/dhjj6m2tlaXXXaZdu7caSc9AAAAAACm2Gp6o9GohoaGJEmdnZ0aHx+fFzM1NaWmpqZ5je1dd92lBx54QA8++KAeeOAB/f73v7dTAgAAAADALIcjs//lsLSXNycSCTU3Nysej2tyclKhUEhdXV1yOp0qLS1Vb2+v+vv7VVJSovr6enm9Xnm9XjU2NmrLli3auHGjhoeHU9s766yzFI/HtWzZMhmGwe/1AgAAAAAyJm3TOzIyopqaGoVCIY2NjamhoUG1tbXyeDwKBAKqqKhQdXW1ioqKNDo6qkgkIrfbLUnatGmTIpHInO2tX79el112mVatWqXNmzfrxBNPzMzMAAAAAADHvbSXN3s8Hj355JNqamrSN7/5TSWTyQVj3W53quE9koGBAf3kJz/Rj3/8Yz311FN666239MMf/tBe5QAAAAAApJH2TG9bW5vKysq0bds27d69W08//bQcDodmZmYkSQ6HQ4ZhSJLy8t67h3a5XFq5cqVWrFih/Px8nXzyydq3b98iTAMAAAAAsBCH4/j9tdq0TW8wGFRLS4u6u7tVWFio/Px8+f1+tba2yufzKRAIKBwOq7i4OG2yU089VZdffrm2bdsmp9Op008/XbW1tYsyEQAAAAAA/ljapre8vFw9PT3zHq+srEw9X1dXJ0nq6+ubF7d169Y5f19xxRW64oorbBULAAAAALDhOP6dXodx+NpkAAAAAMCSdPA3ezO6/RXrfRnd/tE4fi/sBgAAAAAseWkvb84F8XjcdKzL5bIcn40c1LT4OQ7HJ8cmTOdYtnaN4k/sMl/T5qClmg7XZTU+OfGm6fhlazxZqcnOfno79l+mx5yw4fysrA+r88jJNZuDr9Xp4ZdNxy/3rputKQePRS6+p2WjpkNvjJke4zxlbVZqSjw9/ytSCym4+CM68PMXTcevKjtXknTgxX5z8eeeLWlprI/BN8x/vpx1SpY+X3LwPc1OTc+NvG56zHlnFNnKYfZz9YQN50vKrTV7eIyVs4sr1vtycn3g2HdMNL0AAAAAgKOQ5pd2ljJbMx8cHFQsFpMkxWIxDQwMLBi7Y8cOhcPh1N/f//739YlPfELbtm3TI488Yic9AAAAAACm2Gp6o9GohoaGJEmdnZ0aHx+fFzM1NaWmpibt3Lkz9dhbb72lu+++Wx0dHfrOd76j7u5uvfbaazZLBwAAAACY4XA4MvpfLkt7eXMikVBzc7Pi8bgmJycVCoXU1dUlp9Op0tJS9fb2qr+/XyUlJaqvr5fX65XX61VjY6O2bNmijRs3anh4WJL02muvye/3q7CwUJJ07rnn6vnnnzf1G78AAAAAAFiVtukdGRlRTU2NQqGQxsbG1NDQoNraWnk8HgUCAVVUVKi6ulpFRUUaHR1VJBKR2+2WJG3atEmRSCS1rTPOOENDQ0N68803tXr1aj377LNat25dxiYHAAAAADi+pW16PR6P2tvbFY1GVVBQoGQyuWCs2+1ONbxHctJJJ+nv/u7v9L//9//WKaecorPPPvs94wEAAAAAOBppv9Pb1tamsrIyhcNhVVVVyTAMORwOzczMSJq9NtwwjNmNpbkjWDKZ1PPPP6/vfve72r59u4aHh3X++ecvwjQAAAAAAAvKy8vsfzks7ZneYDColpYWdXd3q7CwUPn5+fL7/WptbZXP51MgEFA4HDb1vdxly5bJ6XRq69atWrFihf7qr/5KJ5988qJMBAAAAACwgBy/2VQmpW16y8vL1dPTM+/xysrK1PN1dXWSpL6++T8sv3Xr1jl/NzY2qrGx0VaxAAAAAABY4TAOX5sMAAAAAFiSpkdezej2l59xWka3fzRy++JrAAAAAACOQtrLm3NBPB43HetyuSzHZyMHNS1+Dts1PbHLfPzmoL0cObSf7OZ4O/ZfpuNP2DB7Qzqr+zYb68PqPHJyzebg+rDzOuJY5E5Nid5nTI8pqNiYnf1k9f3Dznu5yTHZeu+3ksPu+tj81W+ajn/ipv/XVo5c2k/ZrOljLfeYHvOTlsaMvi5ycc2mcuTQvw1SNdnIsSQcx9/p5UwvAAAAAGDJOibO9AIAAAAA7HPkcabXksHBQcViMUlSLBbTwMDAvJjXX39dn/70p9XQ0KArr7xSw8PDkqSnnnpKl112mS6//HJ973vfO4rSAQAAAAB4b7aa3mg0qqGhIUlSZ2enxsfH58XcfffduvLKK9XR0aFrrrlGra2tOnTokO688061tbWpo6NDDz/8sCYmJo5uBgAAAAAALCDt5c2JRELNzc2Kx+OanJxUKBRSV1eXnE6nSktL1dvbq/7+fpWUlKi+vl5er1der1fXX3996ovf77zzjlasWKG9e/fq9NNP10knnSRJuuCCC7Rnzx5deumlmZ0lAAAAAOC4lLbpHRkZUU1NjUKhkMbGxtTQ0KDa2lp5PB4FAgFVVFSourpaRUVFGh0dVSQSkdvtTo0fHh7W9u3bde+99+qtt96acwe01atXK5FIZGZmAAAAAIBZjuP3HsZpm16Px6P29nZFo1EVFBQomUwuGOt2u+c0vLt379ZXvvIV3XXXXfJ6vZqentb+/ftTz+/fv39p3QYcAAAAAJBT0rb7bW1tKisrUzgcVlVVlQzDkMPh0MzMjCTJ4XDIMIzZjeW9u7ndu3fr9ttv1/33369zzz1XkuTz+TQyMqLf/e53mp6e1p49e3TeeedlYl4AAAAAAKQ/0xsMBtXS0qLu7m4VFhYqPz9ffr9fra2t8vl8CgQCCofDKi4unjPujjvu0KFDh3TDDTdIks4880zdeuutuuGGG/SZz3xGhmHosssu09q1azMzMwAAAADAcS9t01teXq6enp55j1dWVqaer6urkyT19fWlnn/ssceOuL1LLrlEl1xyia1iAQAAAAA2OI7f3+l1GIevTQYAAAAALEmHXn8jo9t3Fp2S0e0fjbRnenNBPB43HetyuSzHZyMHNS1+jlysyW6O47amJ3aZj98ctJQjF9dHLtZkN0cu1mQlBzWxPqjp/ctBTYufIxdrspsjF2taMvKO3zO9x+99qwEAAAAASx5NLwAAAABgybLV9A4ODioWi0mSYrGYBgYG5sW8/vrr+vSnP62GhgZdeeWVGh4eTj134MAB1dXVae/evTbLBgAAAAAgPVtNbzQa1dDQkCSps7NT4+Pj82LuvvtuXXnllero6NA111yj1tZWSdKLL76o+vp6vfrqq0dRNgAAAADALIcjL6P/5bK0N7JKJBJqbm5WPB7X5OSkQqGQurq65HQ6VVpaqt7eXvX396ukpET19fXyer3yer26/vrrU1/8fuedd7RixQpJ0vT0tO69915dd911mZ0ZAAAAAGDWcXwjq7RN78jIiGpqahQKhTQ2NqaGhgbV1tbK4/EoEAiooqJC1dXVKioq0ujoqCKRiNxud2r88PCwtm/frnvvvVeSdMEFF2RuNgAAAAAA/IG0Ta/H41F7e7ui0agKCgqUTCYXjHW73XMa3t27d+srX/mK7rrrLnm93sWpGAAAAAAAk9JefN3W1qaysjKFw2FVVVXJMAw5HA7NzMxIkhwOhwzDmN1Y3rub2717t26//Xbdf//9OvfcczNUPgAAAAAAC0t7pjcYDKqlpUXd3d0qLCxUfn6+/H6/Wltb5fP5FAgEFA6HVVxcPGfcHXfcoUOHDumGG26QJJ155pm69dZbMzMLAAAAAMCCDqxckdHtuzK69aOTtuktLy9XT0/PvMcrKytTz9fV1UmS+vr6Us8/9thj77ndjo4OS4UCAAAAAGCVwzh8bTIAAAAAYEmKx+MZ3f7hX+7JRbn9g0oAAAAAAByFtJc354L4E7tMx7o2By39vxiH/x8Jq2MyHS9JB3+z1/SYFet92dlPFnO8Hfsv0/EnbDjfUg7X5qAkaWrg16ZzrPR/MCvztprD6rGWbKzZJfA6kqyvj0zvJ0ka27ffVPzaE1dLkpITb5rOsWyNJyePRS7WJNlYH1ZfFzbWR6bXrK332eN1fZgck9M1LYH3cjs1HfrtqOkxzlP/p+X4XNxPVnLYrenQ62+YjncWnSJJmh5+2fSY5d51WfkcnuofMD1m5dl+W8cCxzbO9AIAAAAAlixbZ3oHBwe1b98+bdiwQbFYTC6XS36/f07M66+/rhtvvFHvvPOODMPQ/8/e/UfFWd75/38OMEEKWCYO1UU0yQxRiqaDsWmxLo10yZSF2jimGhJCYrc9jT2x262mJQ21ncQmTdIpp2lFzyebxVJsIo2ADW2kozVVlhS/aP1RaYMlWKoVAQ01Q0wIk5nvH9mMSzHOj2TYIfN6nMM5Ovf7uq/3Nfc1A+9c94+NGzdisVj45S9/SX19PYmJiVxxxRU4nc4JjzoSEREREREROVciqjbdbje9vb0ANDU1MTQ0NClm+/btrFixgoaGBlavXk1NTQ3Hjx/nhz/8IT/96U956KGHGB0dZf/+0E9hEBEREREREQlH0JXe0dFRqqur8Xg8jIyMYLfbaWlpwWg0kpeXR3t7O93d3eTk5FBRUYHFYsFisVBVVRU4B/7kyZMkJyczY8YMHnroIVJSUgDwer0kJ0f3eVEiIiIiIiISv4IWvf39/ZSVlWG32xkcHKSyshKHw4HZbMZms1FYWEhpaSlZWVkMDAzQ3NyMyWQKtO/r62Pr1q3U1taSkJCA2WwGTj2n95133uH666+P3uhEREREREQkrgUtes1mM/X19bjdbtLS0vB6vWeMNZlMEwrezs5ONmzYwLZt27BYLAD4fD6+//3v88orr/DjH/8Yg8FwDoYhIiIiIiIiMiw1bGMAACAASURBVFnQoreuro78/HyWL19OZ2cnTz75JAaDAZ/PB4DBYMDv9wNMuCFVZ2cnmzZtYufOnVx66aWB17/97W8zY8YM7rvvPt3ASkREREREJA74fD6cTic9PT3MmDGD7373u8yaNSuw/ec//zkPPfQQSUlJfPnLX6aoqIjDhw+zdu1ajh8/zoc+9CG+973vBS6VDUfQoreoqAin00lraysZGRkkJiaSm5tLTU0NVqsVm82Gy+UiOzt7QrvNmzczPj7OunXrAJgzZw5Lly7l4Ycf5qMf/SirVq0CYOXKlSxatCjsxEVERERERGR6ePzxxzlx4gSNjY08//zzbNmyhfvvvx+A4eFhGhoaaGpqYmxsjOXLl3P99ddz33338ZnPfIabb76ZHTt20NjYyG233RZ230GL3oKCAtra2ia9XlxcHNheXl4OQEdHR2D73r1733N/Bw+G/vBoERERERERmf6effZZCgsLAcjPz+ell14KbHvxxRe55pprmDFjBjNmzODyyy/n4MGDPPvss6xevRqAT37yk9TU1ERU9Br8p89NFhERERERkfOSx+OJ6v5PP7nnTKqrq7Hb7SxcuBCAG264gccff5ykpCR+8Ytf8PLLL/P1r38dgG984xvcdNNNfOc736G1tZULLriAV199lW984xvs3r077Nx0Ua2IiIiIiIhEVVpaGkePHg38v8/nIykp6T23HT16lPT09AmvHz16lAsvvDCivoOe3hwLwvlXifT09LDjp6IP5XTu+zgd7x0cDrmPpIsz8Ty2P/ScFhWFldPpvMKN9w6/GXJ8UqZ5SnKK5H16p+v3Ibf5wIL5UzI/wh1HTM7ZGPysjh16JeT4ZOucUznF4LGIxe+0qchp/I3BkNsYL7l4SnIafbIjSOS70hZez7Hn/xByfEr+PACO/aE7tPh5VwHnx/zoeSP03y9XXjJFv19i8Dstkpye63895DbXzMqKqI9Qf69+YMF8ILbm7Ok2Y38+FHJ88lxrTM4POTfmz5/P/v37KS0t5fnnn+eKK64IbPvIRz7CD3/4Q8bGxjhx4gSHDh3iiiuuYP78+Tz55JPcfPPNPPXUU1x77bUR9T0til4RERERERGZvhYtWkRHRwfl5eX4/X42b97MAw88wOWXX86//Mu/UFlZyfLly/H7/Xzta18jOTmZL3/5y1RVVfHzn/8ck8nED37wg4j6VtErIiIiIiIiUZWQkMDGjRsnvGa1WgP/feutt3LrrbdO2G42m/mv//qvs+87kkY9PT10dXUB0NXV9Z53ZH799de57bbbqKysZMWKFfT19QHw61//miVLlvC5z32OPXv2nEXqIiIiIiIiIu8voqLX7XbT29sLQFNTE0NDQ5Nitm/fzooVK2hoaGD16tXU1NRw8uRJfvCDH/CTn/yExsZGdu7cyeHDh89uBCIiIiIiIiJnEPT05tHRUaqrq/F4PIyMjGC322lpacFoNJKXl0d7ezvd3d3k5ORQUVGBxWLBYrFQVVUVuPD75MmTJCcnk5iYyL59+0hKSuKtt94CIDU1NbojFBERERERkbgVtOjt7++nrKwMu93O4OAglZWVOBwOzGYzNpuNwsJCSktLycrKYmBggObmZkwmU6B9X18fW7dupba29lSHSUm43W42btzIwoULA7epFhERERERETnXgp7ebDabefzxx1m7di33338/Xq/3jLEmk2lCwdvZ2cmaNWvYtm0bFosl8Lrdbuepp55ifHycRx555CyHICIiIiIiIvLegha9dXV15Ofn43K5KCkpwe/3YzAY8Pl8ABgMBvx+/6mdJby7u87OTjZt2sTOnTuZN+/U8/FGR0dZsWIFJ06cICEhgZSUlAltRERERERERM6loOcWFxUV4XQ6aW1tJSMjg8TERHJzc6mpqcFqtWKz2XC5XGRnZ09ot3nzZsbHx1m3bh0Ac+bMYePGjdx4441UVFSQlJTElVdeyWc/+9nojExERERERETiXtCit6CggLa2tkmvFxcXB7aXl5cD0NHREdi+d+/e99zf0qVLWbp0aUTJioiIiIiIiITD4D99brKIiIiIiIiclzweT1T3f/rJPbFoWtw62fPY/pBj0xcV8erIkZDjLzNdCMCJv74WcpsZl2eHNWnS09PDjofwxx12fAzmFE4fp+OjfezCyel0XuH2Mf63gZDjjZf+06mcwp2DUzE/YuhzEZhPUX6fwuljSt+nKPcRizlBBPMjBr/TYmnOBvo4X+bHNP6sTtXvo3ByOt3m9b+PhhyflZEWUU4n+l8Nuc2MWZeFHR+vc9Y7/GbI8UmZZoCw/2aZiu+0490HQ25zwVW5ER0Lmd6mRdErIiIiIiIikRtPNP5fp/B/RrdOFhERERERkfNWREVvT08PXV1dAHR1dXHw4ORTCl5//XVuu+02KisrWbFiBX19fRO233333bhcrki6FxEREREREQlJREWv2+2mt7cXgKamJoaGhibFbN++nRUrVtDQ0MDq1aupqakJbHvooYd4+eWXI0xZREREREREJDRBr+kdHR2luroaj8fDyMgIdrudlpYWjEYjeXl5tLe3093dTU5ODhUVFVgsFiwWC1VVVYELv0+ePElycjIAzz33HC+88AJLly6dtPorIiIiIiIi5148P7MnaNHb399PWVkZdrudwcFBKisrcTgcmM1mbDYbhYWFlJaWkpWVxcDAAM3NzZhMpkD7vr4+tm7dSm1tLUNDQ9x7773ce++9PProo1EdmIiIiIiIiEjQotdsNlNfX4/b7SYtLQ2v13vGWJPJNKHg7ezsZMOGDWzbtg2LxcJPf/pTRkZG+NKXvsTw8DDHjx/HYrFw8803n5vRiIiIiIiIyCS+OF7qDVr01tXVkZ+fz/Lly+ns7OTJJ5/EYDDg8/kAMBgM+P/nDUxIePcS4c7OTjZt2sTOnTu59NJLAVi5ciUrV64EoLm5mb6+PhW8IiIiIiIiEjVBi96ioiKcTietra1kZGSQmJhIbm4uNTU1WK1WbDYbLpeL7OzsCe02b97M+Pg469atA2DOnDls3LgxOqMQEREREREReQ9Bi96CggLa2tomvV5cXBzYXl5eDkBHR0dg+969e993v1rhFRERERERkWgz+P1xfHK3iIiIiIhIHBjyvBPV/X8o/QNR3f/ZCLrSKyIiIiIiItNbPK91Touid8/TL4Yce8vHP8LuA8+FHL/sE9cA4Hlsf8ht0hcV4fF4Qo9PTw87PpKchkePhRyfmZYyJTmFGw9w+J3jIcXP/MAFp3KagmMR7fc2omMRZpspmR9TMO7BI0dDir/4wtQpyynUeX56jsfi/Dgfcgqnj0hzGn9jMOR44yUXn+oj2vMjgu9ZzY/pm1Pf8EjIfVgyTVNyLH7W8fuQ4yuunx9RTu90hd7HBxbMDzteczZ2+piK+RFJHzK9JQQPEREREREREZmeVPSKiIiIiIjIeSui05t7eno4cuQICxYsoKuri/T0dHJzcyfEvP7666xfv56TJ0/i9/vZuHEjFouFBx54gIcffpiZM2cCsGHDBiwWy9mPRERERERERN6TT9f0hsftdmM2m1mwYAFNTU2UlpZOKnq3b9/OihUrKC4upr29nZqaGu699166u7vZunUrV1999TkZgIiIiIiIiMiZBC16R0dHqa6uxuPxMDIygt1up6WlBaPRSF5eHu3t7XR3d5OTk0NFRQUWiwWLxUJVVVXgwu+TJ0+SnJwMQHd3Nzt27GB4eJgbbriB1atXR3eEIiIiIiIicS6OF3qDF739/f2UlZVht9sZHByksrISh8OB2WzGZrNRWFhIaWkpWVlZDAwM0NzcjMlkCrTv6+tj69at1NbWAlBWVsby5ctJS0vjjjvuYP/+/RQVFUVvhCIiIiIiIhK3gha9ZrOZ+vp63G43aWlpeL3eM8aaTKYJBW9nZycbNmxg27ZtWCwW/H4/q1atCqwAL1y4kD/+8Y8qekVERERERCQqgt69ua6ujvz8fFwuFyUlJfj9fgwGAz6fDwCDwRB40HFCwru76+zsZNOmTezcuZN58+YBp06V/sxnPsPRo0fx+/08/fTTurZXREREREREoiboSm9RURFOp5PW1lYyMjJITEwkNzeXmpoarFYrNpsNl8tFdnb2hHabN29mfHycdevWATBnzhw2btzI1772NVauXMmMGTO47rrrWLhwYXRGJiIiIiIiIgCBhcp4FLToLSgooK2tbdLrxcXFge3l5eUAdHR0BLbv3bv3Pfd30003cdNNN0WUrIiIiIiIiEg4DP54LvlFRERERETiwKsjR6K6/8tMF0Z1/2cjouf0TrVjf+gOOTZl3lV4PJ6Q40/fVCvcNmN/PhRyfPJc65Tk5B1+M+T4pEwznsf2h77/RUUR5RTJuEMdR1Km+VROYY4jmjmdzivs9ymSYxFmm3DH4B0cDj3+4sxTOcXg/Ah3HNH8rAbip2DORvtYRHTspuBYnOj7S0jxMyyzATja8XTIfaRe//GwP0cQ/vyI9vdNODmdbjPW0xtyfPKVOTE5Z8PpY6r+Nogkp6mYH8cPvhxymwtyr4ipYxdpH7GYUzh9xGJOkfZxov/VkONnzLpsSnKS6S/ojaxEREREREREpisVvSIiIiIiInLeiuj05p6eHo4cOcKCBQvo6uoiPT2d3NzcCTGvv/4669ev5+TJk/j9fjZu3IjFYuHFF19ky5Yt+P1+MjMz+f73v09ycvI5GYyIiIiIiIhMFs+3copopdftdtPbe+r6nqamJoaGhibFbN++nRUrVtDQ0MDq1aupqanB7/dz9913873vfY/du3dTWFjI3/72t7MbgYiIiIiIiMgZBF3pHR0dpbq6Go/Hw8jICHa7nZaWFoxGI3l5ebS3t9Pd3U1OTg4VFRVYLBYsFgtVVVWBC79PnjxJcnIyr7zyChkZGdTX1/Pyyy+zcOFCLBZL1AcpIiIiIiISz3xxvNIbtOjt7++nrKwMu93O4OAglZWVOBwOzGYzNpuNwsJCSktLycrKYmBggObmZkwmU6B9X18fW7dupba2lpGREZ577jnuvvtuZs2axe23387VV1/NddddF9VBioiIiIiISHwKWvSazWbq6+txu92kpaXh9XrPGGsymSYUvJ2dnWzYsIFt27ZhsVg4dOgQs2bNIicnB4DCwkJeeuklFb0iIiIiIiISFUGv6a2rqyM/Px+Xy0VJSQl+vx+DwYDP5wPAYDAELopOSHh3d52dnWzatImdO3cyb948AC677DKOHj1Kf38/AM888wxz584954MSERERERGRd/l8/qj+xLKgK71FRUU4nU5aW1vJyMggMTGR3NxcampqsFqt2Gw2XC4X2dnZE9pt3ryZ8fFx1q1bB8CcOXPYuHEjmzZt4q677sLv93PNNddwww03RGVgIiIiIiIiIkGL3oKCAtra2ia9XlxcHNheXl4OQEdHR2D73r1733N/1113HQ8//HBEyYqIiIiIiEj44vg+Vhj88fzAJhERERERkThwaGgkqvu3fsgUPOj/SNCV3ljg8XhCjk1PTw87fir6UE7nvo9YzCnSPpTTue/jvMrpsf2hxy8qiqyPGHqfpqIP5aT5oZz+7/pQTue+j1jMKdI+YjEnmf6C3shKREREREREZLqaFiu9IiIiIiIiErl4vqo1oqK3p6eHI0eOsGDBArq6ukhPTyc3N3dCzOuvv8769es5efIkfr+fjRs3kp6ezp133hmI+dOf/sRdd93FsmXLzm4UIiIiIiIiIu8hotOb3W43vb29ADQ1NTE0NDQpZvv27axYsYKGhgZWr15NTU0NmZmZNDQ00NDQwJ133kleXh633nrr2Y1ARERERERE3pcPf1R/YlnQld7R0VGqq6vxeDyMjIxgt9tpaWnBaDSSl5dHe3s73d3d5OTkUFFRgcViwWKxUFVVFbjw++TJkyQnJwf26ff7ueeee3C5XCQmJkZvdCIiIiIiIhLXgha9/f39lJWVYbfbGRwcpLKyEofDgdlsxmazUVhYSGlpKVlZWQwMDNDc3IzJ9O7tqvv6+ti6dSu1tbWB15544gnmzp2LxWKJzqhERERERERECKHoNZvN1NfX43a7SUtLw+v1