Arcene Cancer Classification.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# **ARCENE Cancer Classification**\n",
    "\n",
    "**Dataset:** ARCENE dataset available at https://archive.ics.uci.edu/ml/datasets/Arcene\n",
    "\n",
    "**Code:** Chapter 5 SVM from `Hands-On Machine Learning with Scikit-Learn and TensorFlow` book (Aurélien Géron)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Common imports\n",
    "import numpy as np\n",
    "import os\n",
    "\n",
    "import pandas as pd\n",
    "\n",
    "%matplotlib inline\n",
    "import matplotlib as mpl\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Get the data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Classify the healthy vs. cancer patients."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Dataset as described on UCI:\n",
    "ARCENE was obtained by merging three mass-spectrometry datasets to obtain enough training and test data for a benchmark. The original features indicate the abundance of proteins in human sera having a given mass value. Based on those features one must separate cancer patients from healthy patients. We added a number of distractor feature called 'probes' having no predictive power. The order of the features and patterns were randomized.\n",
    "\n",
    "<table>\n",
    "    <thead>\n",
    "        <tr>\n",
    "            <th colspan=\"4\">ARCENE Dataset</th>\n",
    "        </tr>\n",
    "    </thead>\n",
    "    <tbody>\n",
    "        <tr>\n",
    "            <td>ARCENE</td>\n",
    "            <td>Positive ex</td>\n",
    "            <td>Negative ex</td>\n",
    "            <td>Total</td>\n",
    "        </tr>\n",
    "         <tr>\n",
    "            <td>Training set</td>\n",
    "            <td>44</td>\n",
    "            <td>56</td>\n",
    "            <td>100</td>\n",
    "        </tr>\n",
    "        <tr>\n",
    "            <td>Validation set</td>\n",
    "            <td>44</td>\n",
    "            <td>56</td>\n",
    "            <td>100</td>\n",
    "        </tr>\n",
    "         <tr>\n",
    "            <td>Test set</td>\n",
    "            <td>310</td>\n",
    "            <td>390</td>\n",
    "            <td>700</td>\n",
    "        </tr>\n",
    "         <tr>\n",
    "            <td>All</td>\n",
    "            <td>398</td>\n",
    "            <td>502</td>\n",
    "            <td>900</td>\n",
    "        </tr>\n",
    "    </tbody>\n",
    "</table>\n",
    "\n",
    "Number of variables/features/attributes: Real: 7000    Probes: 3000   Total: 10000\n",
    "\n",
    "This dataset is one of five datasets used in the NIPS 2003 feature selection challenge."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "ARCENE_PATH = \"dataset\"\n",
    "\n",
    "def load_arcene_data(arcene_path=ARCENE_PATH, data_file=\"data.csv\"):\n",
    "    csv_path = os.path.join(arcene_path, data_file)\n",
    "    return pd.read_csv(csv_path, sep='\\s+', header=None)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "      <th>...</th>\n",
       "      <th>9990</th>\n",
       "      <th>9991</th>\n",
       "      <th>9992</th>\n",
       "      <th>9993</th>\n",
       "      <th>9994</th>\n",
       "      <th>9995</th>\n",
       "      <th>9996</th>\n",
       "      <th>9997</th>\n",
       "      <th>9998</th>\n",
       "      <th>9999</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>71</td>\n",
       "      <td>0</td>\n",
       "      <td>95</td>\n",
       "      <td>0</td>\n",
       "      <td>538</td>\n",
       "      <td>404</td>\n",
       "      <td>20</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>255</td>\n",
       "      <td>570</td>\n",
       "      <td>86</td>\n",
       "      <td>0</td>\n",
       "      <td>36</td>\n",
       "      <td>0</td>\n",
       "      <td>80</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>524</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>41</td>\n",
       "      <td>82</td>\n",
       "      <td>165</td>\n",
       "      <td>60</td>\n",
       "      <td>554</td>\n",
       "      <td>379</td>\n",
       "      <td>0</td>\n",
       "      <td>71</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>213</td>\n",
       "      <td>605</td>\n",
       "      <td>69</td>\n",
       "      <td>7</td>\n",
       "      <td>473</td>\n",
       "      <td>0</td>\n",
       "      <td>57</td>\n",
       "      <td>0</td>\n",
       "      <td>284</td>\n",
       "      <td>423</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>40</td>\n",
       "      <td>0</td>\n",
       "      <td>451</td>\n",
       "      <td>402</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>235</td>\n",
       "      <td>593</td>\n",
       "      <td>28</td>\n",
       "      <td>0</td>\n",
       "      <td>24</td>\n",
       "      <td>0</td>\n",
       "      <td>90</td>\n",
       "      <td>0</td>\n",
       "      <td>34</td>\n",
       "      <td>508</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>56</td>\n",
       "      <td>44</td>\n",
       "      <td>275</td>\n",
       "      <td>14</td>\n",
       "      <td>511</td>\n",
       "      <td>470</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>91</td>\n",
       "      <td>600</td>\n",
       "      <td>0</td>\n",
       "      <td>26</td>\n",
       "      <td>86</td>\n",
       "      <td>0</td>\n",
       "      <td>102</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>469</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>105</td>\n",
       "      <td>0</td>\n",
       "      <td>141</td>\n",
       "      <td>348</td>\n",
       "      <td>0</td>\n",
       "      <td>268</td>\n",
       "      <td>329</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>...</td>\n",
       "      <td>813</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>190</td>\n",
       "      <td>301</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>354</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 10000 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   0     1     2     3     4     5     6     7     8     9     ...  9990  \\\n",
       "0     0    71     0    95     0   538   404    20     0     0  ...   255   \n",
       "1     0    41    82   165    60   554   379     0    71     0  ...   213   \n",
       "2     0     0     1    40     0   451   402     0     0     0  ...   235   \n",
       "3     0    56    44   275    14   511   470     0     0     0  ...    91   \n",
       "4   105     0   141   348     0   268   329     0     0     1  ...   813   \n",
       "\n",
       "   9991  9992  9993  9994  9995  9996  9997  9998  9999  \n",
       "0   570    86     0    36     0    80     0     0   524  \n",
       "1   605    69     7   473     0    57     0   284   423  \n",
       "2   593    28     0    24     0    90     0    34   508  \n",
       "3   600     0    26    86     0   102     0     0   469  \n",
       "4     0     0     0     0   190   301     0     0   354  \n",
       "\n",
       "[5 rows x 10000 columns]"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train = load_arcene_data(data_file=\"arcene_train.data\")\n",
    "X_train.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 100 entries, 0 to 99\n",
      "Columns: 10000 entries, 0 to 9999\n",
      "dtypes: int64(10000)\n",
      "memory usage: 7.6 MB\n"
     ]
    }
   ],
   "source": [
    "X_train.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "      <th>...</th>\n",
       "      <th>9990</th>\n",
       "      <th>9991</th>\n",
       "      <th>9992</th>\n",
       "      <th>9993</th>\n",
       "      <th>9994</th>\n",
       "      <th>9995</th>\n",
       "      <th>9996</th>\n",
       "      <th>9997</th>\n",
       "      <th>9998</th>\n",
       "      <th>9999</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>100.000000</td>\n",
       "      <td>100.000000</td>\n",
       "      <td>100.00000</td>\n",
       "      <td>100.000000</td>\n",
       "      <td>100.000000</td>\n",
       "      <td>100.000000</td>\n",
       "      <td>100.000000</td>\n",
       "      <td>100.000000</td>\n",
       "      <td>100.000000</td>\n",
       "      <td>100.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>100.000000</td>\n",
       "      <td>100.000000</td>\n",
       "      <td>100.000000</td>\n",
       "      <td>100.000000</td>\n",
       "      <td>100.000000</td>\n",
       "      <td>100.000000</td>\n",
       "      <td>100.000000</td>\n",
       "      <td>100.000000</td>\n",
       "      <td>100.000000</td>\n",
       "      <td>100.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>37.060000</td>\n",
       "      <td>35.330000</td>\n",
       "      <td>56.50000</td>\n",
       "      <td>329.860000</td>\n",
       "      <td>17.110000</td>\n",
       "      <td>412.090000</td>\n",
       "      <td>232.330000</td>\n",
       "      <td>6.510000</td>\n",
       "      <td>48.170000</td>\n",
       "      <td>63.690000</td>\n",
       "      <td>...</td>\n",
       "      <td>363.780000</td>\n",
       "      <td>396.480000</td>\n",
       "      <td>45.920000</td>\n",
       "      <td>18.210000</td>\n",
       "      <td>106.070000</td>\n",
       "      <td>165.350000</td>\n",
       "      <td>197.650000</td>\n",
       "      <td>0.360000</td>\n",
       "      <td>88.530000</td>\n",
       "      <td>365.750000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>49.763647</td>\n",
       "      <td>37.690943</td>\n",
       "      <td>68.64011</td>\n",
       "      <td>164.264508</td>\n",
       "      <td>26.348057</td>\n",
       "      <td>108.375152</td>\n",
       "      <td>186.106142</td>\n",
       "      <td>13.149102</td>\n",
       "      <td>57.624009</td>\n",
       "      <td>91.626349</td>\n",
       "      <td>...</td>\n",
       "      <td>262.439741</td>\n",
       "      <td>233.176462</td>\n",
       "      <td>47.886161</td>\n",
       "      <td>23.914112</td>\n",
       "      <td>141.747054</td>\n",
       "      <td>145.380752</td>\n",
       "      <td>113.122118</td>\n",
       "      <td>2.110675</td>\n",
       "      <td>109.698619</td>\n",
       "      <td>106.055386</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>27.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>181.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>61.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>41.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>213.750000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>314.250000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>192.500000</td>\n",
       "      <td>309.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>94.500000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>314.500000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>15.000000</td>\n",
       "      <td>25.000000</td>\n",
       "      <td>19.50000</td>\n",
       "      <td>333.500000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>441.000000</td>\n",
       "      <td>266.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>30.000000</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>...</td>\n",
       "      <td>250.000000</td>\n",
       "      <td>468.500000</td>\n",
       "      <td>34.500000</td>\n",
       "      <td>2.500000</td>\n",
       "      <td>48.500000</td>\n",
       "      <td>186.000000</td>\n",
       "      <td>167.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>38.500000</td>\n",
       "      <td>364.500000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>57.500000</td>\n",
       "      <td>59.750000</td>\n",
       "      <td>107.25000</td>\n",
       "      <td>443.000000</td>\n",
       "      <td>29.000000</td>\n",
       "      <td>501.250000</td>\n",
       "      <td>422.750000</td>\n",
       "      <td>4.500000</td>\n",