njHWZDJNKHg7OzvZsGED27Ztm1Dg7t27l5UrV55l6iIiIiIiIiLvL2jRW1dXR35+PsuXL6ezs5Mnn3wSg8GAz+cDwGAwBO4ElpDw7iXCnZ2dbNq0iZ07d3LppZdO2Gd3dzfz588/l+MQERERERGRM9Ddm99HUVERTqeT1tZWMjIySExMJDc3l5qaGqxWKzabDZfLRXZ29oR2mzdvZnx8nHXr1gEwZ84cNm7cyOHDh0lNTcVgMERnRCIiIiIiIiL/I2jRW1BQQFtb26TXi4uLA9vLy8sB6OjoCGzfu3fve+5v5syZ/OIXv4goWREREREREZFwGPzxvM4tIiIiIiISB3reeDOq+7/yEnNU9382gq70ioiIiIiIyPQWz2ud06Lo9Xg8Icemp6eHHQ8w/sZgyG2Ml1w8JTlFu49YzCmcPmIxp0j7UE7nvo+4z+mx/aG3WVQUU+/TVPShnPT9oZzO3Id3OPTVoKRM85TkdOIvfw25zYzZlzN26JWQ45OtczRnY6iPWPydJ9PftCh6RUREREREJHK++F3oJSF4iIiIiIiIiMj0FFHR29PTQ1dXFwBdXV0cPHhwUszrr7/ObbfdRmVlJStWrKCvrw84dVdnh8PBkiVL2LVr11mkLiIiIiIiIvL+Iip63W43vb29ADQ1NTE0NDQpZvv27axYsYKGhgZWr15NTU0NANu2beOBBx5g9+7dPPDAA7z99ttnkb6IiIiIiIgE4/f7o/oTy4Je0zs6Okp1dTUej4eRkRHsdjstLS0YjUby8vJob2+nu7ubnJwcKioqsFgsWCwWqqqqAhd+nzx5kuTkZACuvPJKPB4PSUlJ+P1+DAZDdEcoIiIiIiIicSto0dvf309ZWRl2u53BwUEqKytxOByYzWZsNhuFhYWUlpaSlZXFwMAAzc3NmEymQPu+vj62bt1KbW0tAHPnzmXJkiWkpKSwaNEiLrzwwuiNTkRERERERGJ+NTaagp7ebDabefzxx1m7di33338/Xq/3jLEmk2lCwdvZ2cmaNWvYtm0bFouFgwcP8tvf/pbf/OY3PPHEExw+fJhHH3303IxERERERERE5B8ELXrr6urIz8/H5XJRUlISOCXZ5/MBYDAYAv9qkJDw7u46OzvZtGkTO3fuZN68ecCp51xdcMEFJCcnk5iYyMyZMzly5Eg0xiUiIiIiIiIS/PTmoqIinE4nra2tZGRkkJiYSG5uLjU1NVitVmw2Gy6Xi+zs7AntNm/ezPj4OOvWrQNgzpw5bNy4kaVLl7J8+XKMRiOXX345DocjOiMTERERERGRuBe06C0oKKCtrW3S68XFxYHt5eXlAHR0dAS279279z33t2zZMpYtWxZRsiIiIiIiIhI+Xxxf02vwx/MVzSIiIiIiInHgxVffiOr+P3LZJVHd/9kIutIbC7zDb4Ycm5Rp5q2jx0OOvyj1glN9DA6H3sfFmXg8npDj09PTw44HGH9jMOQ2xksunpKcBo8cDbnNxRemcvid0I/FzA+cOhah5nU6p6kYd7T7UE7nvg/lFGYfj+0PPX5RUczOj+HRYyHFZ6alhBV/uk00vzcvvjAVCP93XjS/Z0HfH8opeJup+BxFe9zHnv9DyPEp+fOmJKdw+ojl+RHt+QSE/bd/JH2cD+J5pTfojaxEREREREREpisVvSIiIiIiInLeiqjo7enpoaurC4Curi4OHjw4Keb111/ntttuo7KykhUrVtDX1wfAI488wo033sjy5cvZs2fPWaQuIiIiIiIi8v4iKnrdbje9vb0ANDU1MTQ0NClm+/btrFixgoaGBlavXk1NTQ2HDx9m+/btNDQ08OCDD9La2sprr712diMQERERERGR9+X3+6P6E8uC3shqdHSU6upqPB4PIyMj2O12WlpaMBqN5OXl0d7eTnd3Nzk5OVRUVGCxWLBYLFRVVQUueD958iTJycm89tpr5ObmkpGRAcC8efN44YUXJj3jV0RERERERORcCFr09vf3U1ZWht1uZ3BwkMrKShwOB2azGZvNRmFhIaWlpWRlZTEwMEBzczMmkynQvq+vj61bt1JbW8tFF11Eb28vb775Jqmpqfzud79j9uzZ0RyfiIiIiIhI3IvnuzcHLXrNZjP19fW43W7S0tLwer1njDWZTBMK3s7OTjZs2MC2bduwWCwAfPOb3+QrX/kKl1xyCVddddWEeBEREREREZFzKeg1vXV1deTn5+NyuSgpKcHv92MwGPD5fAAYDIbAOdwJCe/urrOzk02bNrFz507mzTv1PDOv18sLL7zAz372M7Zu3UpfXx/z58+PxrhEREREREREgq/0FhUV4XQ6aW1tJSMjg8TERHJzc6mpqcFqtWKz2XC5XJOuy928eTPj4+OsW7cOgDlz5rBx40aMRiM333wzycnJfP7zn2fmzJnRGZmIiIiIiIjEvaBFb0FBAW1tbZNeLy4uDmwvLy8HoKOjI7B9796977m/O+64gzvuuCOiZEVERERERCR8cXxJLwZ/rN9fWkRERERERM7KM6/8Lar7/+icS6O6/7MRdKVXREREREREprd4XuucFkWvx+MJOTY9PT3s+KnoQzmd+z4izumx/aHHLyqKrI8Yep8i7eNE319Cjp9hmX0qpzDf26mYH2OHXgm5j2TrnNicszE4PyL5HOlYxE5OJ/pfDbnNjFmXTc37FO73RyTf5SG2marv/nD6iHR+lGz+fyHHt61fHVEfsfQ+TWVODlddyG1a1v5bVD8XsThnA33E0N8GgZwi6EOmt6B3bxYRERERERGZriJa6e3p6eHIkSMsWLCArq4u0tPTyc3NnRAzPDzM2rVrGR8fJzMzky1btpCSksITTzxBbW0tSUlJLFmyhFtvvfWcDERERERERETemy8GT28+fvw4X//613nrrbdITU1l69atk57us3XrVn7/+9/j9XpZunQpt956K3//+9/59Kc/zRVXXAGcusnyqlWrzthPRCu9breb3t5eAJqamhgaGpoUs2PHDhwOB7t27SInJ4fGxkbGx8f53ve+R11dHQ0NDTQ2NjI8PBxJCiIiIiIiIjKN7d69myuuuIJdu3Zx0003cd99903Y3tnZyV//+lcaGxvZvXs3//mf/8nbb7/NH//4Rz7zmc/Q0NBAQ0PD+xa8EMJK7+joKNXV1Xg8HkZGRrDb7bS0tGA0GsnLy6O9vZ3u7m5ycnKoqKjAYrFgsVhYv349fr8fn8/HwMAAs2fP5tChQ1x++eV88IMfBODaa6/lmWee4V//9V/P4q0SERERERGR9xOLN7J69tln+eIXvwjAJz/5yUlF7zXXXMOHP/zhwP+fPHmSpKQkXnrpJbq7u1mxYgUzZ87kW9/6Fh/60IfO2E/Qore/v5+ysjLsdjuDg4NUVlbicDgwm83YbDYKCwspLS0lKyuLgYEBmpubMZlMAHi9XhYvXszY2Bhr1qxhYGBgwsXgqampjI6OhvfOiIiIiIiIyLSyZ88e6uvrJ7x20UUXBerD1NTUSTcZS05OJjk5mfHxcdatW8fSpUtJTU3FYrFw9dVX84lPfIK9e/fy3e9+lx/96Edn7Dto0Ws2m6mvr8ftdpOWlobX6z1jrMlkChS8AEajkX379nHgwAGqqqr41re+xdGjRwPbjx49qjuiiYiIiIiInOduueUWbrnllgmv3XHHHYH68OjRo1x44YWT2r399tv8+7//Ox/72MdYvfrUXeYLCgpISUkBYNGiRe9b8EII1/TW1dWRn5+Py+WipKQEv9+PwWDA5/MBYDAYAkvlCQnv7s7pdNLZ2QmcqtoNBgNWq5X+/n7+/ve/c+LECZ555hmuueaaYCmIiIiIiIjIeWb+/Pk8+eSTADz11FNce+21E7YfP36c2267jSVLlrBmzZrA69/61rf49a9/DcDvfvc7rrrqqvftJ+hKb1FREU6nk9bWVjIyMkhMTCQ3N5eamhqsVis2mw2Xy0V2dvaEdpWVlTidTmpra0lISMDpdGI0Glm3bh1f+MIX8Pv9LFmyhIsvvji0d0REREREREQiEoOX9LJs2TKqqqpYtmwZRqORH/zgBwBs27aNkpISfv/73/Pqq6+yZ88e9uzZA8DmzZu56667WL9+Pbt37yYlJYXvfve779tP0KK3oKCAtra2Sa8XFxcHtpeXlwPQ0dER2G61WmloaJjU7lOf+hSf+tSngnUrIiIiIiIi57GUlJT3PDX5G9/4BgAf+chHuO22296z7XvVmmdi8MfibbxERERERETknOl4uT+q+7/+illR3f/ZiOg5vSIiIiIiIiLTQdDTm2PBP966+v2kp6eHHT8VfSinc9/H6Xjv4HDIfSRdnInnsf2h57SoKKycTucVbrx3+M2Q45MyzVOSUyTv0ztdvw+5zQcWzJ+S+RHuOGJyzsbgZ3Xs0Cshxydb55zKKQaPRSx+p01FTtGeg5HkNNp+IOQ2aYWf4Njzfwg5PiV/HgDH/tAdWvy8UzdEOR/mR88bof9+ufKSKfr9EoPfaZHk9PxfB0Juk3/5P0XUR6i/Vz+wYD4QW3P2dJuxnt6Q45OvzInJ+SHTn1Z6RURERERE5Lw1LVZ6RUREREREJHLxfCuniFZ6e3p66OrqAqCrq4uDBw9OihkeHmbVqlUsX76cr371qxw7diyw7dixY5SXl3Po0KEI0xYREREREREJLqKi1+1209t76vz8pqYmhoaGJsXs2LEDh8PBrl27yMnJobGxEYA//OEPVFRU8Oqrr55F2iIiIiIiIiLBBT29eXR0lOrqajweDyMjI9jtdlpaWjAajeTl5dHe3k53dzc5OTlUVFRgsViwWCysX78ev9+Pz+djYGCA2bNnA3DixAlqa2sDz14StEhWgQAAIABJREFUERERERERiZagRW9/fz9lZWXY7XYGBweprKzE4XBgNpux2WwUFhZSWlpKVlYWAwMDNDc3YzKZAPB6vSxevJixsTHWrFkDwLXXXhvdEYmIiIiIiMgEvji+pjdo0Ws2m6mvr8ftdpOWlobX6z1jrMlkChS8AEajkX379nHgwAGqqqp48MEHz03WIiIiIiIiIiEIek1vXV0d+fn5uFwuSkpK8Pv9GAwGfD4fAAaDIXAnsISEd3fndDrp7OwEIDU1FYPBEI38RUREREREJAif3x/Vn1gWdKW3qKgIp9NJa2srGRkZJCYmkpubS01NDVarFZvNhsvlIjs7e0K7yspKnE4ntbW1JCQk4HQ6ozUGERERERERkfcUtOgtKCigra1t0uvFxcWB7eXl5QB0dHQEtlutVhoaGs643/fbJiIiIiIiInIuGPzx/JRiERERERGROLD/j4eiuv+iPGtU9382gq70xgKPxxNybHp6etjxU9GHcjr3fcRiTpH2oZzOfR/K6TyYH4/tDz1+UVFYfcTisYjFnCLtIxZzCqePKc0pxHke7hw/q5zCbOMdfjPk+KRMc1Q/26dzisX36byZs3F6LGR6mxZFr4iIiIiIiEQunk/wDXr3ZhEREREREZHpKqKV3p6eHo4cOcKCBQvo6uoiPT2d3NzcCTHDw8OsXbuW8fFxMjMz2bJlCykpKfzyl7+kvr6exMRErrjiCpxO54RHHYmIiIiIiIicKxFVm263m97eXgCampoYGhqaFLNjxw4cDge7du0iJyeHxsZGjh8/zg9/+EN++tOf8tBDDzE6Osr+/aFf1yEiIiIiIiISjqArvaOjo1RXV+PxeBgZGcFut9PS0oLRaCQvL4/29na6u7vJycmhoqICi8WCxWJh/fr1+P1+fD4fAwMDzJ49mxkzZvDQQw+RkpICgNfrJTk5OeqDFBERERERiWe++L2kN3jR29/fT1lZGXa7ncHBQSorK3E4HJjNZmw2G4WFhZSWlpKVlcXAwADNzc2YTCbgVFG7ePFixsbGWLNmDQkJCZjNZuDUc3rfeecdrr/++uiOUEREREREROJW0KLXbDZTX1+P2+0mLS0Nr9d7xliTyRQoeAGMRiP79u3jwIEDVFVV8eCDD+Lz+fj+97/PK6+8wo9//GMMBsO5GYmIiIiIiIi8J929+X3U1dWRn5+Py+WipKQEv9+PwWDA5/MBYDAYAm/g/74hldPppLOzE4DU1NRAcfvtb3+bsbEx7rvvvsBpziIiIiIiIiLREHSlt6ioCKfTSWtrKxkZGSQmJpKbm0tNTQ1WqxWbzYbL5SI7O3tCu8rKSpxOJ7W1tSQkJOB0Ounu7ubhhx/mox/9KKtWrQJg5cqVLFq0KDqjExERERERkbgWtOgtKCigra1t0uvFxcWB7eXl5QB0dHQEtlutVhoaGia1O3jwYMTJioiIiIiIiITD4I/nk7tFRERERETigPvFl6O6f/tHrojq/s9GRM/pFREREREREZkOgp7eHAs8Hk/Isenp6WHHR9LH2KFXQo5Pts6Zkpw8j+0PPX5RESf6Xw05fsasy07lFGYfEY07xD7SFxUBcOyFl0LuI8V2dVRzOp1XuH288/89G3L8Bz52LUDYx29Kjt1UfFbDnB/R/hyF08fpMZzo+0vIfcywzJ6SYzE8eizk+My0lPNnfoT7uYhkfkR7zk7BsZiS79kYGvdU/W0QUU5T8F0e7e+oWJwfsZhTOH1EmlMkfwue+MtfQ28z+/Ip+T08+tv/DrlN2g3/HNGxOB/4iN8TfLXSKyIiIiIiIuetiIrenp4eurq6AOjq6nrPm1MNDw+zatUqli9fzle/+lWOHTu1ivDrX/+aJUuW8LnPfY49e/acReoiIiIiIiIi7y+iotftdtPb2wtAU1MTQ0NDk2J27NiBw+Fg165d5OTk0NjYyMmTJ/nBD37AT37yExobG9m5cyeHDx8+uxGIiIiIiIiInEHQa3pHR0eprq7G4/EwMjKC3W6npaUFo9FIXl4e7e3tdHd3k5OTQ0VFBRaLBYvFwvr16/H7/fh8PgYGBpg9ezaJiYns27ePpKQk3nrrLQBSU1OjPkgREREREZF4Fs8P7Qla9Pb391NWVobdbmdwcJDKykocDgdmsxmbzUZhYSGlpaVkZWUxMDBAc3MzJpMJAK/Xy+LFixkbG2PNmjWnOkxKwu12s3HjRhYuXEhS0rS4l5aIiIiIiIhMQ0FPbzabzTz++OOsXbuW+++/H6/Xe8ZYk8kUKHgBjEYj+/bt45577qGqqirwut1u56mnnmJ8fJxHHnnkLIcgIiIiIiIi78fnj+5PLAta9NbV1ZGfn4/L5aKkpAS/34/BYMDn8wFgMBgCS+UJCe/uzul00tnZCZw6hdlgMDA6OsqKFSs4ceIECQkJpKSkTGgjIiIiIiIici4FPbe4qKgIp9NJa2srGRkZJCYmkpubS01NDVarFZvNhsvlIjs7e0K7yspKnE4ntbW1JCQk4HQ6SUtL48Ybb6SiooKkpCSuvPJKPvvZz0ZtcCIiIiIiIhLfgha9BQUFtLW1TXq9uLg4sL28vByAjo6OwHar1UpDQ8OkdkuXLmXp0qURJywiIiIiIiLh8cX6OchRZPDH8228RERERERE4kDr7/8U1f3fOP/DUd3/2dCtk0VERERERM5z8bzWOS2K3vHX/hZyrDH7UjweT8jx6enpAGG3iXZ8RDk9tj/0+EVFkeUUQ32kLyqKuZwi7SPc/UOMzo+p+FyEOz9i6FicHsPRjqdD7iP1+o9PybEYHj0WcnxmWsr5Mz/C/VxEMj+iPWen4Fic+MtfQ46fMfvy2J0fYX5WYzKn8+C7XDmd+z4i/mz3vxpy/IxZlwGE3WZKfg9PwedCpjfdOllERERERETOWyp6RURERERE5LwV0enNPT09HDlyhAULFtDV1UV6ejq5ubkTYoaHh1m7di3j4+NkZmayZcsWUlJSAtvvvvtuPvjBD7J27dqzG4GIiIiIiIi8r3i+pjeilV63201vby8ATU1NDA0NTYrZsWMHDoeDXbt2kZOTQ2NjY2DbQw89xMsvvxxhyiIiIiIiIiKhCbrSOzo6SnV1NR6Ph5GREex2Oy0tLRiNRvLy8mhvb6e7u5ucnBwqKiqwWCxYLBbWr1+P3+/H5/MxMDDA7NmzAXjuued44YUXWLp0KX19fdEen4iIiIiISNzzEb8rvUGL3v7+fsrKyrDb7QwODlJZWYnD4cBsNmOz2SgsLKS0tJSsrCwGBgZobm7GZDIB4PV6Wbx4MWNjY6xZs4ahoSHuvfde7r33Xh599NGoD05ERERERETiW9Ci12w2U19fj9vtJi0tDa/Xe8ZYk8kUKHgBjEYj+/bt48CBA1RVVWG32xkZGeFLX/oSw8PDHD9+HIvFws0333xuRiMiIiIiIiLyvwQteuvq6sjPz2f58uV0dnby5JNPYjAY8Pl8ABgMhsBF0QkJ714i7HQ6KSkpoaCggNTUVAwGAytXrmTlypUANDc309fXp4JXREREREREoiZo0VtUVITT6aS1tZWMjAwSExPJzc2lpqYGq9WKzWbD5XKRnZ09oV1lZSVOp5Pa2loSEhJwOp3RGoOIiIiIiIi8j3i+e3PQoregoIC2trZJrxcXFwe2l5eXA9DR0RHYbrVaaWhoOON+tcIrIiIiIiIi0Wbwx3PJLyIiIiIiEgeau/4Q1f3fvGBeVPd/NoKu9MYCz2P7Q45NX1SEx+MJPT49/VQfYbaJdjzA8YOhP8v4gtwrpuZ9CrOP0d/+d8jxaTf8c1h9pC8qAuD4n3pC7uOCD185JeMOt4+xPx8KOT55rvVUTuHOwfPgcwThz49ov08A428MhhRvvORiALzDb4bcR1KmOSaPRSzmBBHMj3A/FxHMj2jP2dEnO4IH/o+0hddH1Md5Mz9CbBPTOZ0H3+WR5DT++hshtzFmXRJ2fCy+T+H0EWlO4b5PACf6/hJymxmW2VPye/h498GQ21xwVW5Ex0Kmt2lR9IqIiIiIiEjk4vn83oTgISIiIiIiIiLTU0QrvT09PRw5coQFCxbQ1dVFeno6ubm5E2KGh4dZu3Yt4+PjZGZmsmXLFlJSUnjggQd4+OGHmTlzJgAbNmzAYrGc/UhERERERETkPfnieKk3opVet9tNb28vAE1NTQwNDU2K2bFjBw6Hg127dpGTk0NjYyMA3d3dbN26lYaGBhoaGlTwioiIiIiISNQEXekdHR2luroaj8fDyMgIdrudlpYWjEYjeXl5tLe3093dTU5ODhUVFVgsFiwWC+vXr8fv9+Pz+RgYGGD27NnAqaJ3x44dDA8Pc8MNN7B69epoj1FERERERETiVNCit7+/n7KyMux2O4ODg1RWVuJwODCbzdhsNgoLCyktLSUrK4uBgQGam5sxmUwAeL1eFi9ezNjYGGvWrAGgrKyM5cuXk5aWxh133MH+/fspKiqK7ihFRERERETiWDw/qTZo0Ws2m6mvr8ftdpOWlobX6z1jrMlkChS8AEajkX379nHgwAGqqqpoaGhg1apVgVt/L1y4kD/+8Y8qekVERERERCQqgl7TW1dXR35+Pi6Xi5KSEvx+PwaDAZ/PB4DBYAj8q0FCwru7czqddHZ2ApCamorBYGB0dJTPfOYzHD16FL/fz9NPP83VV18djXGJiIiIiIjI//D7/VH9iWVBV3qLiopwOp20traSkZFBYmIiubm51NTUYLVasdlsuFwusrOzJ7SrrKzE6XRSW1tLQkICTqeT9PR0vva1r7Fy5UpmzJjBddddx8KFC6M2OBEREREREYlvQYvegoIC2traJr1eXFwc2F5eXg5AR0dHYLvVaqWhoWFSu5tuuombbrop4oRFREREREREQmXwx/patIiIiIiIiJyVxt89H9X9L70uP6r7PxtBV3pFRERERERkevPF8VrntCh6/7vnLyHH/vOVs/F4PCHHn76TdLhtoh0fUU6P7Q89flFRZDmF20eY8eH0cTr+RN9fQu5jhiXC+RHl9zbcMUCMzo+p+FyEOT+i/T5NSU5TcCwGjxwNOf7iC1PPn/kRg99pUzJnz4ffeVHsI6ZzOg++yyPJaezPh0JukzzXGnZ8LL5P4fQRaU7jfxsIOd546T8B0f+7K5LvtNH2AyG3SSv8RETHQqa3aVH0ioiIiIiISOTieaU36COLRERERERERKariFZ6e3p6OHLkCAsWLKCrq4v09HRyc3MnxAwPD7N27VrGx8fJzMxky5YtpKSk8OKLL7Jlyxb8fj+ZmZl8//vfJzk5+ZwMRkREREREROR/i2il1+1209vbC0BTUxNDQ0OTYnbs2IHD4WDXrl3k5OTQ2NiI3+/n7rvv5nvf+x67d++msLCQv/3tb2c3AhEREREREZEzCLrSOzo6SnV1NR6Ph5GREex2Oy0tLRiNRvLy8mhvb6e7u5ucnBwqKiqwWCxYLBbWr1+P3+/H5/MxMDDA7NmzeeWVV8jIyKC+vp6XX36ZhQsXYrFYpmKcIiIiIiIicSuen1QbtOjt7++nrKwMu93O4OAglZWVOBwOzGYzNpuNwsJCSktLycrKYmBggObmZkwmEwBer5fFixczNjbGmjVreOONN3juuee4++67mTVrFrfffjtXX3011113XdQHKiIiIiIiIvEnaNFrNpupr6/H7XaTlpaG1+s9Y6zJZAoUvABGo5F9+/Zx4MABqqqq2LBhA7NmzSInJweAwsJCXnrpJRW9IiIiIiIiUeSL34Xe4Nf01tXVkZ+fj8vloqSkBL/fj8FgwOfzAWAwGAJL5QkJ7+7O6XTS2dkJQGpqKgaDgcsuu4yjR4/S398PwDPPPMPcuXPP+aBEREREREREIISV3qKiIpxOJ62trWRkZJCYmEhubi41NTVYrVZsNhsul4vs7OwJ7SorK3E6ndTW1pKQkIDT6WTGjBls2rSJu+66C7/fzzXXXMMNN9wQrbGJiIiIiIhInAta9BYUFNDW1jbp9eLi4sD28vJyADo6OgLbrVYrDQ0Nk9pdd911