       "      <td>79.250000</td>\n",
       "      <td>125.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>482.000000</td>\n",
       "      <td>579.250000</td>\n",
       "      <td>75.250000</td>\n",
       "      <td>34.250000</td>\n",
       "      <td>140.000000</td>\n",
       "      <td>276.500000</td>\n",
       "      <td>295.750000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>165.500000</td>\n",
       "      <td>446.500000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>188.000000</td>\n",
       "      <td>218.000000</td>\n",
       "      <td>236.00000</td>\n",
       "      <td>677.000000</td>\n",
       "      <td>99.000000</td>\n",
       "      <td>583.000000</td>\n",
       "      <td>493.000000</td>\n",
       "      <td>55.000000</td>\n",
       "      <td>202.000000</td>\n",
       "      <td>391.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>863.000000</td>\n",
       "      <td>654.000000</td>\n",
       "      <td>195.000000</td>\n",
       "      <td>83.000000</td>\n",
       "      <td>656.000000</td>\n",
       "      <td>485.000000</td>\n",
       "      <td>422.000000</td>\n",
       "      <td>18.000000</td>\n",
       "      <td>392.000000</td>\n",
       "      <td>533.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>8 rows × 10000 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "             0           1          2           3           4           5     \\\n",
       "count  100.000000  100.000000  100.00000  100.000000  100.000000  100.000000   \n",
       "mean    37.060000   35.330000   56.50000  329.860000   17.110000  412.090000   \n",
       "std     49.763647   37.690943   68.64011  164.264508   26.348057  108.375152   \n",
       "min      0.000000    0.000000    0.00000   27.000000    0.000000  181.000000   \n",
       "25%      0.000000    0.000000    0.00000  213.750000    0.000000  314.250000   \n",
       "50%     15.000000   25.000000   19.50000  333.500000    0.000000  441.000000   \n",
       "75%     57.500000   59.750000  107.25000  443.000000   29.000000  501.250000   \n",
       "max    188.000000  218.000000  236.00000  677.000000   99.000000  583.000000   \n",
       "\n",
       "             6           7           8           9     ...        9990  \\\n",
       "count  100.000000  100.000000  100.000000  100.000000  ...  100.000000   \n",
       "mean   232.330000    6.510000   48.170000   63.690000  ...  363.780000   \n",
       "std    186.106142   13.149102   57.624009   91.626349  ...  262.439741   \n",
       "min      0.000000    0.000000    0.000000    0.000000  ...   61.000000   \n",
       "25%      0.000000    0.000000    0.000000    0.000000  ...  192.500000   \n",
       "50%    266.000000    0.000000   30.000000    0.500000  ...  250.000000   \n",
       "75%    422.750000    4.500000   79.250000  125.000000  ...  482.000000   \n",
       "max    493.000000   55.000000  202.000000  391.000000  ...  863.000000   \n",
       "\n",
       "             9991        9992        9993        9994        9995        9996  \\\n",
       "count  100.000000  100.000000  100.000000  100.000000  100.000000  100.000000   \n",
       "mean   396.480000   45.920000   18.210000  106.070000  165.350000  197.650000   \n",
       "std    233.176462   47.886161   23.914112  141.747054  145.380752  113.122118   \n",
       "min      0.000000    0.000000    0.000000    0.000000    0.000000   41.000000   \n",
       "25%    309.000000    0.000000    0.000000    0.000000    0.000000   94.500000   \n",
       "50%    468.500000   34.500000    2.500000   48.500000  186.000000  167.000000   \n",
       "75%    579.250000   75.250000   34.250000  140.000000  276.500000  295.750000   \n",
       "max    654.000000  195.000000   83.000000  656.000000  485.000000  422.000000   \n",
       "\n",
       "             9997        9998        9999  \n",
       "count  100.000000  100.000000  100.000000  \n",
       "mean     0.360000   88.530000  365.750000  \n",
       "std      2.110675  109.698619  106.055386  \n",
       "min      0.000000    0.000000    0.000000  \n",
       "25%      0.000000    0.000000  314.500000  \n",
       "50%      0.000000   38.500000  364.500000  \n",
       "75%      0.000000  165.500000  446.500000  \n",
       "max     18.000000  392.000000  533.000000  \n",
       "\n",
       "[8 rows x 10000 columns]"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>-1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>-1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   0\n",
       "0  1\n",
       "1 -1\n",
       "2  1\n",
       "3  1\n",
       "4 -1"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_train = load_arcene_data(data_file=\"arcene_train.labels\")\n",
    "y_train.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "-1    56\n",
       " 1    44\n",
       "Name: 0, dtype: int64"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_train[0].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_train.rename( columns={0:'label'}, inplace=True )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Data visualization"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Join train attributes and labels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train_with_labels = pd.concat([X_train, y_train], axis=1, sort=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "      <th>...</th>\n",
       "      <th>9991</th>\n",
       "      <th>9992</th>\n",
       "      <th>9993</th>\n",
       "      <th>9994</th>\n",
       "      <th>9995</th>\n",
       "      <th>9996</th>\n",
       "      <th>9997</th>\n",
       "      <th>9998</th>\n",
       "      <th>9999</th>\n",
       "      <th>label</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>71</td>\n",
       "      <td>0</td>\n",
       "      <td>95</td>\n",
       "      <td>0</td>\n",
       "      <td>538</td>\n",
       "      <td>404</td>\n",
       "      <td>20</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>570</td>\n",
       "      <td>86</td>\n",
       "      <td>0</td>\n",
       "      <td>36</td>\n",
       "      <td>0</td>\n",
       "      <td>80</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>524</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>41</td>\n",
       "      <td>82</td>\n",
       "      <td>165</td>\n",
       "      <td>60</td>\n",
       "      <td>554</td>\n",
       "      <td>379</td>\n",
       "      <td>0</td>\n",
       "      <td>71</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>605</td>\n",
       "      <td>69</td>\n",
       "      <td>7</td>\n",
       "      <td>473</td>\n",
       "      <td>0</td>\n",
       "      <td>57</td>\n",
       "      <td>0</td>\n",
       "      <td>284</td>\n",
       "      <td>423</td>\n",
       "      <td>-1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>40</td>\n",
       "      <td>0</td>\n",
       "      <td>451</td>\n",
       "      <td>402</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>593</td>\n",
       "      <td>28</td>\n",
       "      <td>0</td>\n",
       "      <td>24</td>\n",
       "      <td>0</td>\n",
       "      <td>90</td>\n",
       "      <td>0</td>\n",
       "      <td>34</td>\n",
       "      <td>508</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>56</td>\n",
       "      <td>44</td>\n",
       "      <td>275</td>\n",
       "      <td>14</td>\n",
       "      <td>511</td>\n",
       "      <td>470</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>600</td>\n",
       "      <td>0</td>\n",
       "      <td>26</td>\n",
       "      <td>86</td>\n",
       "      <td>0</td>\n",
       "      <td>102</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>469</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>105</td>\n",
       "      <td>0</td>\n",
       "      <td>141</td>\n",
       "      <td>348</td>\n",
       "      <td>0</td>\n",
       "      <td>268</td>\n",
       "      <td>329</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>190</td>\n",
       "      <td>301</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>354</td>\n",
       "      <td>-1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 10001 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "     0   1    2    3   4    5    6   7   8  9  ...  9991  9992  9993  9994  \\\n",
       "0    0  71    0   95   0  538  404  20   0  0  ...   570    86     0    36   \n",
       "1    0  41   82  165  60  554  379   0  71  0  ...   605    69     7   473   \n",
       "2    0   0    1   40   0  451  402   0   0  0  ...   593    28     0    24   \n",
       "3    0  56   44  275  14  511  470   0   0  0  ...   600     0    26    86   \n",
       "4  105   0  141  348   0  268  329   0   0  1  ...     0     0     0     0   \n",
       "\n",
       "   9995  9996  9997  9998  9999  label  \n",
       "0     0    80     0     0   524      1  \n",
       "1     0    57     0   284   423     -1  \n",
       "2     0    90     0    34   508      1  \n",
       "3     0   102     0     0   469      1  \n",
       "4   190   301     0     0   354     -1  \n",
       "\n",
       "[5 rows x 10001 columns]"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train_with_labels.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "corr_matrix = X_train_with_labels.corr()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "label    1.000000\n",
       "2643     0.375672\n",
       "9278     0.371418\n",
       "7733     0.370503\n",
       "6480     0.368229\n",
       "           ...   \n",
       "9165          NaN\n",
       "9356          NaN\n",
       "9538          NaN\n",
       "9895          NaN\n",
       "9913          NaN\n",
       "Name: label, Length: 10001, dtype: float64"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "corr_matrix[\"label\"].sort_values(ascending=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train_no_cancer = pd.DataFrame(X_train, columns=X_train.columns, index=X_train.index[(X_train_with_labels[\"label\"] == -1)])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train_cancer = pd.DataFrame(X_train, columns=X_train.columns, index=X_train.index[(X_train_with_labels[\"label\"] == 1)])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x292621cd508>"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 792x504 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "X_train_no_cancer.T.plot.line(legend=False,  figsize=(11,7), title=\"Attribute values for the patients without cancer\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x29264d33f48>"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 792x504 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "X_train_cancer.T.plot.line(legend=False,  figsize=(11,7), title=\"Attribute values for the patients with cancer\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Read the valid dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "      <th>...</th>\n",
       "      <th>9990</th>\n",
       "      <th>9991</th>\n",
       "      <th>9992</th>\n",
       "      <th>9993</th>\n",
       "      <th>9994</th>\n",
       "      <th>9995</th>\n",
       "      <th>9996</th>\n",
       "      <th>9997</th>\n",
       "      <th>9998</th>\n",
       "      <th>9999</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>156</td>\n",
       "      <td>138</td>\n",
       "      <td>2</td>\n",
       "      <td>635</td>\n",
       "      <td>444</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>88</td>\n",
       "      <td>692</td>\n",
       "      <td>42</td>\n",
       "      <td>0</td>\n",
       "      <td>50</td>\n",
       "      <td>67</td>\n",
       "      <td>87</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>465</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>7</td>\n",
       "      <td>0</td>\n",
       "      <td>7</td>\n",
       "      <td>0</td>\n",
       "      <td>251</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>22</td>\n",
       "      <td>222</td>\n",
       "      <td>...</td>\n",
       "      <td>698</td>\n",
       "      <td>673</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>18</td>\n",
       "      <td>201</td>\n",
       "      <td>317</td>\n",
       "      <td>0</td>\n",
       "      <td>34</td>\n",
       "      <td>199</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>32</td>\n",
       "      <td>0</td>\n",
       "      <td>470</td>\n",
       "      <td>53</td>\n",
       "      <td>493</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>140</td>\n",
       "      <td>147</td>\n",
       "      <td>...</td>\n",
       "      <td>267</td>\n",
       "      <td>461</td>\n",
       "      <td>20</td>\n",
       "      <td>34</td>\n",
       "      <td>165</td>\n",
       "      <td>225</td>\n",
       "      <td>53</td>\n",
       "      <td>0</td>\n",
       "      <td>47</td>\n",
       "      <td>219</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>77</td>\n",
       "      <td>0</td>\n",
       "      <td>202</td>\n",
       "      <td>0</td>\n",
       "      <td>432</td>\n",
       "      <td>472</td>\n",
       "      <td>0</td>\n",
       "      <td>60</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>181</td>\n",
       "      <td>555</td>\n",
       "      <td>13</td>\n",
       "      <td>0</td>\n",
       "      <td>488</td>\n",
       "      <td>0</td>\n",
       "      <td>71</td>\n",
       "      <td>0</td>\n",
       "      <td>177</td>\n",
       "      <td>487</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>34</td>\n",
       "      <td>102</td>\n",
       "      <td>522</td>\n",
       "      <td>2</td>\n",
       "      <td>474</td>\n",
       "      <td>473</td>\n",
       "      <td>0</td>\n",
       "      <td>39</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>114</td>\n",
       "      <td>630</td>\n",
       "      <td>0</td>\n",
       "      <td>29</td>\n",
       "      <td>192</td>\n",
       "      <td>0</td>\n",
       "      <td>74</td>\n",
       "      <td>0</td>\n",
       "      <td>5</td>\n",
       "      <td>416</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 10000 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   0     1     2     3     4     5     6     7     8     9     ...  9990  \\\n",
       "0     0     0   156   138     2   635   444     0     1     0  ...    88   \n",
       "1     0     7     0     7     0   251     0     0    22   222  ...   698   \n",
       "2     0    32     0   470    53   493     0     0   140   147  ...   267   \n",
       "3     0    77     0   202     0   432   472     0    60     0  ...   181   \n",
       "4     0    34   102   522     2   474   473     0    39     0  ...   114   \n",
       "\n",
       "   9991  9992  9993  9994  9995  9996  9997  9998  9999  \n",
       "0   692    42     0    50    67    87     0     0   465  \n",
       "1   673     0     0    18   201   317     0    34   199  \n",
       "2   461    20    34   165   225    53     0    47   219  \n",
       "3   555    13     0   488     0    71     0   177   487  \n",
       "4   630     0    29   192     0    74     0     5   416  \n",
       "\n",
       "[5 rows x 10000 columns]"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_valid = load_arcene_data(data_file=\"arcene_valid.data\")\n",
    "X_valid.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_valid = load_arcene_data(data_file=\"arcene_valid.labels\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Data preparation"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Shuffle the training data:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "np.random.seed(42)\n",
    "\n",
    "shuffle_index = np.random.permutation(X_train.shape[0])\n",
    "X_train, y_train = X_train.iloc[shuffle_index], y_train.iloc[shuffle_index]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Expecting no incomplete rows, and **no imputation** later in the pipeline:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "      <th>...</th>\n",
       "      <th>9990</th>\n",
       "      <th>9991</th>\n",
       "      <th>9992</th>\n",
       "      <th>9993</th>\n",
       "      <th>9994</th>\n",
       "      <th>9995</th>\n",
       "      <th>9996</th>\n",
       "      <th>9997</th>\n",
       "      <th>9998</th>\n",
       "      <th>9999</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>0 rows × 10000 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "Empty DataFrame\n",
       "Columns: [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, ...]\n",
       "Index: []\n",
       "\n",
       "[0 rows x 10000 columns]"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sample_incomplete_rows = X_train[X_train.isnull().any(axis=1)].head()\n",
    "sample_incomplete_rows"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.pipeline import Pipeline\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "\n",
    "scaler_pipeline = Pipeline([\n",
    "        ('std_scaler', StandardScaler()),\n",
    "    ])\n",
    "\n",
    "X_train_prepared = scaler_pipeline.fit_transform(X_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Binary Classification"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Train a `Stochastic Gradient Descent (SGD) Classifier`:**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_train = y_train.values.ravel()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "SGDClassifier(alpha=0.0001, average=False, class_weight=None,\n",
       "              early_stopping=False, epsilon=0.1, eta0=0.0, fit_intercept=True,\n",
       "              l1_ratio=0.15, learning_rate='optimal', loss='hinge',\n",
       "              max_iter=1000, n_iter_no_change=5, n_jobs=None, penalty='l2',\n",