PPzwwxEnLCIiIiIiIhIqgz+eb+MlIiIiIiISB+qfeiaq+1/1yY9Gdf9nI+hKbyzweDwhx6anp4cdPxV9xGJO3uE3Q45PyjRPSU7h9DFV79NU9BHJsQi3zfnwPoXTRyznFO1jF05Op9vEYk7xOj88j+0PvY9FRTF5LGIxp3D6iPec4vH3i3dwOOT4pIszpySncPqIxTkbaR+xmNP5Ip7XOoPeyEpERERERERkulLRKyIiIiIiIuetiE5v7unp4ciRIyxYsICuri7S09PJzc2dEDM8PMzatWsZHx8nMzOTLVu2MDo6yp133hmI+dOf/sRdd93FsmXLzm4UIiIiIiIiIu8hopVet9tNb28vAE1NTQwNDU2K2bFjBw6Hg127dpGTk0NjYyOZmZk0NDTQ0NDAnXfeSV5eHrfeeuvZjUBERERERETel8/vj+pPLAu60js6Okp1dTUej4eRkRHsdjstLS0YjUby8vJob2+nu7ubnJwcKioqsFgsWCwW1q9fj9/vx+fzMTAwwOzZswP79Pv93HPPPbhcLhITE6M5PhEREREREYljQYve/v5+ysrKsNvtDA4OUllZicPhwGw2Y7PZKCwspLS0lKysLAYGBmhubsZkMgHg9XpZvHgxY2NjrFmzJrDPJ554grlz52KxWKI3MhEREREREQEgxhdjoypo0Ws2m6mvr8ftdpOWlobX6z1jrMlkChS8AEajkX379nHgwAGqqqp48MEHAdi7dy8rV648B+mLiIiIiIiInFnQa3rr6urIz8/H5XJRUlKC3+/HYDDg8/kAMBgMgWc+JSS8uzun00lnZycAqampGAyGwLbu7m7mz59/TgciIiIiIiIi8o+CrvQWFRXhdDppbW0lIyODxMREcnNzqampwWq1YrPZcLlcZGdnT2hXWVmJ0+mktraWhIQEnE4nAIcPH55UBIuIiIiIiIhEQ9Cit6CggLa2tkmvFxcXB7aXl5cD0NHREdhutVppaGiY1G7mzJn84he/iDhhERERERERCY8/Bi/qPX78OF//+td56623SE1NZevWrcycOXNCzO23387f//53jEYjycnJ7Ny5k/7+ftatW4fBYGDu3Ll85zvfmXDW8T8y+GNx9CIiIiIiInLO7Hzi6aju/4uf+njYbR544AFGR0f5yle+wq9+9Suee+45vvWtb02IKS0t5Ve/+tWEM4Vvv/12Pv/5z/Pxj3+cb3/72xQWFrJo0aIz9hN0pTcWdPW9FnLsAks2jb97PuT4pdflA+B5bH/IbdIXFeHxeEKPT08POz6SnI49/4eQ41Py501JTuHGA4z+9r9Dik+74Z9P5TQFxyLa721ExyLMNlMyP6Zg3O90/T6k+A8smB9ZThHM2VDbBOJjcH4cfud4yPEzP3BBzM6PUNtEmtPokx3BA/9H2sLrT/UR7fkRyZyN8rGI1/kxFTmN/flQyH0kz7VOybHY++wfQ47/7LV5U/P3x2+eDD3+XxZqzsZQHxHlNAXH+3wQi8/SffbZZ/niF78IwCc/+Unuu+++CdvffPNNjhw5wu23386RI0f40pe+RFFREd3d3XzsYx8LtOvo6Jj+Ra+IiIiIiIhMX3v27KG+vn7CaxdddFHgHxZSU1Mn/YPE+Pg4//Zv/8bKlSt5++23WbZsGR/5yEcCN1c+U7t/pKJXREREREREouqWW27hlltumfDaHXfcwdGjRwE4evQoF1544YTtZrOZ8vJykpKSuOiii/jwhz/MK6+8MuH63fdq94+CPrLovfT09NDV1QVAV1cXBw8enBQzPDzMqlWrWL58OV/96lc5duwYcOoZvQ6HgyVLlrBr165IuhcREREREZFpbv78+Tz55KnT05966imuvfbaCdsPHDjAf/zHfwCnits///nPWCwW8vLyePrppwPtPvrRj75vPxEVvW63m97eXgCampoYGhqaFLNjxw4cDge7du0iJyeHxsZGALZt28YDDzzA7t27eeCBB3j77bcjSUFERERERERC5Pf7o/oTiWXLlvHnP/+ZZcuW0djYyB133AGcqhlffPFFFi5cyKxZs7j11lv5whe+wJ133snMmTOpqqrixz/+MUuXLmV8fJxPf/rT79tP0NObR0dHqa6uxuPxMDIygt1up6WlBaPRSF5eHu3t7XR3d5OTk0NFRQUWiwWLxcL69evx+/34fD4GBgaYPXs2AFdeeSUej4ekpKQJ52KLiIiIiIhI/EhJSeFHP/rRpNe/8Y1vBP67urp60vY5c+bw4IMPhtxP0KK3v7+fsrIy7HY7g4ODVFZW4nA4MJvN2Gw2CgsLKS0tJSsri4GBAZqbmzGZTAB4vV4WL17M2NgYa9asAWDu3LksWbKElJQUFi1aFPT8axERERERETk7sXj35qkS9PRms9nM448/ztq1a7n//vvxer1njDWZTIGCF8BoNLJv3z7uueceqqqqOHjwIL/97W/5zW9+wxNPPMHhw4d59NFHz81IRERERERERP5B0KK3rq6O/Px8XC4XJSUlgVOSfT4fAAaDIXAO9/++i5bT6aSzsxM4dRtpg8FAeno6F1xwAcnJySQmJjJz5kyOHDkSjXGJiIiIiIiIBD+9uaioCKfTSWtrKxkZGSQmJpKbm0tNTQ1WqxWbzYbL5SI7O3tCu8rKSpxOJ7W1tSQkJOB0Orn00ktZunQpy5cvx2g0cvnll+NwOKI2OBEREREREYlvQYvegoIC2traJr1eXFwc2F5eXg5AR0dHYLvVaqWhoWFSu2XLlrFs2bKIExYREREREZHwxPM1vQZ/pPeXFhERERERkWmh1t0RPOgsrLFfH9X9n42IntMrIiIiIiIiMh0EPb05Fng8npBj09PTw46fij6U07nvI9KcjncfDDn+gqtyI+ojlt6nSPsYf2Mw5HjjJRcDMNp+IOQ2aYWfmJL5Ee44YnHOxuL8ONH/asjxM2ZdBoB3cDjkNkkXZ8bdsZjSnB7bH3qbRUUxmdPRjqdDjk+9/uMAHO18JrT4go+eyuk8mB99wyMhx1syTRH1EUvv01Tm1Nn715DbFORcHlEfof5eTSv8BBBbc/Z0m7E/Hwo5PnmuNSbnx/kink/w1UqviIiIiIiInLemxUqviIiIiIiIRC6OF3ojW+nt6emhq6sLgK6uLg4enHy66PDwMKtWrWL58uV89atf5dixYwA88sgj3HjjjSxfvpw9e/acReoiIiIiIiIi7y+iotftdtPb2wtAU1MTQ0NDk2J27NiBw+Fg165d5OTk0NjYyOHDh9m+fTsNDQ08+OCDtLa28tprr53dCERERERERETOIOjpzaOjo1RXV+PxeBgZGcFut9PS0oLRaCQvL4/29na6u7vJycmhoqICi8WCxWJh/fr1+P1+fD4fAwMDzJ49m9dee43c3FwyMjIAmDdvHi+88ALZ2dlRH6iIiIiIiIjEn6BFb39/P2VlZdjtdgYHB6msrMThcGA2m7HZbBQWFlJaWkpWVhYDAwM0NzdjMp26+5/X62Xx4sWMjY2xZs0a0tLS6O3t5c033yQ1NZXf/e53zJ49O9pjFBERERERiWu+OL6oN2jRazabqa+vx+12k5aWhtfrPWOsyWQKFLwARqORffv2ceDAAaqqqnjwwQf55je/yVe+8hUuueQSrrrqqgnxIiIiIiIiIudS0Gt66+rqyM/Px+VyUVJSgt/vx2Aw4PP5ADAYDIFnPiUkvLs7p9NJZ2cnAKmpqRgMBrxeLy+88AI/+9nP2Lp1K319fcyfPz8a4xIREfn/2bv/sLjqM+/jn4GQiIGWYYfQIvHHgHGWpBnUUqfJhXFamGSHdFviL4SMWWO7ddu0z5Mm2+5mSMM2tL1skXRj0X1ai6W0ja0CNmiFkRojF5t0J9q6dtLgRi1pXTqJlSSD1Sgyzx9ZR1kSZ+aEmQ7wfl0Xf+Sc+3zv+3vmOyR3zpkzAADgf4TD4YT+pLKoV3qdTqcaGhrU3d2tnJwcpaeny2azqbm5WUVFRbLb7Wpqapr0uVyPx6OGhga1tLQoLS1NDQ0NmjNnjjIyMrRmzRrNmzdPt9xyi3JzcxM2OQAAAADA7Ba16XU4HOrp6Zm0vaKiIrK/pqZGkjQwMBDZX1RUpPb29knHbdiwQRs2bDBcMAAAAAAAsTKFU/1aNAAAAADgnHzr508kdPz/6746oeOfi6hXelNBKBSKOTY7OzvueEM5Ht0Te3ylMzk1JXgOyagpnhxvxZ/6r+dizjHv0qKUfC1eO/RszPHn2RadrinBazAVz1M8OVK5plODh2POMe+y4pR8LVKxJkl69enfxBSfaV8iSXr9d0dizjH34gtn3ZqN5JgBf+fFkyNZ5ykpNRk5T7Pw75dUrCmeHKlYk9Ec8f67Lhk1zRSz+Vpn1AdZAQAAAAAwXU2LK70AAAAAAONm8/f0GrrSOzg4KL/fL0ny+/06dOjQWWP9fr9WrFgR+fNjjz2ma6+9VjfeeKN++tOfGkkPAAAAAEBMDDW9Pp9Phw+f/kxaR0eHjh49esa44eFhtba2amxsTJL0xhtv6Otf/7paW1vV3t6un/zkJzp27JjB0gEAAAAAeHdRb28eHR2V1+tVKBTSyMiIXC6Xurq6lJGRoZKSEvX39ysQCKi4uFh1dXWyWq2yWq3avHmztm3bpu3bt2vNmjWSpOeee04XXnih3vve90qSrrzySh04cEB/8zd/k9hZAgAAAABmpahN79DQkKqqquRyuRQMBuXxeFRdXS2LxSK73a7y8nK53W4VFBRoeHhYnZ2dMpvN8nq9Wr9+vfLz8yNjjY6OTngC2vz58zU6OpqYmQEAAAAAJEmz9xO9MTS9FotFbW1t8vl8ysrKityqfCZms1lms1nBYFAHDhzQkSNH1NLSohMnTmjjxo369Kc/rVdeeSUS/8orr8yox4ADAAAAQCqazQ/VmqwVAAAgAElEQVSyitr0tra2qrS0VLW1tdq/f7/27t0rk8mk8fFxSZLJZIp851Na2umPCOfn56u3tzcyxvLly7Vjxw698cYbGhoa0vHjx3X++efrwIEDuvXWWxMxLwAAAAAAoje9TqdTDQ0N6u7uVk5OjtLT02Wz2dTc3KyioiLZ7XY1NTWpsLAwarKMjAz90z/9k2699VaFw2Fde+21E25/BgAAAABgKkVteh0Oh3p6eiZtr6ioiOyvqamRJA0MDJxxjHdu/8hHPqKPfOQjhooFAAAAACAepnB4Ft/cDQAAAACzwO27H0vo+F/629S9sGnoe3oBAAAAAJgOot7enApCoVDMsdnZ2XHHJyMHNU19jlSsyWgOapr6HNTE+qCmv1yO0KN7Yo+vdCalpnhypOJrkYo1Gc1BTVOfIxVrMpojFWuaKcbHZ+8NvlzpBQAAAADMWIaa3sHBQfn9fkmS3+/XoUOHzhrr9/u1YsWKCdteffVV1dTU6LnnnjOSHgAAAACAmBhqen0+nw4fPixJ6ujo0NGjR88YNzw8rNbWVo2NjUW2PfPMM6qrq9Pvf/97I6kBAAAAAIhZ1M/0jo6Oyuv1KhQKaWRkRC6XS11dXcrIyFBJSYn6+/sVCARUXFysuro6Wa1WWa1Wbd68Wdu2bdP27du1Zs2ayHivv/66Wlpa9MUvfjGhEwMAAAAAnDabv7QnatM7NDSkqqoquVwuBYNBeTweVVdXy2KxyG63q7y8XG63WwUFBRoeHlZnZ6fMZrO8Xq/Wr1+v/Pz8CeNdeeWVCZsMAAAAAADvFLXptVgsamtrk8/nU1ZW1oRblf83s9kss9msYDCoAwcO6MiRI2ppadGJEye0ceNG7dixY0qLBwAAAABEN86V3rNrbW1VaWmpamtrtX//fu3du1cmk0nj4+OSJJPJFLlUnpZ2+iPC+fn56u3tjYyxfPlyGl4AAAAAQNJFbXqdTqcaGhrU3d2tnJwcpaeny2azqbm5WUVFRbLb7WpqalJhYWEy6gUAAAAAIGZRm16Hw6Genp5J2ysqKiL7a2pqJEkDAwNnHONM29vb2+MqFAAAAACAeJnCs/kxXgAAAAAwC2zvfDSh429dU5nQ8c+Foe/pBQAAAABgOoh6e3MqCIVCMcdmZ2fHHZ+MHNQ09TkM1/TontjjK53GcqTQeTKaY3TvmT+ucCZZK5afrinOc5uM9RHvPFJyzabg+jDyPhrt//eYj8kqXzbrXouk1pTg92pSajLyuzzGY5L1uz+eHEbXx/2//M+Y46+/aqmhHKl0npJZU8MDvVEi39Zw3cqEvi9Scc1GcqTQvw0iNRnIMRPM5ht8udILAAAAAJixaHoBAAAAADOWoaZ3cHBQfr9fkuT3+3Xo0KGzxvr9fq1YsSLy54ceekjXX3+9ampq9OUvfznyfb8AAAAAgMQYD4cT+pPKDDW9Pp9Phw8fliR1dHTo6NGjZ4wbHh5Wa2urxsbGJEmvvfaavvWtb+kHP/iB7rvvPo2OjmrPntjv8wcAAAAAIB5RH2Q1Ojoqr9erUCikkZERuVwudXV1KSMjQyUlJerv71cgEFBxcbHq6upktVpltVq1efNmbdu2Tdu3b9eaNWskSXPnztV9992nzMxMSdLY2JjmzZuX2BkCAAAAAGatqE3v0NCQqqqq5HK5FAwG5fF4VF1dLYvFIrvdrvLycrndbhUUFGh4eFidnZ0ym83yer1av3698vPzI2OlpaXJYrFIktrb2/XnP/9Zy5cvT9zsAAAAAACzWtSm12KxqK2tTT6fT1lZWZFblc/EbDbLbDYrGAzqwIEDOnLkiFpaWnTixAlt3LhRO3bs0Pj4uL75zW/qhRde0J133imTyTSlEwIAAAAATDSbv7IoatPb2tqq0tJS1dbWav/+/dq7d69MJlPkAVQmkylyAtPSTn9EOD8/X729b3932fLly7Vjxw5J0pe//GXNnTtXd911VyQeAAAAAIBEiNr0Op1ONTQ0qLu7Wzk5OUpPT5fNZlNzc7OKiopkt9vV1NSkwsLCqMkCgYAeeOABffCDH9S6deskSTfffLMqKyvPfSYAAAAAgDNK9ScsJ1LUptfhcKinp2fS9oqKisj+mpoaSdLAwMAZx3hr++LFi9/1640AAAAAAJhKpvBsvrkbAAAAAGaBrT99JKHjb7/hbxI6/rmIeqU3FYRCoZhjs7Oz445PRg5qmvochmt6NPbvhs6udBrLkULnyWiOV3/9TMzxmaUfOF1TnOc2Gesj3nmk5JpNwfVh5H3Ea5E6Nb0y8MuYj5m//KrknKd4f38Y+V0e4zHJ+t0fTw6j66Ni+10xx/dt/YyhHKl0npJZ0zUN3475mMcbNiT0fZGKazaSI4X+bRCpyUAOTG/ToukFAAAAABg3m+/v5fHJAAAAAIAZy1DTOzg4KL/fL0ny+/3v+nAqv9+vFStWRP7c29ura6+9Vtddd53uv/9+I+kBAAAAAHEIh8MJ/Ullhppen8+nw4cPS5I6Ojp09OjRM8YNDw+rtbVVY2NjkqQ333xTd9xxh77//e/rJz/5ie655x69/PLLBksHAAAAAODdRf1M7+joqLxer0KhkEZGRuRyudTV1aWMjAyVlJSov79fgUBAxcXFqqurk9VqldVq1ebNm7Vt2zZt375da9askSSlp6fr5z//uebMmaM//elPkqT58+cndoYAAAAAgFkratM7NDSkqqoquVwuBYNBeTweVVdXy2KxyG63q7y8XG63WwUFBRoeHlZnZ6fMZrO8Xq/Wr1+v/Pz8iQnnzJHP59NXvvIVrVixQnPm8CwtAAAAAEBiRL292WKxqK+vT5s3b9bdd98duVX5TMxms8xms4LBoA4cOKCWlhZ5PB6dOHFCGzdujMS5XC498cQTeuONN/Tggw9OzUwAAAAAAGc0Hg4n9CeVRb3M2traqtLSUtXW1mr//v3au3evTCaTxsfHJUkmkynyweW0tNM9dH5+vnp7eyNjLF++XDt27NDo6Khuu+02tba2au7cucrMzIwcAwAAAADAVIva9DqdTjU0NKi7u1s5OTlKT0+XzWZTc3OzioqKZLfb1dTUpMLCwqjJsrKy9LGPfUx1dXWaM2eOLrvsMv3t3/7tlEwEAAAAAHBmqX41NpGiNr0Oh0M9PT2TtldUVET219TUSJIGBgbOOMY7t99444268cYbDRULAAAAAEA8TOFU/1IlAAAAAMA5+dKPH0ro+LfXrk7o+OeCRycDAAAAwAw3m691ToumNxQKxRybnZ0dd3wyclDT1OcwXNOje2KPr3Qay5FC58lojlefCcQcn/mBxadrivPcJmN9xDuPlFyzKbg+jLyPeC1Sp6ZX9h+I+Zj5jg8m5zzF+/vDyO/yGI9J1u/+eHIYXR8f/cpdMcf/4sufMZQjlc5TMmty/ktLzMfs2fbZhL4vUnHNRnKk0L8NIjUZyIHpbVo0vQAAAAAA42bzlV5D3xc0ODgov98vSfL7/Tp06NBZY/1+v1asWDFp+9atW9XU1GQkPQAAAAAAMTHU9Pp8Ph0+fFiS1NHRoaNHj54xbnh4WK2trRobG5uw/b777tOzzz5rJDUAAAAAADGLenvz6OiovF6vQqGQRkZG5HK51NXVpYyMDJWUlKi/v1+BQEDFxcWqq6uT1WqV1WrV5s2btW3bNm3fvl1r1qyJjPerX/1KTz/9tG688UY9//zzCZ0cAAAAAGB2i9r0Dg0NqaqqSi6XS8FgUB6PR9XV1bJYLLLb7SovL5fb7VZBQYGGh4fV2dkps9ksr9er9evXKz8/PzLW0aNH9e1vf1vf/va39cgjjyR0YgAAAACA08Zn70d6oze9FotFbW1t8vl8ysrKmnSr8juZzWaZzWYFg0EdOHBAR44cUUtLi06cOKGNGzfq8ssv18jIiP7+7/9ex44d02uvvSar1TrhSjAAAAAAAFMlatPb2tqq0tJS1dbWav/+/dq7d69MJpPGx8clSSaTKfIksLS00x8Rzs/PV29vb2SM5cuXa8eOHZKkm2++WZLU2dmp559/noYXAAAAABJsNj+9OWrT63Q61dDQoO7ubuXk5Cg9PV02m03Nzc0qKiqS3W5XU1OTCgsLk1EvAAAAAAAxi9r0OhwO9fT0TNpeUVER2V9TUyNJGhgYOOMYZ9rOFV4AAAAAQKKZwrP5OjcAAAAAzAJf+MHPEjp+880fT+j45yLqlV4AAAAAwPQ2m691ToumNxQKxRybnZ0dd3wyclDT1OdIxZqM5qCmqc9BTawPavrL5UjFmuLJEYl/dE/sOSqdrA9q+ovlMFxTnGvcUI4UOk/nkgPT27RoegEAAAAAxo3P4iu9aX/pAgAAAAAASBRDTe/g4KD8fr8kye/369ChQ2eN9fv9WrFiReTP9957r6qqquTxeOTxePT8888bKQEAAAAAgKgM3d7s8/lksVhUVlamjo4Oud1u2Wy2SXHDw8NqbW3V2NhYZFsgENDtt9+uJUuWGK8aAAAAAIAYRG16R0dH5fV6FQqFNDIyIpfLpa6uLmVkZKikpET9/f0KBAIqLi5WXV2drFarrFarNm/erG3btmn79u0TvpM3EAjoO9/5jo4dO6ZrrrlGn/70pxM6QQAAAACY7VLx6c2vvfaa/vEf/1F/+tOfNH/+fN1+++3Kzc2N7H/iiSf03e9+V9Lp+p988kk99NBDeu2113Tbbbfp4osvliTddNNNcrvdZ80TtekdGhpSVVWVXC6XgsGgPB6PqqurZbFYZLfbVV5eLrfbrYKCAg0PD6uzs1Nms1ler1fr169Xfn7+hPGqqqpUW1urrKwsbdiwQXv27JHT6TRyjgAAAAAA09SuXbu0aNEife5zn9PDDz+su+66S/X19ZH9V199ta6++mpJ0j333KMrrrhCRUVFuv/++3XLLbdo/fr1MeWJ+plei8Wivr4+bd68WXffffeEW5X/N7PZLLPZrGAwqAMHDqilpUUej0cnTpzQxo0bFQ6HtW7dOuXm5mru3LlasWKFDh48GFOhAAAAAABjxsOJ/THiySefVHl5uaTTDe6+ffvOGPfHP/5RP/vZz7RhwwZJ0m9+8xs9/vjjqqur05YtWzQ6OvqueaJe6W1tbVVpaalqa2u1f/9+7d27VyaTSePj45Ikk8kUuVSelna6h87Pz1dvb29kjOXLl2vHjh0KhUJavXq1fv7zn+v888/XL3/5S1177bXRSgAAAAAATGP333+/2traJmz7q7/6q8h3Ic+fP/+s36F877336u/+7u80d+5cSdLSpUt1/fXXa8mSJbr77rvV0tKiL33pS2fNHbXpdTqdamhoUHd3t3JycpSeni6bzabm5mYVFRXJbrerqalJhYWFUSeanZ2tjRs36uabb9bcuXP14Q9/eMKTnQEAAAAAM8/111+v66+/fsK2DRs26JVXXpEkvfLKK3rPe94z6bjx8XE9/vjj2rhxY2RbZWVlJLayslLbt29/19xRm16Hw6Genp5J2ysqKiL7a2pqJEkDAwNnHOOd2z/xiU/oE5/4RLS0AAAAAIApMh4e/0uXMMkVV1yhvXv3aunSpXriiSd05ZVXTop59tlndckll+i8886LbLv11lu1detWLV26VPv27dPixYvfNY8pnIqP8QIAAAAATJnPfO+BhI5/163XxX3Mq6++qi996Us6duyYMjIydMcddygvL0/f+MY3tGrVKi1dulSPPPKInnrqKXm93shxgUBA27dvV0ZGhiwWi7Zv366srKyz5pkWTe/Z7u0+k+zs7LjjDeV4dE/s8ZXO5NSU4Dkko6Z4crwV/9pvB2POcd5fX5aSr0W8c5CU8DWYiucpnhypXNOp/3ou5hzzLi1KydciFWuSpFPPvRBT/LyiS5JW03Res5EcM+DvvHhyJOs8JaUmI+dpFv79koo1xZMjFWsymmPs2Esxx8/JsySlppkiFZveZIn69GYAAAAAAKarqJ/pBQAAAABMb6l/f2/iGLrSOzg4KL/fL0ny+/06dOjQWWP9fv+EJzT/53/+p2pra3XTTTfp85//vE6dOmWkBAAAAAAAojLU9Pp8Ph0+fFiS1NHRoaNHj54xbnh4WK2trRobG5MkhcNhbd26VV//+te1a9culZeX68UXXzRYOgAAAAAgFuFwOKE/qSzq7c2jo6Pyer0KhUIaGRmRy+VSV1eXMjIyVFJSov7+fgUCARUXF6uurk5Wq1VWq1WbN2/Wtm3btH37dq1Zs0aS9MILLygnJ0dtbW169tlntWLFClmt1oRPEgAAAAAwO0VteoeGhlRVVSWXy6VgMCiPx6Pq6mpZLBbZ7XaVl5fL7XaroKBAw8PD6uzslNlsltfr1fr165Wfnx8Za2RkRL/61a+0detWXXTRRbrtttu0ZMkSffjDH07oJAEAAAAAs1PUptdisaitrU0+n09ZWVmRW5XPxGw2y2w2KxgM6sCBAzpy5IhaWlp04sQJbdy4URs2bNBFF12k4uJiSVJ5ebl+85vf0PQCAAAAABIiatPb2tqq0tJS1dbWav/+/dq7d69MJpPGx8clSSaTKXIPd1ra6Y8I5+fnq7e3NzLG8uXLtWPHDr3++ut65ZVXNDQ0pIsuukgHDhzQddel7vc5AQAAAMBMMJ7in7tNpKhNr9PpVENDg7q7u5WTk6P09HTZbDY1NzerqKhIdrtdTU1NKiwsjJps7ty5+upXv6pNmzYpHA7r8ssv1zXXXDMV8wAAAAAAYJKoTa/D4VBPT8+k7RUVFZH9NTU1kqSBgYEzjvHO7R/+8If1wAMPGCoWAAAAABC/VH/CciKZwrN59gAAAAAwC3zq//0koeN/99M3JnT8c2Hoe3oBAAAAAJgOot7enApCoVDMsdnZ2XHHJyMHNU19jlSsyWgOapr6HNQ0A9bHo3tij690xpUjFV+LVKzJaI5UrCmeHEmtKcZ1Hu8aP6eaEr0+EvjeNlxTCp6neHKkYk1Gc6RiTZj+pkXTCwAAAAAwbjZ/qpXbmwEAAAAAM5ahpndwcFB+v1+S5Pf7dejQobPG+v1+rVixQpJ07NgxeTyeyM8HP/hB7dq1y0gJAAAAAIAYjYcT+5PKDDW9Pp9Phw8fliR1dHTo6NGjZ4wbHh5Wa2urxsbGJEl5eXlqb29Xe3u7vvCFL6ikpEQ33HCDwdIBAAAAAHh3UT/TOzo6Kq/Xq1AopJGREblcLnV1dSkjI0MlJSXq7+9XIBBQcXGx6urqZLVaZbVatXnzZm3btk3bt2/XmjVrJowZDoe1fft2NTU1KT09PWGTAwAAAADMblGb3qGhIVVVVcnlcikYDMrj8ai6uloWi0V2u13l5eVyu90qKCjQ8PCwOjs7ZTab5fV6tX79euXn508a87HHHtOll14qq9WakEkBAAAAAN42mx9kFbXptVgsamtrk8/nU1ZWVuRW5TMxm80ym80KBoM6cOCAjhw5opaWFp04cUIbN27Ujh07JEm7d+/WzTffPHWzAAAAAADgDKI2va2trSotLVVtba3279+vvXv3ymQyaXx8XJJkMpki/2uQlnb6I8L5+fnq7e2NjLF8+fJIwytJgUBAV1xxxZROBAAAAABwZuPiSu9ZOZ1ONTQ0qLu7Wzk5OUpPT5fNZlNzc7OKiopkt9vV1NSkwsLCmBK+/PLLmj9/vkwm0zkXDwAAAADAu4na9DocDvX09EzaXlFREdlfU1MjSRoYGDjjGO/cnpubq5/97GeGigUAAAAAIB6m8Gz+RDMAAAAAzALr7vpxQsdv+0xtQsc/F1Gv9KaC0KN7Yo7NrnQqFArFHp+dfTpHnMckOl6SXjv0bMzHnGdblJzzFGeOU//1XMzx8y4tiitHdqVTkvTabwdjznHeX1+WlHnHm8PQeYp3Dc6A95EU//pI9HmSpOePjcQUb80zS5LGjr0Uc445eZaUfC1SsSbJwPqI931hYH0kes0m5ffHTFkfMR6T0jXNgN/l1DT1OYzW9MYfgzHHZ7zv9DeyvH7kDzEfM/fCwqT8PfzGH16M+ZiMwgsMvRYzwWy+1pn2ly4AAAAAAIBEmRZXegEAAAAAxo2Pc6U3LoODg/L7/ZIkv9+vQ4cOnTXW7/drxYoVkT/v3r1b1dXVuvbaa/XjHyf2vnIAAAAAwOxmqOn1+Xw6fPiwJKmjo0NHjx49Y9zw8LBaW1s1NjYW2faNb3xD9957r3bt2qV7771XJ06cMFICAAAAAABRRb29eXR0VF6vV6FQSCMjI3K5XOrq6lJGRoZKSkrU39+vQCCg4uJi1dXVyWq1ymq1avPmzdq2bZu2b9+uNWvWRMa77LLTDxKaM2eOwuEw39cLAAAAAEiYqE3v0NCQqqqq5HK5FAwG5fF4VF1dLYvFIrvdrvLycrndbhUUFGh4eFidnZ0ym83yer1av3698vPzJ4x36aWX6tprr1VmZqYqKyv1nve8J2GTAwAAAADw9OZ3ZbFY1NfXp82bN+vuu++ecKvy/2Y2m2U2mxUMBnXgwAG1tLTI4/HoxIkT2rhxow4dOqTHH39cv/jFL/TYY4/p5Zdf1iOPPDKlEwIAAAAA4C1Rr/S2traqtLRUtbW12r9/v/bu3SuTyaTx8XFJkslkivyvQVra6R46Pz9fvb29kTGWL1+uHTt26MUXX9R5552nefPmKT09Xbm5uTp58mQi5gUAAAAA+B+z+OHN0Ztep9OphoYGdXd3KycnR+np6bLZbGpublZRUZHsdruamppUWFgYNdkFF1ygG2+8UbW1tcrIyNCFF16o6urqKZkIAAAAAAD/W9Sm1+FwqKenZ9L2ioqKyP6amhpJ0sDAwBnHeOf2m266STfddJOhYgEAAAAAiIcpPJs/0QwAAAAAs8BN/9qe0PF3/R9PQsc/F1Gv9KaCUCgUc2x2dnbc8cnIQU1Tn8NoTf+86+GY479+U5WhHKl0nozmGPzjSzHHX/Y+iySpsvHumI95tP4fkrI+4p1HKq7ZVFwfq772/2KO79nyaUm8FqlU0/PHRmI+xppnTkpN9//yP2M+5vqrlqpi+10xx/dt/Ywk6aNfie2YX3z5dPyMWB+P7ok9vtJpLEcKnaek1hTnuTWSY+VX/y2m+F7vbadrSsHX4hPf/F7M8Q/+460puT5mitl8rTPq05sBAAAAAJiupsWVXgAAAACAcWFxpTcug4OD8vv9kiS/369Dhw6dNdbv92vFihWRPz/44IP62Mc+ptraWt1///1G0gMAAAAAEBNDTa/P59Phw4clSR0dHTp69OgZ44aHh9Xa2qqxsTFJ0ssvv6x//dd/VXt7u374wx+qu7tbf/jDHwyWDgAAAADAu4t6e/Po6Ki8Xq9CoZBGRkbkcrnU1dWljIwMlZSUqL+/X4FAQMXFxaqrq5PVapXVatXmzZu1bds2bd++XWvWrJEk/eEPf5DNZlNOTo4k6QMf+ICefvrpmL7jFwAAAABgzPgsfpBV1KZ3aGhIVVVVcrlcCgaD8ng8qq6ulsVikd1uV3l5udxutwoKCjQ8PKzOzk6ZzWZ5vV6tX79e+fn5kbEuuugiHT58WC+99JLmz5+vffv26eKLL07k/AAAAAAAs1jUptdisaitrU0+n09ZWVmRW5XPxGw2y2w2KxgM6sCBAzpy5IhaWlp04sQJbdy4UTt27NA///M/63Of+5ze9773afHixTKbzVM6IQAAAAAA3hK16W1tbVVpaalqa2u1f/9+7d27VyaTSePj45Ikk8kU+c6ntLTTHxHOz89Xb29vZIzly5drx44dGhsb09NPP60f/ehHGhsb0y233KKNGzcmYl4AAAAAAERvep1OpxoaGtTd3a2cnBylp6fLZrOpublZRUVFstvtampqiulzuXPmzFFGRobWrFmjefPm6ZZbblFubu6UTAQAAAAAcGZhPtN7dg6HQz09PZO2V1RURPbX1NRIkgYGBs44xju3b9iwQRs2bDBULAAAAAAA8TCFZ3PLDwAAAACzwJo77k3o+J2bbkno+OfC0Pf0AgAAAAAwHUS9vTkVhEKhmGOzs7PjjjeU49E9scdXOpNTU4LnkIya4snxVvxrvx2MOcd5f31ZSr4WrwUOxRx/3mLb6ZoSvAZT8TzFkyOVa3rt0LMx5zjPtshQTWPHXor5mDl5lpQ6T+eS4/XnfxdT/FzrxUmrKd71kejXLp6a3jpmJvydF0+OZJ2npNRk5Dyl4N8v8b4vUvJ9lIprNgXnHe9rl4yaMP1xpRcAAAAAMGNNiyu9AAAAAADjZvOjnAw1vYODgzp58qTKysrk9/uVnZ0tm802Ieb48eNauXKlFi1aJOn0057XrVunn/70p7rvvvs0Z84c/cM//IOcTue5zwIAAAAAgDMw1PT6fD5ZLBaVlZWpo6NDbrd7UtN78OBBrV69Wlu3bo1sO3bsmNrb29XR0aFTp06ptrZWy5cv19y5c89tFgAAAACAs+JK77sYHR2V1+tVKBTSyMiIXC6Xurq6lJGRoZKSEvX39ysQCKi4uFh1dXWyWq2yWq3Ky8tTIBDQ2rVrlZubq/r6ej3zzDO6/PLLNXfuXM2dO1cXXnihDh06pKVLlyZjrgAAAACAWSZq0zs0NKSqqiq5XC4Fg0F5PB5VV1fLYrHIbrervLxcbrdbBQUFGh4eVmdnp8xms/r6+rRkyRItW7ZMu3fvVmNjoz760Y9OeALa/PnzNTo6mtAJAgAAAABmr6hNr8ViUVtbm3w+n7KysjQ2NnbWWLPZLLPZLElyOBzKzMyUJFVWVmrnzp36+Mc/rldeeSUS/8orr/AYcAAAAABAwkT9yqLW1laVlpaqqalJq1atUjgclslk0vj4uCTJZDJF7g9PS3t7uPr6evX29kqS9u3bp8WLF2vp0qV68sknderUKYVCIT333HORB10BAAAAABJjPBxO6E8qi3ql1+l0qqGhQd3d3crJyVF6erpsNpuam5tVVFQku92upqYmFRYWTjhu06ZN2rJli3bt2qXMzO0RXJwAACAASURBVEw1NjYqLy9PHo9HtbW1CofD2rhxo+bNm5ewyQEAAAAAZreoTa/D4VBPT8+k7RUVFZH9NTU1kqSBgYHI/oULF6q9vX3ScTfccINuuOEGwwUDAAAAAOKT6ldjE8kUns3PrgYAAACAWWD17d9N6PgPfelTCR3/XBj6nt5kOzb6asyxeVmZCj26J+b47EqnJOmahm/HfMzjDRsUCoViz5GdHXe8JH3mew/EfMxdt16XlJriPWb3kwdjjv/bK0viymG0pmTMm5r+8jmM1vTfx2N/onxBTpYk6UcDT8UUX7f8CkM1JeO1iPv3bBJqGjv2Uszxc/IsceWYSWs2kb9n38qRSucpGTmoKc4ccf67KxXXRyrWFE+OVKzprWPGgsdijp+Tn5eyrwWmt2nR9AIAAAAAjJvNN/hGfXozAAAAAADTlaGmd3BwUH6/X5Lk9/t16NChSTHHjx/XVVddJY/HI4/Ho7a2tsi+l19+WS6XS6dOnTJYNgAAAAAgVuFwYn9SmaHbm30+nywWi8rKytTR0SG32y2bzTYh5uDBg1q9erW2bt06YXt/f7/uuOMOvfRS7J/VAgAAAADAiKhN7+joqLxer0KhkEZGRuRyudTV1aWMjAyVlJSov79fgUBAxcXFqqurk9VqldVqVV5engKBgNauXavc3FzV19drwYIFSktL07333qtrr702GfMDAAAAAMxiUZveoaEhVVVVyeVyKRgMyuPxqLq6WhaLRXa7XeXl5XK73SooKNDw8LA6OztlNpvV19enJUuWaNmyZdq9e7caGxu1c+dOLV++PBnzAgAAAAAgetNrsVjU1tYmn8+nrKwsjY2NnTXWbDbLbDZLkhwOhzIzMyVJlZWV2rlz5xSVDAAAAACIx3iqf/A2gaI+yKq1tVWlpaVqamrSqlWrFA6HZTKZND4+LkkymUyRx1+npb09XH19vXp7eyVJ+/bt0+LFixNRPwAAAAAAZxX1Sq/T6VRDQ4O6u7uVk5Oj9PR02Ww2NTc3q6ioSHa7XU1NTSosLJxw3KZNm7Rlyxbt2rVLmZmZamxsTNgkAAAAAABnN5u/pzdq0+twONTT0zNpe0VFRWR/TU2NJGlgYCCyf+HChWpvbz/ruI899ljcxQIAAAAAEA9TeDa3/AAAAAAwC6z86r8ldPxe720JHf9cGPqe3mQLhUIxx2ZnZ8cdn4wc1DT1OVKxJqM5qGnqc1DTDFgfj+6JPb7SGVeOVHwtUrEmozlSsaZ4ciS1phjXebxr/JxqSvT6SOB723BNKXie4smRijUZzZGKNWH6mxZNLwAAAADAOJ7eDAAAAADADGToSu/g4KBOnjypsrIy+f1+ZWdny2azTYg5fvy4Vq5cqUWLFkk6/eCrdevW6fvf/74efvhhSdKKFSu0YcOGc5wCAAAAAODdzOZHORlqen0+nywWi8rKytTR0SG32z2p6T148KBWr16trVu3Rrb9/ve/1+7du3X//ffLZDKptrZWFRUVk44FAAAAAGAqRG16R0dH5fV6FQqFNDIyIpfLpa6uLmVkZKikpET9/f0KBAIqLi5WXV2drFarrFar8vLyFAgEtHbtWuXm5qq+vl7ve9/7dM899yg9PV2SNDY2pnnz5iV8kgAAAACA2Slq0zs0NKSqqiq5XC4Fg0F5PB5VV1fLYrHIbrervLxcbrdbBQUFGh4eVmdnp8xms/r6+rRkyRItW7ZMu3fvVmNjo3bu3Knc3FyFw2F94xvfUElJiS655JJkzBMAAAAAMAtFbXotFova2trk8/mUlZWlsbGxs8aazWaZzWZJksPhUGZmpiSpsrJSO3fulCSdOnVKW7Zs0fz587Vt27apmAMAAAAA4F3M4o/0Rn96c2trq0pLS9XU1KRVq1YpHA7LZDJpfHxckmQymSIfik5Le3u4+vp69fb2SpL27dunxYsXKxwO6zOf+Ywuu+wyfeUrX4nc5gwAAAAAQCJEvdLrdDrV0NCg7u5u5eTkKD09XTabTc3NzSoqKpLdbldTU5MKCwsnHLdp0yZt2bJFu3btUmZmphobG9XX16f/+I//0Ouvv67+/n5J0he+8AVdfvnliZkdAAAAAGBWf09v1KbX4XCop6dn0vaKiorI/pqaGknSwMBAZP/ChQvV3t4+4ZjKyko988wz51QwAAAAAACxMoVn8xc2AQAAAMAs4PyXloSOv2fbZxM6/rkw9D29AAAAAIDpYzZf65wWTW8oFIo5Njs7O+74ZORIRk2vH/lDzPFzLyw0VNOp516I+Zh5RZco9Oie2HNUOiVJr//uSEzxcy++UFJy1keiz+1MWR/JmPcbLw7HFJ9xwfsN1fT60O9jjp970UJJ0p/9T8UUf37ZFYZqMrRmY3wfSaffS6m4ZhOZw2hNb/z3H2OOzyh4nyTp1H89F1P8vEuLDNVk5PfsTPg7LxXXx0ypaezYSzEfMyfPkpzzFOc6jzueNRtzjlefCcR8TOYHFqfm+jCQA9Nb1Kc3AwAAAAAwXdH0AgAAAABmLEO3Nw8ODurkyZMqKyuT3+9Xdna2bDbbhJjjx49r5cqVWrRokaTTT3tet26dfvSjH6mzs1Mmk0mf/exn5XQ6z30WAAAAAICzerxhw1+6hL8YQ02vz+eTxWJRWVmZOjo65Ha7JzW9Bw8e1OrVq7V169bItpdfflk//vGP9eCDD+rUqVOqqqrSNddcI5PJdG6zAAAAAADgDKI2vaOjo/J6vQqFQhoZGZHL5VJXV5cyMjJUUlKi/v5+BQIBFRcXq66uTlarVVarVXl5eQoEAlq7dq1yc3NVX1+vBQsW6Gc/+5nmzJmjF198Ue95z3toeAEAAAAACRO16R0aGlJVVZVcLpeCwaA8Ho+qq6tlsVhkt9tVXl4ut9utgoICDQ8Pq7OzU2azWX19fVqyZImWLVum3bt3q7GxUTt37tScOXP0wx/+UHfeeac8Hk8y5ggAAAAAmKWiPsjKYrGor69Pmzdv1t13362xsbGzxprNZpnNZkmSw+HQVVddJUmqrKzUwYMHI3Fr165Vf3+//H6/9u/ff65zAAAAAADgjKI2va2trSotLVVTU5NWrVqlcDgsk8mk8fFxSZLJZIp80XFa2tvD1dfXq7e3V5K0b98+LV68WM8//7w2bNigcDisjIwMzZ07d8IxAAAAAABMpai3NzudTjU0NKi7u1s5OTlKT0+XzWZTc3OzioqKZLfb1dTUpMLCwgnHbdq0SVu2bNGuXbuUmZmpxsZGLViwQDabTTfeeKNMJpPKy8v1oQ99KGGTAwAAAADMblGbXofDoZ6enknbKyoqIvtramokSQMDA5H9CxcuVHt7+6TjNmzYoA0bZu/jsgEAAAAAyWMKv3VvMgAAAAAAM4yh7+lNtlAoFHNsdnZ23PHJyJGMml4/8oeY4+deWGioplPPvRDzMfOKLlHo0T2x56h0SpJe/92RmOLnXnyhpOSsj0Sf25myPpIx7zdeHI4pPuOC9xuq6fWh38ccP/eihZKkP/ufiin+/LIrDNWUlPdqCq7ZROYwWtMb//3HmOMzCt4nSTr1X8/FFD/v0iJDNYV+sTf2+I+uMJYjhV67ZOSY7TWNHXsp5mPm5FmSc57i/PdE3PGs2ZhzvPpMIOZjMj+wODXXh4EcmN54ihQAAAAAYMai6QUAAAAAzFiGmt7BwUH5/X5Jkt/v16FDhybFHD9+XFdddZU8Ho88Ho/a2toi+8bHx/XJT35Su3btMlg2AAAAAADRGfpMr8/nk8ViUVlZmTo6OuR2u2Wz2SbEHDx4UKtXr9bWrVsnHf+tb31LJ06cMFYxAAAAAAAxitr0jo6Oyuv1KhQKaWRkRC6XS11dXcrIyFBJSYn6+/sVCARUXFysuro6Wa1WWa1W5eXlKRAIaO3atcrNzVV9fb0WLFignp4emUwmXX311cmYHwAAAABgFova9A4NDamqqkoul0vBYFAej0fV1dWyWCyy2+0qLy+X2+1WQUGBhoeH1dnZKbPZrL6+Pi1ZskTLli3T7t271djYqA0bNuihhx7Szp071dLSkoz5AQAAAABmsahNr8ViUVtbm3w+n7KysjQ2NnbWWLPZLLPZLElyOBzKzMyUJFVWVmrnzp168MEHFQwGtW7dOr344ovKyMjQBRdcwFVfAAAAAEBCRG16W1tbVVpaqtraWu3fv1979+6VyWTS+Pi4JMlkMikcDkuS0tLefi5WfX29XC6X3G639u3bp8WLF+uLX/xiZP+dd94pi8VCwwsAAAAASJioTa/T6VRDQ4O6u7uVk5Oj9PR02Ww2NTc3q6ioSHa7XU1NTSosLJxw3KZNm7Rlyxbt2rVLmZmZamxsTNgkAAAAAAA4k6hNr8PhUE9Pz6TtFRUVkf01NTWSpIGBgcj+hQsXqr29/azjfu5zn4u7WAAAAAAA4mEKv3VvMgAAAAAAM0xa9BAAAAAAAKanqLc3p4JQKBRzbHZ2dtzxychBTVOfIxVrMpqDmqY+BzXFd8yrv34m5vjM0g+wPhJY01jwWMw55uTnpeRrkYo1xZODmmbf+gg9uif2+EpnUmqKJwfrI/E5ML1xpRcAAAAAMGMZutI7ODiokydPqqysTH6/X9nZ2bLZbBNijh8/rpUrV2rRokWSTj/4at26dWpsbNRTTz2l+fPnS5Luuusu/gcFAAAAAJAQhppen88ni8WisrIydXR0yO12T2p6Dx48qNWrV2vr1q0TtgcCAd1zzz3Kzc01XjUAAAAAADGI2vSOjo7K6/UqFAppZGRELpdLXV1dysjIUElJifr7+xUIBFRcXKy6ujpZrVZZrVbl5eUpEAho7dq1ys3NVX19vSwWi4aGhvTlL39ZL730kq677jpdd911yZgnAAAAAGAWitr0Dg0NqaqqSi6XS8FgUB6PR9XV1bJYLLLb7SovL5fb7VZBQYGGh4fV2dkps9msvr4+LVmyRMuWLdPu3bvV2Nior33ta1q7dq1uueUWvfnmm7r55pu1ZMmSSVeJAQAAAACYClGbXovFora2Nvl8PmVlZWlsbOyssWazWWazWZLkcDiUmZkpSaqsrNTOnTuVmZmpm2++ObLd4XDo0KFDNL0AAAAAgISI+vTm1tZWlZaWqqmpSatWrVI4HJbJZNL4+LgkyWQyKRwOnx4s7e3h6uvr1dvbK0nat2+fFi9erN/97neqra3Vm2++qTfeeENPPfWUFi9enIh5AQAAAAAQ/Uqv0+lUQ0ODuru7lZOTo/T0dNlsNjU3N6uoqEh2u11NTU0qLCyccNymTZu0ZcsW7dq1S5mZmWpsbNSCBQv0sY99TDfccIMyMjL08Y9/XJdeemnCJgcAAAAAmN2iNr0Oh0M9PT2TtldUVET219TUSJIGBgYi+xcuXKj29vZJx33qU5/Spz71KcMFAwAAAAAQK1P4rXuTAQAAAACYYaJ+phcAAAAAgOkq6u3NqSAUCsUcm52dHXd8MnJQ09TnMFrT//l+V8zx//p31YZypNJ5MprjV0P/HXP85RcVSJKuafh2zMc83rAhKesj3nmk4ppNxfVR3dQac3zX5vWSpF8fGY75mNIL3z/rXotk1rT/8JGYj3EUX5iUmhoe6I35mIbrVsb9+0aSnP/SElP8nm2flTRD1seje2KPr3Qay5FC5ympNcV5bo3kiHWdv7XGU/G1uPFbbTHH/+T/rkvJ9YHpjyu9AAAAAIAZi6YXAAAAADBjGbq9eXBwUCdPnlRZWZn8fr+ys7Nls9kmxBw/flwrV67UokWLJJ1+2vO6deu0d+9etbScvr2opKRE27Ztk8lkOsdpAAAAAAAwmaGm1+fzyWKxqKysTB0dHXK73ZOa3oMHD2r16tXaunVrZNvo6Ki++c1v6gc/+IFyc3P13e9+VyMjI8rNzT23WQAAAAAAcAZRm97R0VF5vV6FQiGNjIzI5XKpq6tLGRkZKikpUX9/vwKBgIqLi1VXVyer1Sqr1aq8vDwFAgGtXbtWubm5qq+v1+DgoBYtWqTbb79dv//973X99dfT8AIAAAAAEiZq0zs0NKSqqiq5XC4Fg0F5PB5VV1fLYrHIbrervLxcbrdbBQUFGh4eVmdnp8xms/r6+rRkyRItW7ZMu3fvVmNjoyoqKvTLX/5SDz74oM4//3zV1dWptLRUl1xySTLmCgAAAACYZaI+yMpisaivr0+bN2/W3XffrbGxsbPGms1mmc1mSZLD4dBVV10lSaqsrNTBgweVk5OjD3zgA8rLy9P8+fP1wQ9+UL/97W+naCoAAAAAAEwUteltbW1VaWmpmpqatGrVKoXDYZlMJo2Pj0uSTCaTwuHw6cHS3h6uvr5evb2nv3Nv3759Wrx4sZYsWaJnn31WL7/8ssbGxvT000+ruLg4EfMCAAAAACD67c1Op1MNDQ3q7u5WTk6O0tPTZbPZ1NzcrKKiItntdjU1NamwsHDCcZs2bdKWLVu0a9cuZWZmqrGxUbm5udq0aZM++clPSpJWrVoVebozAAAAAABTLWrT63A41NPTM2l7RUVFZH9NTY0kaWBgILJ/4cKFam9vn3RcVVWVqqqqDBcMAAAAAECsTOG37k0GAAAAAGCGMfQ9vckWCoVijs3Ozo47Phk5qGnqc6RiTUZzUNPU56Am1gc1/eVypGJN8eSgptm3PpJS06N7Yo+vdMaV462anvzdizHnuPLiC2bva2EgB6a3qA+yAgAAAABguqLpBQAAAADMWIZubx4cHNTJkydVVlYmv9+v7Oxs2Wy2CTHHjx/XypUrI09nrqio0Ic+9CF97Wtfi8T8+te/VktLi66++upzmAIAAAAAAGdmqOn1+XyyWCwqKytTR0eH3G73pKb34MGDWr16tbZu3Tph+1tPdH7kkUe0YMECGl4AAAAAQMJEbXpHR0fl9XoVCoU0MjIil8ulrq4uZWRkqKSkRP39/QoEAiouLlZdXZ2sVqusVqvy8vIUCAS0du1a5ebmqr6+XgsWLJAk/fnPf9add96pH/7whwmfIAAAAABg9ora9A4NDamqqkoul0vBYFAej0fV1dWyWCyy2+0qLy+X2+1WQUGBhoeH1dnZKbPZrL6+Pi1ZskTLli3T7t271djYqJ07d0qSHnjgAa1atUq5ubkJnyAAAAAAYPaK2vRaLBa1tbXJ5/MpKytLY2NjZ401m80ym82SJIfDoczMTElSZWVlpOGVpO7u7gl/BgAAAAAgEaI+vbm1tVWlpaVqamrSqlWrFA6HZTKZND4+LkkymUwKh8OnB0t7e7j6+nr19vZKkvbt26fFixdLOv09Wq+//rre//73T/lkAAAAAAB4p6hXep1OpxoaGtTd3a2cnBylp6fLZrOpublZRUVFstvtampqUmFh4YTjNm3apC1btmjXrl3KzMxUY2OjJOmFF17QBRdckJjZAAAAAADwDlGbXofDoZ6enknbKyoqIvtramokSQMDA5H9CxcujDyp+Z2WLl2qu+66y3DBAAAAAADEyhR+695kAAAAAABmmKif6QUAAAAAYLqKentzKgiFQjHHZmdnxx2fjBzUNPU5DNf06J7Y4yudxnKk0HkymuPP/qdijj+/7IrTNcV5bpOxPuKdR0qu2RRcH0beR7wWqVPTaP+/x3xMVvmy5JyneH9/GPldHuMxyfrdH08Oo+tj5Vf/Leb4Xu9thnKk0nlKZk3XNHw75mMeb9iQ0PdFKq7ZSI4U+rdBpCYDOTC9caUXAAAAADBj0fQCAAAAAGYsQ7c3Dw4O6uTJkyorK5Pf71d2drZsNtuEmOPHj2vlypVatGiRpNNPe163bp2+973v6eGHH5bJZNJtt92mysrKc58FAAAAAABnYKjp9fl8slgsKisrU0dHh9xu96Sm9+DBg1q9erW2bt0a2Xby5Em1t7fL5/Pp1Vdf1Sc+8QmaXgAAAABAwkRtekdHR+X1ehUKhTQyMiKXy6Wuri5lZGSopKRE/f39CgQCKi4uVl1dnaxWq6xWq/Ly8hQIBLR27Vrl5uaqvr5eZrNZBQUFevXVV/Xqq6/KZDIlY44AAAAAgFkqatM7NDSkqqoquVwuBYNBeTweVVdXy2KxyG63q7y8XG63WwUFBRoeHlZnZ6fMZrP6+vq0ZMkSLVu2TLt371ZjY6PuuOMOvf/971dVVZXefPNNffrTn07GHAEAAAAAs1TUptdisaitrU0+n09ZWVkaGxs7a6zZbJbZbJYkORwOZWZmSpIqKyu1c+dOPfHEEzp69Kh+8YtfSJJuvfVWXXHFFVq6dOlUzAUAAAAAgAmiPr25tbVVpaWlampq0qpVqxQOh2UymTQ+Pi5JMplMCofDpwdLe3u4+vp69fb2SpL27dunxYsX673vfa/OO+88zZ07V/PmzVN2drZOnjyZiHkBAAAAABD9Sq/T6VRDQ4O6u7uVk5Oj9PR02Ww2NTc3q6ioSHa7XU1NTSosLJxw3KZNm7Rlyxbt2rVLmZmZamxs1IIFC/Tv//7vuuGGG5SWlqYrrrhCy5cvT9jkAAAAAACzW9Sm1+FwqKenZ9L2ioqKyP6amhpJ0sDAQGT/woUL1d7ePum4z3/+8/r85z9vuGAAAAAAAGJlCr91bzIAAAAAADOMoe/pTbZQKBRzbHZ2dtzxychBTVOfIxVrMpqDmqY+BzWxPqjpL5cjFWuKJwc1zb71kYyanvzdizHHX3nxBXHliNT06J7Ya6p0ztrXwkgOTG9RH2QFAAAAAMB0RdMLAAAAAJixDN3ePDg4qJMnT6qsrEx+v1/Z2dmy2WwTYo4fP66VK1dq0aJFkk4/+GrdunX6zne+o4cfflhZWVn65Cc/KafTee6zAAAAAADgDAw1vT6fTxaLRWVlZero6JDb7Z7U9B48eFCrV6/W1q1bI9sGBwf10EMP6f7775ck1dTUyOFwKDMz8xymAAAAAADAmUVtekdHR+X1ehUKhTQyMiKXy6Wuri5lZGSopKRE/f39CgQCKi4uVl1dnaxWq6xWq/Ly8hQIBLR27Vrl5uaqvr5ezz33nD70oQ9p3rx5kqSLLrpIg4ODKi0tTfhEAQAAAACzT9Smd2hoSFVVVXK5XAoGg/J4PKqurpbFYpHdbld5ebncbrcKCgo0PDyszs5Omc1m9fX9//buPi6qMu8f+GdkwBBwGcPcSHsAVG5shRZL78xNSMbH2kVTSEQz3NweLFMjzCdSN0ORvJPW+zbB0FJ0E0yqVXyKUDSxNTFa4SeoaRKQoMyAwADX7w9eM0kwzJzJM84Mn/frta/dnfle5/u9znXOcS7mzHUO4MEHH8Sjjz6KPXv2YOXKlXj11VexceNGaLVa6HQ6nDp1ChEREdboJxEREREREXVBJie9Xl5eSEtLQ3Z2Ntzd3dHU1GQ0VqVSQaVSAUCb25bDwsLw3nvvwdfXF1FRUfjrX/+K++67D4GBgYZ4IiIiIiIiolvN5OrNqampCAoKQmJiIsaMGQMhBBQKBVpaWgAACoUCQojWjXX7ZXOLFy/Gvn37AADHjh3DoEGDUFVVherqamzfvh2LFi1CWVkZ+vfvL0e/iIiIiIiIiEx/0xsSEoL4+HhkZWXB09MTTk5O8Pf3R1JSEnx9fREYGIjExET07du3Tbv58+fjzTffxPbt2+Hq6oqVK1dCpVLh8uXLmDRpEpydnREbGwsnJyfZOkdERERERERdm8lJ77Bhw7B37952r48aNcrwfmRkJADg6NGjhvf79euHrVu3tmu3fPlyi4slIiIiIiIikkIh9PcmExERERERETkYi57Ta20ajcbsWA8PD8nx1shhjZp0P5WbHe/8+z4W1aS7/KP5OfreY1kOM/vh/Ps+NleTvi7JYyexD4BtHh/WOC90V34yK97Z+/cW1dRU+bPZ8creXpJyWHM/OcI1zZIcjRcvmRXvcl8/SfH6NuYef8Avx6DuxzLz4u+5GwCPD1vIYYjff9j8HGEhVqnJEa7lrOnW57C4JonHuEU5JMZLvS5boyayfyYXsiIiIiIiIiKyV5z0EhERERERkcOyaNJbVFSE/Px8AEB+fj7Onj3bLqaurg6xsbGYOnUqJk+ejIKCAgDAoUOHMGnSJERERGDnzp2/oXQiIiIiIiKizlk06c3Ozsa5c+cAALt27UJFRUW7mJSUFPTv3x/btm3DihUrUFpaCp1Oh1WrViE1NRVbt27Fjh07UFlZ+dt6QERERERERGSEyYWstFotFi1aBI1Gg+rqaqjVamRmZsLZ2RkBAQHIzc1FYWEh/Pz8EBUVBR8fH/j4+KCgoABjx45FTEwM3NzcsGzZMpSUlODee+/F7373OwBAcHAwTp48ibFjx8reUSIiIiIiIup6TE56L168iPHjx0OtVqO8vBzR0dEIDw+Hl5cXAgMDMWLECIwbNw7e3t4oKytDRkYGVCoV1Go1ampqkJKSgt27dyMhIQFTpkxpswKam5sbtFqtrB0kIiIiIiKirsvk7c1eXl44cOAAFixYgA0bNqCpqclorEqlgkqlAgB4enoiNDQUABASEoLvvvsO7u7uqK2tNcTX1tZyGXAiIiIiIiKSjclJb2pqKoKCgpCYmIgxY8ZACAGFQoGWlhYAgEKhgBCidWPdftlccHAwcnJyALQuduXn5wdfX19cvHgR165dQ2NjI06ePImHHnpIjn4RERERERERmb69OSQkBPHx8cjKyoKnpyecnJzg7++PpKQk+Pr6IjAwEImJiejbt2+bdrNnz8bixYsREREBpVKJhIQEODs7Iy4uDjExMRBCYNKkSejTp49snSMiIiIiIqKuzeSkd9iwYdi7d2+710eNGmV4PzIyEgBw9OhRw/uenp5ITk5u1y40NNRw2zMRERERERGRnBRCf28yERERERERkYOx6Dm9RERERERERPbA5O3NtkCj0Zgd6+HhITneGjmsUZPu8o9mxzv3vceimhpLL5jdxsXnfllzuPjcD8A6x4fc+9ZRjg9r9Fvu40P3Y5nZ8c733C0phzX3OUHiiQAAIABJREFUkyNc0yw6Pi78YFa8y/33tsZfvGR2Dpf7+km+BgK2eU3rqseHPZ+rhmP8h8tm53C5t69NjoUlNd04U2h2G9c/DJIcb4v7SUoOq9a0/7D5bcJCJOew5Dpbf7bY7DZ3+A+wqN9k3/hNLxERERERETksTnqJiIiIiIjIYVk06S0qKkJ+fj6A1mfwnj17tl1MXV0dYmNjMXXqVEyePBkFBQWG927cuIHIyEiUlJRYWDYRERERERGRaRZNerOzs3Hu3DkAwK5du1BRUdEuJiUlBf3798e2bduwYsUKlJaWAgDOnDmDqKgoXLpk/u+oiIiIiIiIiCxhciErrVaLRYsWQaPRoLq6Gmq1GpmZmXB2dkZAQAByc3NRWFgIPz8/REVFwcfHBz4+PigoKMDYsWMRExMDNzc3LFu2DADQ2NiI999/H7GxsbJ3joiIiIiIiLo2k5PeixcvYvz48VCr1SgvL0d0dDTCw8Ph5eWFwMBAjBgxAuPGjYO3tzfKysqQkZEBlUoFtVqNmpoapKSkYPfu3UhISMDq1asRHBxsjX4RERERERERmZ70enl5IS0tDdnZ2XB3d0dTU5PRWJVKBZVKBQDw9PREaGgoACAkJAQbN268RSUTERERERERmcfkb3pTU1MRFBSExMREjBkzBkIIKBQKtLS0AAAUCgWEEK0b6/bL5oKDg5GTkwOgdbErPz8/OeonIiIiIiIiMsrkN70hISGIj49HVlYWPD094eTkBH9/fyQlJcHX1xeBgYFITExE375927SbPXs2Fi9ejIiICCiVSiQkJMjWCSIiIiIiIqKOmJz0Dhs2DHv37m33+qhRowzvR0ZGAgCOHj1qeN/T0xPJyclGt7t161bJxRIRERERERFJoRD6e5OJiIiIiIiIHIzJb3ptgUajMTvWw8MDTeWVZscr+/QGAFTV1ZvdplePOyTXJDUekN5vR6hJSg6La9p/2Pz4sBDLctjQfrI0R8P/KzE7vnt/39aaJO5baxwfUvthk8esDR4flpxHDUXnzG7TfaBflxsLa9Yk93lh0X6Sev2w5FpuZhtrXful5LD0+PjLmhSz43e/HmNRDqmfu2zymmZBTRHr0sxus2PuDFnPC1s8Zg05bOizgaEmC3KQfTO5kBURERERERGRveKkl4iIiIiIiByWRbc3FxUVoaamBg8//DDy8/Ph4eEBf3//NjF1dXWIj4/H5cuXodPpsGTJEgwePBifffYZ0tLS4OTkhAEDBiA+Pr7No46IiIiIiIiIbhWLZpvZ2dk4d671t1m7du1CRUVFu5iUlBT0798f27Ztw4oVK1BaWor6+nqsW7cOW7ZsQXp6OrRaLQ4fNv8+fyIiIiIiIiIpTH7Tq9VqsWjRImg0GlRXV0OtViMzMxPOzs4ICAhAbm4uCgsL4efnh6ioKPj4+MDHxwcFBQUYO3YsYmJi4ObmhmXLlsHFxQXp6elwdXUFADQ1NaF79+6yd5KIiIiIiIi6JpOT3osXL2L8+PFQq9UoLy9HdHQ0wsPD4eXlhcDAQIwYMQLjxo2Dt7c3ysrKkJGRAZVKBbVajZqaGqSkpGD37t1ISEjA6tWr4eXlBaD1Ob11dXUYPny47J0kIiIiIiKirsnkpNfLywtpaWnIzs6Gu7s7mpqajMaqVCqoVCoAgKenJ0JDQwEAISEh2LhxIwCgpaUFa9aswfnz57F+/XooFIpb0Q8iIiIiIiKidkz+pjc1NRVBQUFITEzEmDFjIISAQqFAS0sLAEChUEAI0bqxmxakCg4ORk5ODgAgPz8ffn5+AIClS5eioaEB//jHPwy3ORMRERERERHJweQ3vSEhIYiPj0dWVhY8PT3h5OQEf39/JCUlwdfXF4GBgUhMTETfvn3btJs9ezYWL16MiIgIKJVKJCQkoLCwEJ988gmGDBmCGTNmAACmT5+OsLAweXpHREREREREXZrJSe+wYcOwd+/edq+PGjXK8H5kZCQA4OjRo4b3PT09kZyc3KZN3759cfbs2d9UMBEREREREZG5FEJ/bzIRERERERGRgzH5Ta8t0Gg0Zsd6eHhIjrckR1Plz2bHK3t7WaUmzX7zn3nsERYiuQ8AJOewqN9m5vAICwEA3Dj9ndk5XAMflLUmfV1Sc9SdPGV2fI8hDwGA9GPQGmNnjXNV4vEh93kkJYe+D40XL5mdw+W+flYZi0rtDbPje7u7Os7xIfW8sOT4kPuYtcJYNPy/ErPju/f3td3jQ+K5KvdnA4tqssK1XO7POLZ4fNhiTVJyWFqT1H+PAKCx9IL5bXzut8q/w9ovj5jdxn3kYxaNBdk3kwtZEREREREREdkrTnqJiIiIiIjIYVk06S0qKkJ+fj6A1scRdbQ4VV1dHWJjYzF16lRMnjwZBQUFAIB9+/Zh0qRJePrpp/HPf/7zN5RORERERERE1DmLJr3Z2dk4d+4cAGDXrl2oqKhoF5OSkoL+/ftj27ZtWLFiBUpLS9Hc3Iy1a9fiww8/xI4dO7Bp0yZUVVX9th4QERERERERGWFyISutVotFixZBo9GguroaarUamZmZcHZ2RkBAAHJzc1FYWAg/Pz9ERUXBx8cHPj4+KCgowNixYxETEwM3NzcsW7YMTk5O+OKLL6BUKnH16lUAgJubm+ydJCIiIiIioq7J5KT34sWLGD9+PNRqNcrLyxEdHY3w8HB4eXkhMDAQI0aMwLhx4+Dt7Y2ysjJkZGRApVJBrVajpqYGKSkp2L17NxISErB69WoolUpkZ2dj+fLlePzxx6FU2sUC0kRERERERGSHTN7e7OXlhQMHDmDBggXYsGEDmpqajMaqVCqoVCoAgKenJ0JDQwEAISEh+O67Xx4ro1ar8dVXX0Gn02H37t2/tQ9EREREREREHTI56U1NTUVQUBASExMxZswYCCGgUCjQ0tICAFAoFBBCtG6s2y+bCw4ORk5ODoDWxa78/Pyg1Woxbdo0NDY2olu3bnB1dW3ThoiIiIiIiOhWMnlvcUhICOLj45GVlQVPT084OTnB398fSUlJ8PX1RWBgIBITE9G3b9827WbPno3FixcjIiICSqUSCQkJcHd3x5NPPomoqCgolUoMHDgQTz31lGydIyIiIiIioq7N5KR32LBh2Lt3b7vXR40aZXg/MjISAHD06FHD+56enkhOTm7XLiIiAhERERYXTERERERERGQuhdDfm0xERERERETkYPiDWiIiIiIiInJYdvG8oApNndmxd3n0gEajMTvew8MDACS3kTvekprqC8+aHX/HIH+LatLm5pndxn3Eo9AczDE/xxOPAwBunP7ORGQr18AHAQBNlT+bnUPZ28uifsu9b5vKK82OV/bpDcA2jw9rnBfm9uOOQf4W1SR1PwFAXf6/zYrv8fAfLarJorHYf9j8NmEhuFpbb3b8nW532Ozxof3yiFnx7iMfa61J4n6y5Pgw97rpPuLR1pqk7icLrrOO8G+enDlsuSbd5R/NzuHc9x6bHAtrXNMkx9vgfpKSw6r/Dp8tNrvNHf4DJOcw9zoOWH4tt6TfZN/4TS8RERERERE5LE56iYiIiIiIyGFZNOktKipCfn4+gNZn8J492/52r7q6OsTGxmLq1KmYPHkyCgoK2ry/ZMkSJCYmWpKeiIiIiIiIyCwWTXqzs7Nx7tw5AMCuXbtQUVHRLiYlJQX9+/fHtm3bsGLFCpSWlhreS09PR3Gx+b8HICIiIiIiIrKEyYWstFotFi1aBI1Gg+rqaqjVamRmZsLZ2RkBAQHIzc1FYWEh/Pz8EBUVBR8fH/j4+KCgoABjx45FTEwM3NzcsGzZMgDAqVOncPr0aURERLSZCBMRERERERHdaiYnvRcvXsT48eOhVqtRXl6O6OhohIeHw8vLC4GBgRgxYgTGjRsHb29vlJWVISMjAyqVCmq1GjU1NUhJScHu3buRkJCABQsWIDk5GcnJyfjXv/5ljf4RERERERFRF2Zy0uvl5YW0tDRkZ2fD3d0dTU1NRmNVKhVUKhUAwNPTE6GhoQCAkJAQbNy4EXv37kV1dTWef/55VFZWor6+Hj4+Ppg4ceIt6g4RERERERHRL0xOelNTUxEUFISpU6fi+PHjyMnJgUKhQEtLCwBAoVBACAEA6Nbtl58IBwcHIycnBw8++CDy8/Ph5+eH6dOnY/r06QCAjIwMlJaWcsJLREREREREsjE56Q0JCUF8fDyysrLg6ekJJycn+Pv7IykpCb6+vggMDERiYiL69u3bpt3s2bOxePFiREREQKlUIiEhQbZOEBEREREREXXE5KR32LBh2Lt3b7vXR40aZXg/MjISAHD06FHD+56enkhOTja6XX7DS0RERERERHJTCP29yUREREREREQOxuQ3vbZAo9GYHevh4SE53ho5WNOtz2GLNVmagzXd+hysSVqbSu0Ns+N7u7vy+OgiNVmawxZrkpKDNXW948MWa5KSwxZrsjSHVWraf9j8+LAQs2PJdnUzHUJERERERERknzjpJSIiIiIiIodl8aQ3IyMDiYmJt7IWIiIiIiIioluK3/QSERERERGRwzJ7Iav6+nosXLgQV65cgU6nw+jRow3vrV27Ft999x1qa2vh6+uLVatW4ZtvvkFCQgKUSiV69uyJxMREVFZWYuHChVAqlXBycsLq1avRp08fWTpGREREREREZPakNz09Hffccw/effddFBcXIy8vDxqNBlqtFj179sTmzZvR0tKC8ePHo7y8HAcOHEBYWBhiYmJw6NAh1NTUIC8vD4MGDUJcXBxOnjyJ69evc9JLREREREREsjH79ubS0lIEBQUBAAYMGICePXsCALp3746qqirMmzcPS5cuRV1dHXQ6Hf72t7+hqqoKM2bMwN69e6FUKvH0009DpVJh1qxZ+Pjjj+Hk5CRPr4iIiIiIiIggYdLr6+uLM2fOAAAuXbqEpKQkAMBXX32FsrIyJCUlYd68eaivr4cQAllZWQgPD8fWrVvRv39/7Ny5EwcPHkRwcDDS0tIwZswYbNq0SZ5eEREREREREUHC7c2RkZF48803MW3aNDQ3N2PmzJmorq7G4MGD8Y9//ANTpkyBi4sL+vXrh4qKCvzhD39AXFwcevToAWdnZyxfvhxCCLz++utYv349unXrhoULF8rZNyIiIiIiIurizJ70du/eHWvXru3wvV27dnX4ekZGRrvXduzYYW5KIiIiIiIiot9EIYQQt7sIIiIiIiIiIjnwOb1ERERERETksDjpJSIiIiIiIofFSS8RERERERE5LE56iYiIiIiIyGFx0ktEREREREQOi5NeIiIiIiIiclic9BIREREREZHD4qSXiIiIiIiIHBYnvUREREREROSwlLe7ACIiIiKyLTqdDkVFRdBoNOjZsyf69+8PFxeX212WrLRaLdzd3QEAxcXFOHv2LAYNGgRfX99bluPIkSN47LHHbtn2OtLQ0ICioiLU1dVBpVJhwIABUCgUsuYksnVO8fHx8be7CDlcuHABS5YswaFDhzBgwAB4enoCAJYtW4aQkJB28Q0NDfj444/x448/wsPDAy+++CI+/fRTBAYGQqVSmcw3f/58jB49utOYrVu3IjAwEJWVlVi4cCHeffdd/Pvf/8aQIUPQo0ePdvFnzpzB119/jd69eyM+Ph7/8z//gxMnTmDw4MHw8PBoFx8ZGYmgoCD06tXLZL0393vHjh34z3/+g379+uHVV1/Ftm3b8OCDD8LLy6vD+A8++AAbN27Ezp07cfjwYVy/fh0BAQHo1s34jQNXr15Fbm4uvv32W1y+fBkeHh5wc3Mzu05TDh8+jLKyMvTr18/w2oEDB+Dj42O0zbVr19CtWzd069YNu3fvRlFREQYOHGjWPwy7d++Gv7+/0fcvXLhgOOZycnLwr3/9Czdu3MB9991ntE1jYyMuXboET09PHD9+HPv27YNGo8H999/fYfyOHTswaNAgSf+QaTQa6HQ6uLi4YN++fTh16hQGDhzY6dhduXIFubm5KCgowE8//QSVSoU77rjDaLzcYw1IH+/fMtbArR9vqWMNWGe8pY410PXObUD+8bZkrAH5x5vndte5ln/55ZeYO3cuSkpKUFJSgry8PGzYsAF9+vTp9NyTqri4GFqt1rCPAeD06dP4/e9/b7LtiRMnUFZWBm9vb7NyHTlyBPfee2+nMbNmzUJ4eDh27dqFdevWwdXVFR999BGam5vx4IMPdtimuroarq6uuHjxIo4fPw4AnX4OmzBhAkpKSvDII4+YvL525MyZMygpKWlzjtzsyy+/RGxsLM6fP4+PPvoIly5dwocffogHHnig033VFa/lAFBRUYGysjJcv34da9asgbe3N3r37m1ep8iuKIQQ4nYXYUpSUpLR9+bNm9fh69HR0Zg9ezaampqwZs0arFmzBgEBAYiOjsbWrVvbxb/22mvw9fVFWVkZTpw4geXLl6NHjx5Yt24dNm/e3C5+5MiRaGpqMvz/a9euGS7aR44c6bCm6dOnY8uWLZg7dy6eeOIJhIWFIS8vDzt37sT//u//touPiIjA8uXLsWHDBowcORKhoaE4ceIE0tLSOuzD2LFj0bNnTwwfPhzPPfec4a+VnXn55Zfh6+uL2tpa5Obm4s0330Tv3r2xatWqDnPExsbikUcewUMPPYRDhw4ZLkbnz5/H8uXLO8zxz3/+Ezt27EBwcDDc3NxQW1uL/Px8TJ48Gc8880y7+PPnzxut94EHHmj3Wnx8PDQaDZqamnDjxg0kJyfDxcXFsL+N1ZSSkgIAePjhh9HY2AhXV1d069YNS5cubRf/62Pws88+w4QJEwB0fAzqc2/cuBHffPMNHn/8cRw/fhwDBgzAyy+/3GFNr776KkaMGIHr168jLy8PI0aMwDfffIO7774bb775Zrv4hx9+GIMGDcJbb73V6Qcwve3btxuO5ZEjR+Lq1avo1asXtFotVq1a1WGbTz75BFlZWfjDH/6AY8eOYdCgQTh//jyio6OhVqvbxcs91oD08ZY61oD84y11rPW1yzneUsca6JrnNiD/eEsda0D+8ZY61oD8422L5zbgGNfyyMhIbNq0qc3nCI1Gg2effRa7du1qF9/Y2Gi0XmPfDr///vs4cuQImpqaEBAQgPj4eCgUCqPj/eWXXyI+Ph49e/bE6NGjkZ+fDxcXFwQFBeHFF19sF79jx442/3/z5s2YOXMmgNbPVx3R537mmWewadMmuLm5QafTYfr06di+fXu7+OXLl+Oee+7BnXfeibS0NAwZMgSnT5/G6NGjERMT02GO6OhoREVFYf369Rg7diwmT56MPn36dBgLtE4M3377bXTr1g3R0dE4cOAAPDw88MADD+D111/vcPspKSlwcXFBdXU1Vq9ejUWLFuH555/Htm3bOszRVa/lADBz5kzMnj0b27Ztw+jRo5Gent7h519yAMIObN68WajVapGZmSkyMjLa/MeYadOmGf53SUmJGDdunLhy5YqIjo7uMD4qKkoIIURzc7MYN26c4fXp06d3GH/ixAnxwgsviPLy8nb5jNHn/nWsqZqee+65Nq9HREQY3b5OpxOpqali9OjRYsmSJWL//v3iP//5j9Gabq5l/PjxHb7eUU16s2bNEkII8cwzzxjNERERIRobG9u81tDQICZOnNhhfGRkpBg+fLiIjo4W06ZNM/zH2H6KjIw0/O8tW7aIF154odM+CCHE5MmTRXNzs/j555/F8OHDDa9PnTq1w/glS5aIiIgIwzE4ZsyYTo9Bfa3PPPOM0Ol0QgghWlpajG7/5tzTpk0ztBFCiKeffrrD+GnTpolTp06JiRMniri4OPHvf//b6Lb122lsbBQajUaMHDlStLS0dNpnIVrHWx9XV1cn5s2bJxobG40eg3KPtb6NnjnjLXWshZB/vKWOtT5WzvGWOtZCdM1zWwj5x1vqWOu3Jed4Sx1rIeQfb1s8t2/Ob8/X8okTJ7apXYjW8Z40aVKH8Wq1WgQHB4vQ0FAREhLS5r+NmTJliqGmd955RyxbtkwI0fl4a7Vacf78eTF06FCh0+lES0uL0T7MmjVLTJkyRaxfv16sX79ehISEGP63MRMnThTV1dVizpw5or6+XgjROt7G+q3PPXXqVFFbWyuEEEKn03V6XuiPqbq6OrFlyxYxefJk8Ze//EW89NJLHcY//fTT4vr166KsrEw8+uijoqGhoU3uX/vzn/9sOFdra2sNx4WxPui31RWv5foampqaxIwZMzrdPtk/u/hN77PPPovCwkLcddddePTRR81qo1QqcejQITz++OPw8fHBkiVLDN/8Govfs2cPnnrqKXz66acAgK+//hotLS0dxj/88MPo168fli5diueee86sWy2Ki4uxcuVKNDU14dixYxg6dCj27dtnNP6ee+5BSkoK/vSnPyE5ORmhoaHIyckxetuFEAJKpRIzZ87EtGnTkJeXh2PHjuGTTz7p8Jtkve3bt+P69eu4du0a8vLy4O7u3uktUl988QVGjBiBgwcPwtXVFcXFxWhoaDAa39TUhIaGBjg7Oxteq6+vN7rPUlNTMW3aNKxZs6bTv37qNTc3o7GxES4uLoiOjsaVK1ewcuXKTtu0tLTgxo0buPPOO7Fs2TIArX+p1ul0HcYvX74c6enpOHHiBJYtW4bMzEyEh4cb3X5VVRW+//579O7d23DrVn19faf7CQAuXbqEAQMG4IcffoCPjw8uXbpkNFahUCAoKAi7du3CoUOHkJaWhtjYWLi7uyMzM7NdfHNzM+rr63H9+nXU1dWhrq4OLi4unf6FvqamBlqtFh4eHrhx4wZ++uknODs7G+2H3GOt74eU8ZY61oB1xlvKWAPyj7fUsQa65rkNyD/eUscakH+8pY41IP942+q5Ddj/tTwiIgLh4eEIDg6Gh4cHtFotvvnmG0RHR3cYv337dsTExODDDz/E7373u073jZ4QwnD8vPHGG5g/fz42bdpk9JhqaWmBq6sr7r//fsyZMwdKpdKwnY5s3LgR69atQ3NzM1555RV8/fXXRr+d13vooYfw4osv4uLFi9i8eTOio6MxdepUPPXUU0b7cO3aNfTr1w/19fXo0aMHtFqt0ZpurtfV1RXR0dGIjo6GVqs1+u1pc3Oz4TZjhUJh2D/GPp+OGzcOkydPxiOPPIKTJ09i6tSp+OCDDxAQEGC0pq56LQdaf7u+atUqDBkyBMePH0dzc7PJNmSnbueMW4r6+npRU1NjdvyVK1fEG2+8Iaqrqw2vHTt2TDz11FMdxldWVoq///3vbV6Lj48XJSUlneZpaGgQixYtEmPGjDFZ07Vr10ReXp7YuHGjOHjwoNBqtWLu3Lni8uXLHcbX1dWJ9evXi4iICKFWq8WUKVNEQkKCuHbtWofxv67fHFeuXBFvv/22yMzMFHl5eSI8PFw8++yz4ty5cx3GX7p0ScyZM0eMHz9ezJ8/X1RUVIjMzExx+vRpozkOHjwowsLCxKxZs8Rrr70m/vrXvwq1Wi0OHz5stM2ZM2fM+nZDCCGysrJEWFiYuHr1qhCi9a+yixYtEv/1X/9ltM3evXuFWq0Wzc3NhtemTZsmdu7c2WmuU6dOieeee67Tv+IK0Xp3woIFC8TYsWNFcnKy0Gg04vHHHxeffvqp0TanT58Wf/7zn8WMGTNEYGCgePLJJ8UTTzwhjh8/3mG8sb+Q6vfDr3366adi+PDhIiYmRrzzzjtizJgxIjw8XKSnpxutKTMzU4SEhIgXX3xRqNVqcejQIZGUlGT0L+XGxvrQoUNGc0gZayGkj7elYy2EfON981gPHjzYMNbHjh0zmuNWjff27ds7jJc61kJ0zXNbiN823nKc20LIP96WjLUQ8o63LZ7bQjjGtVyI1s9FBw8eFHv27BGHDh0SlZWVRmOFECI3N1fk5eV1GnOzzZs3i0mTJhk+pzU0NIjnnntODB48uMP4jz76SEyYMKHNeL/88ssiOTm50zx79+4VL7/8spgyZYrZtbW0tIja2lrR0tJi9POQEEJ8+eWXYsKECWLevHli+PDh4m9/+5t44oknxOeff260TWd333Vk06ZNYuTIkSIqKkrMmzdPTJ8+XTz//PPivffeM9qmqKhIfPHFF4bPsFVVVZ3m6KrXciGEOH/+vPjoo49EQ0OD+Pzzz8UPP/xgVjuyP3Yz6RVCiLffflv2NpbG629zliOH1MmsNfaT1Jp0Op04e/asOHnypDh79my726Z+q/r6esNtUnqFhYWdtrn5QiqEEBqNxqxc5eXlYsuWLdIKlLD90tJS8c0334iSkpJ2txvdzNQHEFPOnj0rfvzxR5NxVVVV4vTp0+L69etCCCGampo6jZd7rIUQhtvObtbZeFs61kIIUVFRIdt4mzvWQlhnvKWOtRBtx7uoqEiWc/vX5Dq3KyoqxNatW80vTmIOa53bQrSO95UrVzqN+fV4/3q//Zq1zm0p1/Lfcm7zWm78/Nbvl4qKCjFnzhwxatQoMXfu3FtybN7shx9+aFfH/v37jcb/evJWWlpqVp7i4mKxZs0ak3H6fldWVoo5c+aIsLAwk/3WarUiNzdX7NmzR+Tm5nb6x6mbc+j3rTk5ampqhE6nEzqdThw8eFCcPHnylm5fCOtcy631OU3Ktby2tlaUlZWJyspKkZycbPSLKLJ/dvWc3pKSEtTU1MjaxtL4u+66S7YcpaWlsvbBGjUplUoMHDgQwcHBGDhwoOG2pFule/fu7W7D6exWHgDtbuE2Z+EvALjrrruM3uLVGXO3/8ADD+CPf/wjfHx82txq9Gsdra4txcCBA81a9VKlUmHw4MHo2bMnAMDJyclo7IULFzBv3jx88MEHuPPOOw1jrb81qaP4OXPmYMGCBbhw4YLhdWPx+javv/56uza/XrTk5vi5c+e2iXd3dzeZQ19XXV2dYbzN6cfFixdN5rhw4QJeeeUVvP/+++jVq5dhrDurSavVStpX+hz6mvTj3Vn80qVpLdCxAAAK8klEQVRLsWXLFlRXVwNoHevOatKven/u3Dl4e3tjxYoViImJQWlpqdH4tLQ0fP755ygrK0N0dDRmzJjRaXx6enq7+M5WPG1oaMDWrVvbtHnppZfMqqmpqQn79u3rtCZj/TAnx/fff4+7774by5Ytw6xZs4zGe3h4SNpPHdW0cuVKxMXFdVrTnj17cOnSJdTW1iI6OhozZ87sNMfN1/GNGzdKvo7Pnz/fZMzN13J9fGfX8puv4/Pnzzf7Ogu0Xsu//fZbs+Ol5tBfy99//31J13Jz9tPNNm7cKPlaPn/+/E6v5fv37wcA/P3vf0dYWBiysrLw5JNPYvHixR3G6xf++fnnn/HKK69ArVbjtddew88//2w0x9atW9GvXz9UVVW1aRMUFGQ0XqVSobKy0hD/3nvvGc1xc03r16/Hvn37TNak7/fKlSsRFhaGPXv2mOy3m5sbBg4ciP379+Ott97CihUrzMqh37fm5PDw8MC1a9cwb948rFq1Ch999JHRHFK3D7SuCP3FF1+gd+/e2LlzJ+bNm4cFCxbgypUrRuP37NmDqqoqvPHGG5gwYQLmzp3bafy+fftQXV3dJv7mVbs7avPZZ5+1ybF48WKzakpMTMT27ds7rUlvwYIFKCwsxOrVq+Hs7Gx0ETyyf3bxm169kpISDBs2DCqVyvCPorGVki1tI3e8pTmGDh2KXr162VRN5sZLXX1b7njWJF9NN/92/qWXXjKsmm7sg7TU+NuR48UXX5SUQ2q8NfotRx8AIC4uDr6+viguLsa6desMq96vWLGiw1Xv5Y53lBy2WFNHTyz4+uuvARi/9kttI3d8R22uX79ueGaqXDlssd/mtAFaH2Pz5JNPAgBCQ0Px4Ycfdhi3f/9+REdHGyaLq1evRl5eHhYvXmx0TRF9G/3kzFQbS+Ol1GRpv82t6bfkkNoPc7cPtE7yly9fjuXLl2PkyJFYtGgRTpw4gTfeeKPDlYzljrdWDqD19+6hoaHYsmULVq9ejdzcXKOxZN/satK7bds23H333Yb/X1JScsvbyB3fFWvq1asXtm/fjhdeeKHTxR2sFc+a5KsJgOED5L333os5c+Z0ujCJJfGOksMRaqqsrMS7776LlpYWPPnkk/jv//5vAMYXWJE73lFy2GJNa9aswebNmxEfH2+428XUYz2ktpE73lFyWKMmqQtv6kmZaFnaRs54a/Rb7hyWbN/Z2RkDBw6ERqPBX/7yFwDAqFGjsGnTptsSb60cQOtCVqmpqQgICMC5c+dQW1vbaTzZL7uY9BYXF6O8vByJiYmIjY2FEAItLS1Yu3atYaXl39pG7viuWhMgffVtueNZk3w1WbJqupR4R8nhSDVJWfVe7nhHyWGLNVnyxAKpbeSOd5Qc1qhp3759+P7779GnTx/cuHEDN27cQHZ2ttHnAFsy0ZLaRu54a/Vb7hxStw9If1qI3PHWygEAsbGxOHjwIF544QVkZWUhPj6+03iyY1b+DbFF8vPzRVxcnBg+fLiIi4sTcXFxYuHChZ2uVCi1jdzxXbUmPamrb8sdz5rkqUnqqulS4x0lh6PUJHXVe7njHSWHLdakJ+WJBZa2kTveUXJYoyZzF7mU+nQKS9rIHW+tflsrh5QFSqU+LUTueGvl0CsvLxc//vijuHz5sqSnSZB9sYtJr96vH4vz9ddf3/I2csd31ZqEsN5K2raUgzXJE+8oOViTPPGOksOWa7LkiQXmtpE73lFyWKOmmJgYw2rPUrYvhS2eR9bot9w5pG5fCOlP5pA73ho5Fi5cKMaOHStGjhwphg4dKiZPniypPdkPu7i9+eTJkygpKcHmzZsxc+ZMAK0Put62bRs+++yzW9JG7viuWtPN9CtE61cBNkXueNbEmm53DtZkvzVZI4ct12TJEwvMbSN3vKPksFZNUhe5tNVjVmpN1ui3nDksWdRU/2QOc3PIHW+tmj7//HMsXboUr732Gl599VWzayP7YheT3p49e6KyshKNjY2orKwE0PqYAmMrzlrSRu74rlrTzRxlJW3W1DVqskYO1mS/NVkjB2uy35qskcMaNUld5NJR9pM1+i13DksWNS2x4GkhcsZbI4ebmxsUCgXq6urQq1cv6HS6TushO3a7v2qWYsOGDUKtVovQ0FAREhIiwsLCbnkbueO7ak1CtP5m8Gbnzp27rfGsiTXd7hysyX5rskYO1mS/NVkjh5w1FRUVia+++ko89dRT4siRIyI3N1fk5OR0+ht/uWuyRrw1+i13Dku3L7Uf1oi3Ro61a9eKTZs2iXfeeUfMnTtXTJo0yWRNZJ/satIbHh4uysvLRXx8vDh+/Lh44YUXbnkbueO7Yk1SL8Byx7Mm1nS7c7Am+62pq/abNdlODmvUJHXRSkfZT9bot9w5LFlw1BbHwho59LRardDpdOLgwYPi559/NhlP9skubm/WU6lUuOuuu1BbW4uhQ4fivffeu+Vt5I7vijXV1NTgiy++wNWrVw2/+1UoFJg6deptiWdNrOl252BN9ltTV+03a7KdHNaoaciQIRgyZAgKCgowePBgw+snTpy4bTVZYz9Zo99y55C6fUty2OIxKzV+7dq1HT6669tvvzXrJ3tkh273rFuKV199Vezfv1/ExsaK7du3m7XkvtQ2csd31ZqE6JqrVrMm+63JGjlYk/3WZI0crMl+a7JGDjlrys/PF+np6WL06NEiPT1dpKeni48//liMHz/+ttVkjXhr9FvuHJZuX2o/rBEvZ46MjAyRkZEh/u///k98+OGHIiMjQzz22GMiNTXVZE1knxRCCHG7J97m0mq1+OGHH+Dl5YXU1FSEhIRg6NCht7SN3PFdsaauuGo1a7Lfmrpqv1mT7eRgTfZbk6P0u7i4GNnZ2cjIyMDEiRMBtC5a6e/vj9DQUIfdT9bot9w5pG7fkhy2eMxa+nSRp59+Gu+88w78/Pxw6dIlxMXF4eOPPzYaT/bLrm5vdnd3R0BAAAAgLi5OljZyx3fFmrriqtWsyX5r6qr9Zk22k4M12W9NjtLvAQMGYMCAAVAqlcjMzERTUxOEEFAqlR1OnBxlP1mj33LnkLp9S3LY4jFr6dNFlEol/Pz8AAD9+vVDt27dOo0n+2VXk16yT1IvwHLHsybWdLtzsCb7ramr9ps12U4Oa9Skl52dja1bt2LDhg0YM2YM0tLSbtn2bXE/WaPf1sph7vYtyWGLx6yl+8nb2xtJSUkICgpCQUGBpOdek33hnzPIavQX4D/96U9YtWqV4S9rtyueNbGm252DNdlvTdbIwZrstyZr5LBGTb9etPL69eu3vSZr7Cdr9FvuHFK3b0kOWzxmpcavWrUKvXr1Qk5ODnr16oVVq1aZrInsEye9ZDVSL8Byx7Mm1nS7c7Am+63JGjlYk/3WZI0c1qjJw8MDBw4cgEKhQHp6Oqqqqm57TdbYT9bot9w5pG7fkhy2eMxKje/evTueffZZvPXWW5g+fTpcXFxM1kT2iZNeshqpF2C541kTa7rdOViT/dZkjRysyX5rskYOa9S0cuVKeHt7Y/78+bhw4QLi4+Nve03W2E/W6LfcOaRu35IctnjMWpKDuohbtg40kQkajUYUFhaK8vJysWrVKnH8+PHbGs+aWNPtzsGa7Lcma+RgTfZbkzVyWKMmqRxlP0nVVXPY4jFrjf1E9smuHllEREREREREJAVvbyYiIiIiIiKHxUkvEREREREROSxOeomIiIiIiMhhcdJLREREREREDuv/A9Hj7Dpq+8y9AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 1296x1296 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.set(style=\"white\")\n",
    "corr = cb_df.corr()\n",
    "f, ax = plt.subplots(figsize=(18, 18))\n",
    "cmap = sns.diverging_palette(220, 10, as_cmap=True)\n",
    "sns.heatmap(corr, cmap=cmap, vmax=.3, center=0, square=True, linewidths=.5, cbar_kws={\"shrink\": .5})\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>attr1</th>\n",
       "      <th>attr2</th>\n",
       "      <th>attr3</th>\n",
       "      <th>attr4</th>\n",
       "      <th>attr5</th>\n",
       "      <th>attr6</th>\n",
       "      <th>attr7</th>\n",
       "      <th>attr8</th>\n",
       "      <th>attr9</th>\n",
       "      <th>attr10</th>\n",
       "      <th>...</th>\n",
       "      <th>attr56</th>\n",
       "      <th>attr57</th>\n",
       "      <th>attr58</th>\n",
       "      <th>attr59</th>\n",
       "      <th>attr60</th>\n",
       "      <th>attr61</th>\n",
       "      <th>attr62</th>\n",
       "      <th>attr63</th>\n",
       "      <th>attr64</th>\n",
       "      <th>class</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>43405.000000</td>\n",
       "      <td>43405.000000</td>\n",
       "      <td>43405.000000</td>\n",
       "      <td>43405.000000</td>\n",
       "      <td>4.340500e+04</td>\n",
       "      <td>43405.000000</td>\n",
       "      <td>43405.000000</td>\n",
       "      <td>43405.000000</td>\n",
       "      <td>43405.000000</td>\n",
       "      <td>43405.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>4.340500e+04</td>\n",
       "      <td>43405.000000</td>\n",
       "      <td>4.340500e+04</td>\n",
       "      <td>43405.000000</td>\n",
       "      <td>4.340500e+04</td>\n",
       "      <td>43405.000000</td>\n",
       "      <td>4.340500e+04</td>\n",
       "      <td>43405.000000</td>\n",
       "      <td>43405.000000</td>\n",
       "      <td>43405.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>0.035160</td>\n",