       "              power_t=0.5, random_state=42, shuffle=True, tol=0.001,\n",
       "              validation_fraction=0.1, verbose=0, warm_start=False)"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.linear_model import SGDClassifier\n",
    "\n",
    "sgd_clf = SGDClassifier(random_state=42)\n",
    "sgd_clf.fit(X_train_prepared, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_valid_predict = sgd_clf.predict(X_valid)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Performance of the model**:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[12, 44],\n",
       "       [ 0, 44]], dtype=int64)"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix, precision_score, recall_score, f1_score\n",
    "\n",
    "confusion_matrix(y_valid, y_valid_predict)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "def performance_scores(y, y_predict):\n",
    "    print(\"Precision score:\\t\", precision_score(y, y_predict))\n",
    "    print(\"Recall score:\\t\\t\", recall_score(y, y_predict))   \n",
    "    print(\"F1 score:\\t\\t\", f1_score(y, y_predict))   \n",
    "    return"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Precision score:\t 0.5\n",
      "Recall score:\t\t 1.0\n",
      "F1 score:\t\t 0.6666666666666666\n"
     ]
    }
   ],
   "source": [
    "performance_scores(y_valid, y_valid_predict)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Recall vs. Precision**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import cross_val_predict\n",
    "from sklearn.metrics import precision_recall_curve\n",
    "\n",
    "precisions, recalls, thresholds = precision_recall_curve(y_valid, y_valid_predict)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "def show_recall_vs_precision_curve(recalls, precisions, label=None):\n",
    "    plt.plot(recalls, precisions, label=label)\n",
    "    plt.axis([0, 1, 0, 1])\n",
    "    plt.xlabel(\"Recall\")\n",
    "    plt.ylabel(\"Precision\")\n",
    "    plt.show()\n",
    "    return"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "show_recall_vs_precision_curve(recalls, precisions)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "def show_threshold_vs_recall_precision_curve(thresholds, recalls, precisions):\n",
    "    plt.plot(thresholds, precisions[:-1], \"b--\", label=\"Precision\")\n",
    "    plt.plot(thresholds, recalls[:-1], \"g-\", label=\"Recall/Sensitivity\")\n",
    "    plt.xlabel(\"Threshold\")\n",
    "    plt.legend(loc=\"lower left\")\n",
    "    plt.ylim([0, 1])\n",
    "    plt.show()\n",
    "    return"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "show_threshold_vs_recall_precision_curve(thresholds, recalls, precisions)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**ROC Curve**:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import roc_curve\n",
    "\n",
    "fpr, tpr, thresholds = roc_curve(y_valid, y_valid_predict)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "def show_roc_curve(fpr, tpr, label=None):\n",
    "    plt.plot(fpr, tpr, label=None)\n",
    "    plt.plot([0, 1], [0, 1], \"k--\")\n",
    "    plt.axis([0, 1, 0, 1])\n",
    "    plt.xlabel(\"FPR / 1 - TNR / 1 - specificity\")\n",
    "    plt.ylabel(\"TPR / sensitivity\")\n",
    "    plt.show()\n",
    "    return"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYoAAAEKCAYAAAAMzhLIAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjMsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+AADFEAAAgAElEQVR4nO3dd3gU5drH8e9N70VRkY6IQOiw1NCDSFFAQZqEUJJQDxxRj/0Vy9GDDQ+K1IiAFGlSlANYEARpaZSEHiQJvfeElOf9YxeMMYQlZGvuz3V5md2dnbkZyP52npm5HzHGoJRSSt1OLlcXoJRSyr1pUCillMqUBoVSSqlMaVAopZTKlAaFUkqpTGlQKKWUypTDgkJEvhKRUyKy+zavi4hMFJGDIrJTRBo4qhallFJZ58gjiq+Bjpm83gmoavsvGJjswFqUUkplkcOCwhizATiXySLdgNnGagtQQkQedlQ9SimlsiaPC7ddFohL8zje9tzx9AuKSDDWow4KFy7csHr16k4pUCnl/i4nJBN77hqp2mUiQ8kXT5GaeBVSU84YYx7IyjpcGRSSwXMZ/k0bY6YB0wAsFosJDQ11ZF1KKQ9gjCFk42HeX7WHNqWL8WmvupQolNfVZbmFm62ZRIRZIdM4c/o0n/zn30eyuj5XBkU8UD7N43LAMRfVopTyIInJKbzx3W4WhcXTsWZpPulVl8L5Xflx5j6OHj3K8OHD6d27N8899xyvjB0DwCf/+XeW1+nKy2NXAANsVz81BS4aY/427KSUUmmduZLIc9O3sigsntF+VfnyuQYaEliPIqZPn46Pjw8//fQTV65cybZ1O2zvish8oA1QSkTigbeAvADGmCnAKqAzcBC4BgxyVC1KKe8QfewSQbNDOXMlkc/71uepumVcXZJbOHToEEFBQaxbt462bdsyffp0qlSpkm3rd1hQGGP63uF1A4x01PaVUt5l9e4TjF0YSbECeVk8rDm1yxV3dUluY9euXYSFhTFt2jQCAwMRyegUcNbp8ZpSyq0ZY5i07iAfr91P3fIlmO7fkAeLFXB1WS63e/duwsPDGTBgAN27dycmJob777/fIdvSFh5KKbeVkJTC6AWRfLx2P0/XL8u3wU1zfEjcuHGDcePG0aBBA15//XUSEhIAHBYSoEcUSik3deJiAsFzQtl19CIvd6zOsNaPZPuQiqfZunUrQ4YMISoqiv79+zNhwgQKFHB8cGpQKKXcTmTcBYJnh3I1MZnp/hba+zzk6pJc7ujRo7Rs2ZKHHnqI77//ni5dujht2xoUSim3sjzyKC8t3smDRfMzZ4gv1UoXdXVJLrV//34ee+wxypYty7fffoufnx/FihVzag16jkIp5RZSUw0frt7LmAWR1CtfghWjWuTokLhw4QLBwcFUr16dDRs2APD00087PSRAjyiUUm7gSmIyz38byY/RJ+nbuDxvd61Fvjw593vsihUrGD58OCdOnOCll16iUaNGLq1Hg0Ip5VJx564RNDuUA6euMO4pHwKaV8rRJ60DAwMJCQmhdu3aLF++HIvF4uqSNCiUUq6zNeYsw+eGk5ySyteDGtGyapaam3q8tE38LBYLFStW5OWXXyZfvnwursxKg0Ip5RILtsXy5vLdlC9ZiBkBFh55oIirS3KJuLg4hg0bRp8+ffD392fYsGGuLulvcu4goFLKJZJTUnl7ZRSvLN1Fsyql+G6kb44MidTUVCZPnkzNmjX59ddfSUxMdHVJt6VHFEopp7l4PYlR88L57cAZBvtW5rXO1cmTO+d9Xz1w4ACBgYFs2LCB9u3bM23aNCpXruzqsm5Lg0Ip5RQxp68QOCuUuPPXGN+jNr0bVXB1SS4THR3Nzp07+eqrrxg4cKDbn7zXoFBKOdyG/acZNS+cPLlzMTewKY0r3+fqkpxux44dREZGEhAQQLdu3YiJiaFkyZKuLssuOe+YTynlNMYYZm46zMCZ2yhToiDLR/rmuJBITEzkzTffxGKx8Oabb95q4ucpIQEaFEopB7mRnMpr3+3i7ZXR+NV4iCXDm1P+vkKuLsupNm/eTP369Xnvvffo168fERERTmnil9106Ekple3OXklk+Nxwth0+x8i2VXjh8WrkyuXe4/DZ7ejRo7Ru3ZrSpUuzatUqOnXq5OqSskyDQimVrfaeuETgrFBOX07kv33q0a1eWVeX5FR79uyhRo0alC1bloULF+Ln50fRop7ds0qHnpRS2ebH6JP0+PJ3biSnsnBosxwVEufPn2fw4MH4+Pjw22+/AdC9e3ePDwnQIwqlVDYwxvDlr4f4eO0+apctzjR/C6WLe95YfFZ99913jBgxgtOnT/Pqq6+6vIlfdtOgUErdk4SkFF5espPlkcd4qm4ZPupZhwJ5c7u6LKcZPHgwM2fOpF69evzwww80aNDA1SVlOw0KpVSWnbqUQNCcMHbEXeClJ6oxok0Vt795LDukbeLXtGlTqlatyosvvkjevHldXJljaFAopbJkZ/wFgmeHcSkhian+DXmiZmlXl+QUR44cYejQofTr148BAwYQHBzs6pIcTk9mK6Xu2sodx3h2ymZy5xKWDG+eI0IiNTWVSZMmUatWLTZu3EhSUpKrS3IaPaJQStktNdUw4af9fP7LQRpVKsnk/g0pVSS/q8tyuH379hEYGMjGjRvp