       "      <td>0.590212</td>\n",
       "      <td>0.114431</td>\n",
       "      <td>6.314702</td>\n",
       "      <td>-3.853466e+02</td>\n",
       "      <td>-0.056107</td>\n",
       "      <td>0.093478</td>\n",
       "      <td>12.640779</td>\n",
       "      <td>2.652166</td>\n",
       "      <td>0.626868</td>\n",
       "      <td>...</td>\n",
       "      <td>-2.621959e+01</td>\n",
       "      <td>-0.010510</td>\n",
       "      <td>3.002644e+01</td>\n",
       "      <td>1.333288</td>\n",
       "      <td>4.480858e+02</td>\n",
       "      <td>17.033202</td>\n",
       "      <td>1.502328e+03</td>\n",
       "      <td>9.343074</td>\n",
       "      <td>72.788592</td>\n",
       "      <td>0.048174</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>2.993833</td>\n",
       "      <td>5.842209</td>\n",
       "      <td>5.438928</td>\n",
       "      <td>294.978029</td>\n",
       "      <td>6.118020e+04</td>\n",
       "      <td>7.200663</td>\n",
       "      <td>5.712548</td>\n",
       "      <td>505.346176</td>\n",
       "      <td>62.926207</td>\n",
       "      <td>14.669245</td>\n",
       "      <td>...</td>\n",
       "      <td>5.320062e+03</td>\n",
       "      <td>13.672969</td>\n",
       "      <td>5.329289e+03</td>\n",
       "      <td>122.094599</td>\n",
       "      <td>3.153355e+04</td>\n",
       "      <td>552.399187</td>\n",
       "      <td>1.390628e+05</td>\n",
       "      <td>123.985521</td>\n",
       "      <td>2347.072100</td>\n",
       "      <td>0.214137</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>-463.890000</td>\n",
       "      <td>-430.870000</td>\n",
       "      <td>-479.960000</td>\n",
       "      <td>-0.403110</td>\n",
       "      <td>-1.190300e+07</td>\n",
       "      <td>-508.410000</td>\n",
       "      <td>-517.480000</td>\n",
       "      <td>-141.410000</td>\n",
       "      <td>-3.496000</td>\n",
       "      <td>-479.910000</td>\n",
       "      <td>...</td>\n",
       "      <td>-1.108300e+06</td>\n",
       "      <td>-1667.300000</td>\n",
       "      <td>-1.986900e+02</td>\n",
       "      <td>-327.970000</td>\n",
       "      <td>-1.244000e+01</td>\n",
       "      <td>-12.656000</td>\n",
       "      <td>-2.336500e+06</td>\n",
       "      <td>-1.543200</td>\n",
       "      <td>-10677.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>0.003434</td>\n",
       "      <td>0.269010</td>\n",
       "      <td>0.021540</td>\n",
       "      <td>1.050800</td>\n",
       "      <td>-4.949000e+01</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.005783</td>\n",
       "      <td>0.431070</td>\n",
       "      <td>1.018500</td>\n",
       "      <td>0.295530</td>\n",
       "      <td>...</td>\n",
       "      <td>8.959000e-03</td>\n",
       "      <td>0.014632</td>\n",
       "      <td>8.755400e-01</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>5.729400e+00</td>\n",
       "      <td>4.515600</td>\n",
       "      <td>4.220500e+01</td>\n",
       "      <td>3.100200</td>\n",
       "      <td>2.207800</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>0.049636</td>\n",
       "      <td>0.471940</td>\n",
       "      <td>0.196580</td>\n",
       "      <td>1.574000</td>\n",
       "      <td>-1.184000e+00</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.059653</td>\n",
       "      <td>1.074100</td>\n",
       "      <td>1.195500</td>\n",
       "      <td>0.505990</td>\n",
       "      <td>...</td>\n",
       "      <td>5.270100e-02</td>\n",
       "      <td>0.119650</td>\n",
       "      <td>9.511200e-01</td>\n",
       "      <td>0.006389</td>\n",
       "      <td>1.040000e+01</td>\n",
       "      <td>6.649500</td>\n",
       "      <td>7.153500e+01</td>\n",
       "      <td>5.102900</td>\n",
       "      <td>4.389200</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>0.129560</td>\n",
       "      <td>0.688270</td>\n",
       "      <td>0.403260</td>\n",
       "      <td>2.806000</td>\n",
       "      <td>5.046100e+01</td>\n",
       "      <td>0.089441</td>\n",
       "      <td>0.150810</td>\n",
       "      <td>2.639100</td>\n",
       "      <td>2.063000</td>\n",
       "      <td>0.709090</td>\n",
       "      <td>...</td>\n",
       "      <td>1.287800e-01</td>\n",
       "      <td>0.284570</td>\n",
       "      <td>9.928000e-01</td>\n",
       "      <td>0.236280</td>\n",
       "      <td>2.389000e+01</td>\n",
       "      <td>10.443000</td>\n",
       "      <td>1.177400e+02</td>\n",
       "      <td>8.650900</td>\n",
       "      <td>10.461000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>94.280000</td>\n",
       "      <td>480.960000</td>\n",
       "      <td>28.336000</td>\n",
       "      <td>53433.000000</td>\n",
       "      <td>1.250100e+06</td>\n",
       "      <td>543.250000</td>\n",
       "      <td>649.230000</td>\n",
       "      <td>53432.000000</td>\n",
       "      <td>9742.300000</td>\n",
       "      <td>1099.500000</td>\n",
       "      <td>...</td>\n",
       "      <td>2.931500e+02</td>\n",
       "      <td>552.640000</td>\n",
       "      <td>1.108300e+06</td>\n",
       "      <td>23853.000000</td>\n",
       "      <td>4.818700e+06</td>\n",
       "      <td>108000.000000</td>\n",
       "      <td>2.501600e+07</td>\n",
       "      <td>23454.000000</td>\n",
       "      <td>294770.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>8 rows × 65 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "              attr1         attr2         attr3         attr4         attr5  \\\n",
       "count  43405.000000  43405.000000  43405.000000  43405.000000  4.340500e+04   \n",
       "mean       0.035160      0.590212      0.114431      6.314702 -3.853466e+02   \n",
       "std        2.993833      5.842209      5.438928    294.978029  6.118020e+04   \n",
       "min     -463.890000   -430.870000   -479.960000     -0.403110 -1.190300e+07   \n",
       "25%        0.003434      0.269010      0.021540      1.050800 -4.949000e+01   \n",
       "50%        0.049636      0.471940      0.196580      1.574000 -1.184000e+00   \n",
       "75%        0.129560      0.688270      0.403260      2.806000  5.046100e+01   \n",
       "max       94.280000    480.960000     28.336000  53433.000000  1.250100e+06   \n",
       "\n",
       "              attr6         attr7         attr8         attr9        attr10  \\\n",
       "count  43405.000000  43405.000000  43405.000000  43405.000000  43405.000000   \n",
       "mean      -0.056107      0.093478     12.640779      2.652166      0.626868   \n",
       "std        7.200663      5.712548    505.346176     62.926207     14.669245   \n",
       "min     -508.410000   -517.480000   -141.410000     -3.496000   -479.910000   \n",
       "25%        0.000000      0.005783      0.431070      1.018500      0.295530   \n",
       "50%        0.000000      0.059653      1.074100      1.195500      0.505990   \n",
       "75%        0.089441      0.150810      2.639100      2.063000      0.709090   \n",
       "max      543.250000    649.230000  53432.000000   9742.300000   1099.500000   \n",
       "\n",
       "           ...             attr56        attr57        attr58        attr59  \\\n",
       "count      ...       4.340500e+04  43405.000000  4.340500e+04  43405.000000   \n",
       "mean       ...      -2.621959e+01     -0.010510  3.002644e+01      1.333288   \n",
       "std        ...       5.320062e+03     13.672969  5.329289e+03    122.094599   \n",
       "min        ...      -1.108300e+06  -1667.300000 -1.986900e+02   -327.970000   \n",
       "25%        ...       8.959000e-03      0.014632  8.755400e-01      0.000000   \n",
       "50%        ...       5.270100e-02      0.119650  9.511200e-01      0.006389   \n",
       "75%        ...       1.287800e-01      0.284570  9.928000e-01      0.236280   \n",
       "max        ...       2.931500e+02    552.640000  1.108300e+06  23853.000000   \n",
       "\n",
       "             attr60         attr61        attr62        attr63         attr64  \\\n",
       "count  4.340500e+04   43405.000000  4.340500e+04  43405.000000   43405.000000   \n",
       "mean   4.480858e+02      17.033202  1.502328e+03      9.343074      72.788592   \n",
       "std    3.153355e+04     552.399187  1.390628e+05    123.985521    2347.072100   \n",
       "min   -1.244000e+01     -12.656000 -2.336500e+06     -1.543200  -10677.000000   \n",
       "25%    5.729400e+00       4.515600  4.220500e+01      3.100200       2.207800   \n",
       "50%    1.040000e+01       6.649500  7.153500e+01      5.102900       4.389200   \n",
       "75%    2.389000e+01      10.443000  1.177400e+02      8.650900      10.461000   \n",
       "max    4.818700e+06  108000.000000  2.501600e+07  23454.000000  294770.000000   \n",
       "\n",
       "              class  \n",
       "count  43405.000000  \n",
       "mean       0.048174  \n",
       "std        0.214137  \n",
       "min        0.000000  \n",
       "25%        0.000000  \n",
       "50%        0.000000  \n",
       "75%        0.000000  \n",
       "max        1.000000  \n",
       "\n",
       "[8 rows x 65 columns]"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cb_df1 = cb_df.fillna(cb_df.mean())\n",
    "cb_df1.describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We will scale all the features because they are not in the same range. Not doing so might affect the performance of our model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = cb_df1.drop('class', axis=1)\n",
    "y= cb_df1['class']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import StandardScaler\n",
    "ss = StandardScaler()\n",
    "X_std = ss.fit_transform(X)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Split the data using stratify parameter = y to ensure that class distribution for the target is kept constant during the split."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "X_train, X_test, y_train, y_test = train_test_split(X_std, y, test_size=0.2, stratify=y, random_state = 0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Define a cvDictGen function to compare initial models"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import cross_val_score\n",
    "def cvDictGen(functions, scr, X_train, y_train, cv=5, verbose=1):\n",
    "    cvDict = {}\n",
    "    for func in functions:\n",
    "        cvScore = cross_val_score(func, X_train, y_train, cv=cv, verbose=verbose, scoring=scr)\n",
    "        cvDict[str(func).split('(')[0]] = [cvScore.mean(), cvScore.std()]\n",
    "    return cvDict"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Machine Learning"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.ensemble import RandomForestClassifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "lr = LogisticRegression(penalty='l1', dual=False, tol=0.0001, C=1.0, fit_intercept=True, intercept_scaling=1, \n",
    "                        class_weight=None, random_state=None, solver='liblinear', max_iter=100, multi_class='ovr', verbose=2)\n",
    "rf = RandomForestClassifier(n_estimators=10, criterion='gini', max_depth=None, min_samples_split=2, \n",
    "                            min_samples_leaf=1, min_weight_fraction_leaf=0.0, max_features='auto', max_leaf_nodes=None,\n",
    "                            bootstrap=True, oob_score=False, n_jobs=1, random_state=None, verbose=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[LibLinear][LibLinear][LibLinear]"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\Prabodh\\Anaconda2\\envs\\tensorflow\\lib\\site-packages\\sklearn\\svm\\base.py:898: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[LibLinear]"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\Prabodh\\Anaconda2\\envs\\tensorflow\\lib\\site-packages\\sklearn\\svm\\base.py:898: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[LibLinear]"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Done   5 out of   5 | elapsed:  8.4min finished\n",
      "[Parallel(n_jobs=1)]: Done   5 out of   5 | elapsed:   16.9s finished\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "{'LogisticRegression': [0.7009006652287232, 0.014459090110615042],\n",
       " 'RandomForestClassifier': [0.8617584413140886, 0.01221855605486194]}"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cvd = cvDictGen(functions=[lr, rf], X_train = X_train, y_train= y_train, scr='roc_auc')\n",
    "cvd"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Random Forest Classifier\n",
    "rf.fit(X_train, y_train)\n",
    "y_pred = rf.predict(X_test)\n",
    "test_labels = rf.predict_proba(np.array(X_test))[:,1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import accuracy_score, confusion_matrix, precision_score, roc_auc_score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Random Forest\n",
      "ROC AUC Score:  0.8567387217011095\n",
      "Precision: 0.831081081081081\n"
     ]
    }
   ],
   "source": [
    "print('Random Forest')\n",
    "print('ROC AUC Score: ', roc_auc_score(y_test, test_labels))\n",
    "print('Precision:', precision_score(y_test, y_pred))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "def Confusion_Matrix(CMatrix, labels=['operating','bankrupt']):\n",
    "    df = pd.DataFrame(data=CMatrix, index=labels, columns=labels)\n",
    "    df.index.name='TRUE'\n",
    "    df.columns.name='PREDICTION'\n",
    "    df.loc['Total'] = df.sum()\n",
    "    df['Total'] = df.sum(axis=1)\n",
    "    return df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "RF_CMatrix = confusion_matrix(y_test, y_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>PREDICTION</th>\n",
       "      <th>operating</th>\n",
       "      <th>bankrupt</th>\n",
       "      <th>Total</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TRUE</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>operating</th>\n",
       "      <td>8238</td>\n",
       "      <td>25</td>\n",
       "      <td>8263</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>bankrupt</th>\n",
       "      <td>295</td>\n",
       "      <td>123</td>\n",
       "      <td>418</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Total</th>\n",
       "      <td>8533</td>\n",
       "      <td>148</td>\n",
       "      <td>8681</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "PREDICTION  operating  bankrupt  Total\n",
       "TRUE                                  \n",
       "operating        8238        25   8263\n",
       "bankrupt          295       123    418\n",
       "Total            8533       148   8681"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Confusion_Matrix(RF_CMatrix)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### XGBoost"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "from xgboost.sklearn import XGBClassifier\n",
    "xgb1 = XGBClassifier(\n",
    " learning_rate =0.1,\n",
    " n_estimators=1000,\n",
    " max_depth=5,\n",
    " min_child_weight=1,\n",
    " gamma=0,\n",
    " subsample=0.8,\n",
    " colsample_bytree=0.8,\n",
    " objective= 'binary:logistic',\n",
    " nthread=4,\n",
    " scale_pos_weight=35, #obtained from trial and error\n",
    " seed=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "XGBClassifier(base_score=0.5, booster='gbtree', colsample_bylevel=1,\n",
       "       colsample_bytree=0.8, gamma=0, learning_rate=0.1, max_delta_step=0,\n",
       "       max_depth=5, min_child_weight=1, missing=None, n_estimators=1000,\n",
       "       n_jobs=1, nthread=4, objective='binary:logistic', random_state=0,\n",
       "       reg_alpha=0, reg_lambda=1, scale_pos_weight=35, seed=0, silent=True,\n",
       "       subsample=0.8)"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "xgb1.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\Prabodh\\Anaconda2\\envs\\tensorflow\\lib\\site-packages\\sklearn\\preprocessing\\label.py:151: DeprecationWarning: The truth value of an empty array is ambiguous. Returning False, but in future this will result in an error. Use `array.size > 0` to check that an array is not empty.\n",
      "  if diff:\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>PREDICTION</th>\n",
       "      <th>operating</th>\n",
       "      <th>bankrupt</th>\n",
       "      <th>Total</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TRUE</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>operating</th>\n",
       "      <td>8207</td>\n",
       "      <td>56</td>\n",
       "      <td>8263</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>bankrupt</th>\n",
       "      <td>141</td>\n",
       "      <td>277</td>\n",
       "      <td>418</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Total</th>\n",
       "      <td>8348</td>\n",
       "      <td>333</td>\n",
       "      <td>8681</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "PREDICTION  operating  bankrupt  Total\n",
       "TRUE                                  \n",
       "operating        8207        56   8263\n",
       "bankrupt          141       277    418\n",
       "Total            8348       333   8681"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_pred_xgb = xgb1.predict(X_test)\n",
    "predictions = [round(value) for value in y_pred_xgb]\n",
    "Confusion_Matrix(confusion_matrix(y_test, predictions))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8279511131364988"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "roc_auc_score(y_test, predictions)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "xgb2 = XGBClassifier(\n",
    " learning_rate =0.05,\n",
    " n_estimators=1400,\n",
    " max_depth=3,\n",
    " min_child_weight=10,\n",
    " gamma=1.5,\n",
    " subsample=0.6,\n",
    " colsample_bytree=1.0,\n",
    " objective= 'binary:logistic',\n",
    " nthread=8,\n",
    " scale_pos_weight=35, #obtained from trial and error\n",
    " seed=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "XGBClassifier(base_score=0.5, booster='gbtree', colsample_bylevel=1,\n",
       "       colsample_bytree=1.0, gamma=1.5, learning_rate=0.05,\n",
       "       max_delta_step=0, max_depth=3, min_child_weight=10, missing=None,\n",
       "       n_estimators=1400, n_jobs=1, nthread=8, objective='binary:logistic',\n",
       "       random_state=0, reg_alpha=0, reg_lambda=1, scale_pos_weight=35,\n",
       "       seed=0, silent=True, subsample=0.6)"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "xgb2.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\Prabodh\\Anaconda2\\envs\\tensorflow\\lib\\site-packages\\sklearn\\preprocessing\\label.py:151: DeprecationWarning: The truth value of an empty array is ambiguous. Returning False, but in future this will result in an error. Use `array.size > 0` to check that an array is not empty.\n",
      "  if diff:\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>PREDICTION</th>\n",
       "      <th>operating</th>\n",
       "      <th>bankrupt</th>\n",
       "      <th>Total</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TRUE</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>operating</th>\n",
       "      <td>7843</td>\n",
       "      <td>420</td>\n",
       "      <td>8263</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>bankrupt</th>\n",
       "      <td>79</td>\n",
       "      <td>339</td>\n",
       "      <td>418</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Total</th>\n",
       "      <td>7922</td>\n",
       "      <td>759</td>\n",
       "      <td>8681</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "PREDICTION  operating  bankrupt  Total\n",
       "TRUE                                  \n",
       "operating        7843       420   8263\n",
       "bankrupt           79       339    418\n",
       "Total            7922       759   8681"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_pred_xgb_best = xgb2.predict(X_test)\n",
    "predictions_best = [round(value) for value in y_pred_xgb_best]\n",
    "Confusion_Matrix(confusion_matrix(y_test, predictions_best))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8800878939782868"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "roc_auc_score(y_test, predictions_best)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "features_label = X.columns\n",
    "feature_list = dict(zip(features_label, xgb2.feature_importances_*100))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['attr27',\n",
       " 'attr46',\n",
       " 'attr5',\n",
       " 'attr58',\n",
       " 'attr34',\n",
       " 'attr24',\n",
       " 'attr29',\n",
       " 'attr55',\n",
       " 'attr25',\n",
       " 'attr21',\n",
       " 'attr36']"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "top_ten_features = sorted(feature_list, key=feature_list.__getitem__, reverse= True)[:11]\n",
    "top_ten_features"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* **X27**: profit on operating activities / financial expenses\n",
    "* **X46**: (current assets - inventory) / short-term liabilities\n",
    "* **X5**: [(cash + short-term securities + receivables - short-term liabilities) / (operating expenses - depreciation)] * 365\n",
    "* **X58**: total costs /total sales\n",
    "* **X34**: operating expenses / total liabilities\n",
    "* **X24**: gross profit (in 3 years) / total assets\n",
    "* **X29**: logarithm of total assets\n",
    "* **X55**: working capital\n",
    "* **X25**: (equity - share capital) / total assets\n",
    "* **X21**: sales (n) / sales (n-1)\n",
    "* **X36**: total sales / total assets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "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": "3.5.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}