0KEDU6dOpVKlSq4uy2k0KJRSdrmamMzYhZGsiTpJL0s53uteO8dMV7pv3z6ioqL4+uuvGTBgQI44D5OWBoVS6o7iz18jcFYo+09e5s0nfRjs6/3TlUZERBAZGcmgQYPo2rUrMTExlChRwtVluUTO+DqglMqy0D/O0e2LTRy9cJ2ZgxozpEVlrw6JhIQEXnvtNRo1asS4ceNuNfHLqSEBGhRKqUwsDI2j7/QtFCuYl2UjfWn9mHfPab1p0ybq1avHBx98wIABA4iMjPTIJn7ZTYeelFJ/k5Jq+GDVHmZsPEyLR0sxqV8DihfyznsEbjp69Cht27albNmyrFmzhg4dOri6JLehQaGU+otLCUn8Y14E6/efZmDzSrzRpYZXT1caHR2Nj48PZcuWZcmSJbRt25YiRXLeHN6Z8d6/faXUXTt85ipPT9rEpoNneP/p2ozrWtNrQ+LcuXMMHDiQmjVrsmHDBgCeeuopDYkM6BGFUgqATQfPMGJuOLkEvglsQtNH7nd1SQ6zZMkSRo4cydmzZ3n99ddp3Lixq0tyaxoUSuVwxhjmbDnC2yujqfJAYWYMaESF+713JrqBAwcya9YsGjRowOrVq6lXr56rS3J7GhRK5WBJKamMWxHF3K2xtK/xIBN616NoAe87aZ22iV/z5s2pUaMGL7zwAnny6EegPRw6+CgiHUVkn4gcFJFXMni9goisE5EIEdkpIp0dWY9S6k/nr97AP2Qrc7fGMrxNFab6W7wyJA4fPkyHDh2YPXs2AMHBwbz88ssaEnfBYUEhIrmBSUAnwAfoKyI+6RZ7A1hojKkP9AG+dFQ9Sqk/7T95mW6TNhEee4EJvevycsfq5PayOa1TUlKYOHEitWrVYsuWLbeOKtTdc2SkNgYOGmNiAERkAdANiE6zjAGK2X4uDhxzYD1KKeDnPScZsyCSgvly821wU+pXKOnqkrLdnj17GDJkCJs3b6ZTp05MmTKFChUquLosj+XIoCgLxKV5HA80SbfMOGCtiPwDKAy0z2hFIhIMBAP6l61UFhljmLYhhv+s3kvNMsWYPsDCw8ULuroshzh48CD79u1jzpw5PPfcc17dcsQZHHmOIqO/mfTHfn2Br40x5YDOwBwR+VtNxphpxhiLMcbywAPe3UJAKUdISErhhYU7+OB/e+lc+2EWDW3udSERFhbGV199BVjvhzh8+DD9+/fXkMgGjgyKeKB8msfl+PvQ0hBgIYAxZjNQACjlwJqUynFOXU6g7/QtLI04ytjHH+OLvvUpmM975rS+fv06r7zyCk2aNOHdd9+91cSvWLFid3inspcjg2I7UFVEKotIPqwnq1ekWyYW8AMQkRpYg+K0A2tSKkfZffQi3b7YxN7jl5n8XANG+1X1qm/YGzZsoG7duowfP56BAwcSERGhTfwcwGHnKIwxySIyClgD5Aa+MsZEicg7QKgxZgXwAjBdRJ7HOiw10OilCUplix92HueFRZHcVygfi4c3o2aZ4q4uKVsdPXoUPz8/ypcvz08//YSfn5+rS/Ja4mmfyxaLxYSGhrq6DKXcVmqq4b8/H+C/Px+gQYUSTPW38EBR75mudNeuXdSuXRuA77//nrZt21K4cGEXV+X+RCTMGGPJynu9s9uXUjnUtRvJjJofzn9/PkDPhuWYH9zUa0LizJkz+Pv7U6dOnVtN/J588kkNCSfQWxOV8hLHLlwnaHYoe45f4vXONQhs6R0z0RljWLRoEaNGjeL8+fO89dZbNGmS/kp75UgaFEp5gbAj5xk6J4zEpBRCAhrRtvqDri4p2wQEBDBnzhwsFgs///zzrWEn5TwaFEp5uMVh8by2dBcPlyjA/KAmVH2oqKtLumdpm/i1bt2aOnXq8M9//lP7M7mI7nWlPFRKquHD1XuZuiGG5lXuZ1K/BpQsnM/VZd2zmJgYgoKC6N+/P4MGDWLIkCGuLinH05PZSnmgywlJBM0OZeqGGPybVmTW4MYeHxIpKSl89tln1K5dm+3bt5Mrl348uQs9olDKwxw5e5XAWaHEnLnKu91r4d+0oqtLumfR0dEMHjyYrVu30qVLF6ZMmUK5cuVcXZay0aBQyoP8fsg6XSnAnCGNaV7FOzreHD58mEOHDjFv3jz69OnjFVdreRMNCqU8xDdbjjBuRRSVShUmJMBCxfs9+/6B7du3ExkZSVBQEF26dCEmJoaiRT3/RLw30kFApdxcUkoqby7bzRvLdtOyaimWjmju0SFx7do1XnzxRZo2bcoHH3xwq4mfhoT70qBQyo1duHaDgK+2MWfLEYJbPcKMgEYU8+DpSn/99Vfq1KnDJ598QlBQkDbx8xA69KSUmzp46jKBs0I5diGBj5+tS8+Gnn1yNz4+nscff5yKFSvyyy+/0LZtW1eXpOykQaGUG1q37xSj50WQP28u5gc3pWFFz52udMeOHdStW5dy5cqxfPly2rRpQ6FChVxdlroLOvSklBsxxjB9QwxDvt5O+fsKsXxUC48NidOnT9OvXz/q1avH+vXrAejcubOGhAfSIwql3ERicgqvf7ebxWHxdKpVmk961aVQPs/7FTXGsGDBAkaPHs3Fixd5++23adasmavLUvfA8/4VKuWFTl9OZNg3YYQdOc8Yv6qM8atKrlyeeS+Bv78/c+fOpUmTJoSEhFCzZk1Xl6TukQaFUi4WdewiQbNCOXftBpP6NaBLnYddXdJdS01NRUQQEdq2bUvDhg0ZPXo0uXN7z9zcOZmeo1DKhVbvPk7PyZsxwOJhzT0yJA4ePIifnx8zZ84EYMiQITz//PMaEl5Eg0IpFzDGMPHnAwz7JpzqDxdl+ShfapX1rDmtk5OT+fjjj6lduzYRERHky+fZTQnV7enQk1JOdv1GCi8t3sH3O4/zTP2yvP9MbQrk9axv37t372bQoEGEhobSrVs3vvzyS8qUKePqspSDaFAo5UTHL14neHYYu49d5JVO1Rna6hGPbIAXGxvLkSNHWLBgAb169fLIP4Oy3x2DQkRqGWN2O6MYpbxZROx5gueEcS0xmRkDLPjVeMjVJd2VrVu3smPHDoKDg+ncuTMxMTEUKVLE1WUpJ7DnHMUUEdkmIiNEpITDK1LKCy2LOErvaVsokDcX34309aiQuHr1KmPHjqVZs2Z8+OGHJCYmAmhI5CB3DApjTAvgOaA8ECoi80TkcYdXppQXSE01jF+9l39+G0n98iVYPrIFj3nQnNa//PILderUYcKECQwbNozw8HDy58/v6rKUk9l1jsIYc0BE3gBCgYlAfbEOSr5mjFnqyAKV8lRXEpP554IIftpzin5NKjDuqZrky+M5FxrGx8fzxBNPULlyZdavX0+rVq1cXZJyEXvOUdQBBgFdgB+Bp4wx4SJSBtgMaFAolU7cuWsEzgrl4OkrvNOtJv5NK3rMCd+IiAjq169PuXLlWLlyJa1bt6ZgwYKuLku5kD1fb74AwoG6xpiRxphwAGPMMeANRxanlCfaEnOWrl9s5PjF68wa1JgBzSp5REicPHmS3r1706BBg1tN/Dp27KghoewKiqXGmDnGmOs3nxCRMQDGmDkOq0wpDzR/Wyz9Z2ylZOF8LB/VghZV3X9Oa2MM33zzDT4+Pixbtoz33nuP5s2bu7os5UbsCYoBGTw3MJvrUMqjJaekMm5FFK8u3YXvo6X4boQvlUt5xnSl/fr1w9/fn2rVqhEZGcnrr79O3ryeO4ueyn63PUchIn2BfkBlEVmR5qWiwFlHF6aUp7h4LYlR88P57cAZAltU5tXONcjt5p1f0zbx69ChA82aNWPkyJHan0llKLOT2b8Dx4FSwCdpnr8M7HRkUUp5ikOnrxA4K5T489f4sEcdejUq7+qS7mj//v0EBQUxYMAAhgwZwqBBg1xdknJztw0KY8wR4AigM44olYH1+08zal44+XLnYl5QUxpVus/VJWUqOTmZTz/9lLfeeosCBQroSWplt8yGnjYaY1qIyGXApH0JMMaYYg6vTik3ZIxh5qY/eO+HaB57qCgzAiyUK+ne03vu3LmTwYMHExYWxtNPP82kSZN4+GHPa2muXCOzI4oWtv97zm2kSjnYjeRU3ly2m29D4+jg8xATetejcH73760ZHx9PXFwcixYtokePHh5xua5yH3e86klE/isiWRp+EpGOIrJPRA6KyCu3WaaXiESLSJSIzMvKdpRyhrNXEuk/Yyvfhsbxj3aPMqV/Q7cOid9//50pU6YA3Gri17NnTw0JddfsuTw2HHjT9mH/kYhY7FmxiOQGJgGdAB+gr4j4pFumKvAq4GuMqQn8866qV8pJ9hy/RNcvNrEj/gIT+9bnhQ7V3HZO6ytXrjBmzBhatGjBJ598cquJX+HCnnG5rnI/9jQFnGWM6Qw0BvYD40XkgB3rbgwcNMbEGGNuAAuAbumWCQImGWPO27Z16q6qV8oJ1kadoMfk30lOTWXRsGZ0reu+E/SsXbuWWrVq8fnnnzNy5Eht4qeyxd0cNz8KVAcqAdF2LF8WiEvzOB5okm6ZxwBEZBOQGxhnjFmdfkUiEgwEA1SoUOEuSlYq64wxfPnrIT5as4+65YozbYCFh4oVcHVZtxUXF0eXLl2oUqUKGzZsoEWLFq4uSXkJe5oCjgeeAQ4BC4F3jTEX7Fh3RsflJt3jPEBVoA1QDvjNNlHSX9ZvjJkGTAOwWCzp16FUtktISuHlJTtZHnmMbvXKML5HHbedrjQsLIyGDRtSvnx5Vq1aRcuWLSlQwH0DTXkee85RHAaaGWM6GmO+sjMkwHoEkfbuo3LAsQyWWW6MSTLGHAb2YQ0OpVzm5KUEek/dzPLIY7z0RDU+613PLUPixIkTPPvss1gslltN/B5//HENCZXtMruPoroxZi+wDaggIn8Z87nZRTYT24GqIlIZOAr0wdoSJK1lQF/gaxEphXUoKubu/ghKZZ8dcRcInhPK5YRkpvk3pEPN0q4u6W+MMcyePZvnn3+ea9eu8f7772sTP+VQmQ09jcV6XuCTDF4zQLvMVmyMSRaRUcAarOcfvjLGRInIO0CoMWaF7bUOIhINpAAvGWO0j5RyieWRR/nX4p08UDQ/S0c0p3pp97yntE+fPixcuBBfX19mzJhB9erVXV2S8nJiTOZD/iJSwBiTcKfnnMVisZjQ0FBXbFp5qdRUw6c/7ueLdQdpXOk+JvdvwP1F3OtKobRN/GbNmsXly5cZMWIEuXJ5zox5yrVEJMwYY9ftDenZ86/sdzufU8rjXE1MZtg3YXyx7iB9GpXnm8AmbhcSe/fupVWrVoSEhAAQEBDAqFGjNCSU02R2jqI01ktcC4pIff68iqkY4N6NbZSyQ/x563Sl+09e5q2nfBjY3L1moktKSuKjjz7i7bffpnDhwhQpUsTVJakcKrNzFE9gnaCoHPBpmucvA685sCalHG77H+cYNieMGympfD2oMa0ee8DVJf1FZGQkgwYNIjIykp49e/L5559TurT7nVhXOUNmTQFnAbNEpIcxZokTa1LKob7dHssby3ZTvmQhpgdYqPKA+31TP3HiBCdOnGDJkiU888wzri5H5XCZDT31N8Z8A1QSkbHpXzfGfJrB25RyW8kpqby/ai9fbTpMy6ql+KJvA4oXcp8pPzdu3MjOnTsZMWIEHTt25NChQxQqpKO8yvUyOxt2s4NYEazTn6b/TymPcfF6EoNnhfLVpsMM8q3EzIGN3CYkLl++zKhRo2jZsiWfffbZrSZ+GhLKXWQ29DTV9v+3nVeOUtnv8JmrDJm1ndiz1/jgmdr0bew+/cLWrFlDcHAwcXFxjBkzhvfee0+b+Cm3Y898FB+KSDERySsiP4vIGRHp74zilLpXvx04TbcvNnLhWhJzA5u4VUjExcXx5JNPUqhQITZu3Mhnn32mVzYpt2TPhdgdjDGXgCex9mZ6DHjJoVUpdY+MMcz6/Q8GztzOw8ULsnykL00eud/VZWGMYdu2bQCUL1+e//3vf0RERGgLDuXW7AmKmwO5nYH5xphzDqxHqXt2IzmV177bzVsromhb7UGWjGhO+ftcP95//PhxevToQZMmTW418Wvfvr028VNuz575KFaKyF7gOjBCRB4AXNK+Q6k7OXf1BsO/CWPr4XOMaFOFF91gJjpjDF9//TVjx44lISGB8ePH4+vr69KalLobdwwKY8wrtjkpLhljUkTkKn+fqU4pl9t34jKBs7dz8lIin/WuR/f6ZV1dEgC9evVi8eLFtGzZkhkzZvDYY4+5uiSl7oq9M9zVwHo/RdrlZzugHqWy5Kfok4xZEEHh/HlYOLQZ9cqXcGk9KSkpiAi5cuXiqaeeol27dgwdOlT7MymPZM8Md3OAKkAk1lbgYG0zrkGhXM4Yw5T1MXy4Zi+1yhRn+gALpYu7dsx/z549DBkyhEGDBhEUFMSAAQNcWo9S98qeIwoL4GPu1I9cKSdLSErh1aW7+C7iKE/WeZiPetalYD7XzUSXlJTE+PHjeffddylSpAjFixd3WS1KZSd7gmI3UBo47uBalLLbqUsJBM8JIzLuAi88/hij2j3q0s6vERERDBw4kJ07d9K7d28mTpzIgw8+6LJ6lMpO9gRFKSBaRLYBiTefNMZ0dVhVSmViV/xFgmaHcvF6ElP6N6RjLdd3VT158iRnzpxh2bJldOum13oo72JPUIxzdBFK2ev7ncd4cdEO7i+cn8XDm1GzjOuGdzZs2MCuXbsYOXIkHTt25ODBgxQsWNBl9SjlKHe8BMMYsx74A8hr+3k7EO7gupT6i5vTlY6aF0GtMsVZPsrXZSFx6dIlRowYQevWrZk4ceKtJn4aEspb2dPrKQhYDEy1PVUWWObIopRK69qNZEbOC2fizwd4tmE55gY1oZSLpitdtWoVNWvWZOrUqYwdO5bw8HBt4qe8nj1DTyOBxsBWAGPMARHRs3TKKY5euE7QrFD2nrjEG11qMKRFZZedtI6Li6Nbt25Uq1aNxYsX06RJE5fUoZSz2RMUicaYGzd/OW033emlssrhwo6cY+icMBKTUgkZ2Ii21Zz//cQYw9atW2natCnly5dn7dq1+Pr6ki9fPqfXopSr2HOb6HoReQ0oKCKPA4uAlY4tS+V0i0Lj6DttK0Xy5+G7kc1dEhLHjh2je/fuNGvW7FYTv7Zt22pIqBzHniOKV4AhwC5gKLAKmOHIolTOlZJq+M//9jD9t8P4Pno/k/o1oEQh534wG2MICQnhxRdfJDExkY8//lib+KkczZ6mgKnAdGC6iNwHlNO7tJUjXEpIYsz8CNbtO82AZhV580kf8uZ2fm+knj17snTpUlq3bs2MGTN49NFHnV6DUu7Enl5PvwJdbctGAqdFZL0xZqyDa1M5yB9nrhI4O5Q/zlzl30/X4rkmFZ26/bRN/Lp3706HDh0ICgrSJn5KYd85iuK2Ge6eAWYaYxoC7R1blspJfj94hu5fbuLMlUTmDGni9JDYvXs3vr6+hISEAODv76+dXpVKw57fhDwi8jDQC/jewfWoHGbO5j/w/2obDxTJz/KRvjSr4rzpSm/cuMHbb79NgwYNOHToECVLlnTatpXyJPaczH4HWANsNMZsF5FHgAOOLUt5u6SUVN5eGcU3W2Lxq/4gn/WpR9ECee/8xmwSFhbGwIED2b17N/369eOzzz7jgQcecNr2lfIk9pzMXoT1ktibj2OAHo4sSnm381dvMGJuOJtjzjK09SP864nq5HbydKVnz57lwoULrFy5kieffNKp21bK09g7w51S2eLAycsEzg7l+IUEPu1Vl2calHPattetW8euXbsYPXo0HTp04MCBAxQo4NpJjpTyBHq2TjnNur2nePrL37mamML84KZOC4mLFy8ydOhQ2rVrx+TJk2818dOQUMo+tw0KEWkmrpwJRnkNYwzTNhxi8KztVLy/ECtG+dKwonNOHK9cuRIfHx9mzJjBiy++SFhYmDbxU+ouZTb0FABMEpH9wGpgtTHmhHPKUt4iMTmF15buZkl4PJ1rl+bjZ+tSKJ9zRjzj4uLo0aMH1atXZ9myZTRq1Mgp21XK29z2N9YYMwxARKoDnYCvRaQ4sA5rcGwyxqQ4pUrlkU5dTmDYnDDCYy/wz/ZVGd2uKrkcfNLaGMPmzZtp3rz5rSZ+zZs31/5MSt0DeyYu2muMmWCM6Qi0AzYCz2JrO54ZEekoIvtE5KCIvJLJcj1FxIiI5W6KV+5r99GLdP9iE9HHL/Hlcw34Z/vHHB4S8fHxdO3aFV9f31tN/Nq0aaMhodQ9uqsxAGPMdaxNAVfdaVkRyQ1MAh4H4oHtIrLCGBOdbrmiwGjsCB7lGf636zhjF+6gZKG8LB7WnFplHTsTXWpqKtOnT+ell14iOTmZTz/9lBYtWjh0m0rlJI4cLG4MHLTdd4GILAC6AdHplnsX+BB40YG1KCcwxjDx54NM+Gk/9SuUYKp/Qx4s6vgri3r06MGyZcto164d0xJdNkcAABuwSURBVKdP55FHHnH4NpXKSRwZFGWBuDSP44G/TAkmIvWB8saY70XktkEhIsFAMECFChUcUKq6V9dvpPDi4h38sPM4zzQoy/tP16ZA3twO215ycjK5cuUiV65c9OjRgy5dujBkyBCXzX6nlDe76/soRCS3iDxnz6IZPHerPbmI5AImAC/caUXGmGnGGIsxxqJtFtzP8YvXeXbq76zadZzXOlfnk2frOjQkdu7cSbNmzZg+fToA/fv3JzAwUENCKQfJ7D6KYiLyqoh8ISIdxOofQAzWBoF3Eg+UT/O4HHAszeOiQC3gVxH5A2gKrNAT2p4lPPY8Xb/YxB9nrhESYCG4VRWHfWAnJiby1ltv0bBhQ44cOaK9mZRyksyGnuYA54HNQCDwEpAP6GaMibRj3duBqiJSGTgK9AH63XzRGHMRKHXzsW3eixeNMaF3+WdQLrI0PJ5Xlu6idLECzA1swmMPFXXYtrZv387AgQOJjo7G39+fCRMmcP/9zus0q1ROlllQPGKMqQ0gIjOAM0AFY8xle1ZsjEkWkVFYO8/mBr4yxkSJyDtAqDFmxT3WrlwkJdXw0Zp9TFl/iKaP3Mfk5xpSsrBjL0E9f/48V65cYdWqVXTq1Mmh21JK/ZXcblZTEQk3xjS43WNXsVgsJjRUDzpc5XJCEv9cEMnPe0/xXJMKjOta02HTlf7yyy/s2rWLMWPGANahJ22/oVTWiEiYMSZLQ/uZ/YbXFZFLInJZRC4DddI8vpS1UpUniz17jR6Tf+fX/ad5t1tN/v10bYeExIULFwgKCsLPz4+pU6feauKnIaGUa2TWwsNxl60oj7P50FlGzA0j1cDswY3xfbTUnd+UBcuXL2f48OGcPHmSf/3rX4wbN04DQikXu21QiEgBYBjwKLAT6zmGZGcVptzHvK2x/N/y3VS8vxAhAY2oVKqwQ7YTGxvLs88+S40aNVixYgUWi14Ap5Q7yOxk9iwgCfgN6AzUBMY4oyjlHpJTUnn3+2hmbT5Cm2oPMLFvfYpl83Slxhg2btxIy5YtqVChAj/99BNNmzbV/kxKuZHMBph9jDH9jTFTgZ5ASyfVpNzAhWs3CJi5jVmbjxDUsjIhAY2yPSRiY2Pp0qULrVq1utXEr1WrVhoSSrmZzI4okm7+YLvU1QnlKHdw8NQVAmdt59iFBD7qWYdnLeXv/Ka7kJqaypQpU3j55Zet/aEmTtQmfkq5scyCol6aq5sEKGh7LIAxxhRzeHXK6X7dd4p/zI8gf55czAtqgqXSfdm+jWeeeYbly5fz+OOPM23aNCpVqpTt21BKZZ/MgmKHMaa+0ypRLmWMIWTjYd5ftYdqpYsxfUBDypUslG3rT9vEr3fv3nTr1o2BAwdqfyalPEBm5ygyvhNPeZ3E5BReXrKT937YQwef0iwe1ixbQ2LHjh00adKEadOmAdC3b18GDRqkIaGUh8jsiOJBERl7uxeNMZ86oB7lZGeuJDL8mzC2/3Ge0e0ezdaZ6BISEnjvvfcYP3489913H6VLl86W9SqlnCuzoMgNFCHjduHKC0Qfu0TQ7FDOXEnk8771eapumWxb97Zt2wgICGDv3r0EBATw6aefct992X++QynleJkFxXFjzDtOq0Q51ZqoEzz/bSTFClinK61dLnunK7106RLXr19n9erVPPHEE9m6bqWUc2UWFHok4YWMMUxad5CP1+6nbvkSTPdvyIPFsme60rVr1xIVFcXzzz9P+/bt2bdvn7bfUMoLZHYy289pVSinSEhKYfSCSD5eu5/u9crwbXDTbAmJ8+fPM2jQIJ544glCQkK0iZ9SXua2QWGMOefMQpRjnbiYQK+pm/l+5zH+1bEaE3rXy5bpSpcuXYqPjw9z5szh1VdfJTQ0VANCKS+T2dCT8hKRcRcInh3K1cRkpvtbaO/zULasNzY2lj59+lCrVi1WrVpF/fp6241S3sgxM84ot7E88ii9pm4mX55cLB3he88hYYy51ZepQoUK/PLLL2zdulVDQikvpkHhpVJTDR+u3suYBZHUK1+C5SN9qVb63ua0PnLkCJ06daJNmza3wqJFixbkzZu9zQKVUu5Fh5680JXEZJ7/NpIfo0/St3F53u5ai3x5sv6dIDU1lS+//JJXXnkFgM8//5yWLbWZsFI5hQaFl4k7d42g2aEcOHWFcU/5ENC80j23yujevTsrV67kiSeeYOrUqVSsWDGbqlVKeQINCi+yNeYsw+eGk5ySyteDGtGy6gNZXldSUhK5c+cmV65c9O3bl549e+Lv76/9mZTKgfQchZdYsC2W/iFbKVEwL8tG+t5TSISHh9O4cWOmTJkCWJv4DRgwQENCqRxKg8LDJaek8vbKKF5ZuotmVUrx3UhfHnmgSJbWdf36dV599VUaN27MiRMnKF8+eycsUkp5Jh168mAXrycxal44vx04w2DfyrzWuTp5cmct+7ds2UJAQAD79+9n8ODBfPzxx5QsWTKbK1ZKeSINCg8Vc/oKgbNCiTt/jfE9atO7UYV7Wt/Vq1dJSkrixx9/pH379tlUpVLKG2hQeKDfDpxm5Nxw8uTOxdzApjSunLX23atXryYqKooXXngBPz8/9u7dS758+bK5WqWUp9NzFB7EGMPMTYcZOHM7ZUoUZPlI3yyFxNmzZwkICKBTp07MmjWLGzduAGhIKKUypEHhIW4kp/Lad7t4e2U07ao/yOLhzSl/391NV2qMYfHixfj4+DBv3jzeeOMNtm/frgGhlMqUDj15gHNXbzDsmzC2HT7HyLZVeOHxalmarjQ2NpZ+/fpRp04d1q5dS926dR1QrVLK22hQuLm9Jy4ROCuU05cT+W+fenSrV/au3m+MYd26dbRr146KFSvy66+/0rhxY/Lk0b96pZR9dOjJjf0YfZIeX/7OjeRUFg5tdtchcfjwYTp06ICfn9+tJn7NmzfXkFBK3RUNCjd0c7rS4DmhVHmwCCtGtaBu+RJ2vz8lJYX//ve/1KpVi61btzJ58mRt4qeUyjL9aulmEpJSeGXJTpZFHuOpumX4qGedu56Jrlu3bvzwww907tyZKVOm6B3WSql7okHhRk5dSiBoThg74i7w0hPVGNGmit39ldI28fP396dv377069dP+zMppe6ZQ4eeRKSjiOwTkYMi8koGr48VkWgR2SkiP4tIju1fvTP+Al2/2MSBk5eZ0r8hI9s+aveHfGhoKBaLhcmTJwPQu3dvnnvuOQ0JpVS2cFhQiEhuYBLQCfAB+oqIT7rFIgCLMaYOsBj40FH1uLOVO47x7JTN5M4lLBnenI61Stv1vuvXr/Pyyy/TpEkTTp8+rfNEKKUcwpFDT42Bg8aYGAARWQB0A6JvLmCMWZdm+S1AfwfW43ZSUw2f/bSfib8cpFGlkkzu35BSRfLb9d7NmzcTEBDAgQMHCAwM5KOPPqJECftPeCullL0cGRRlgbg0j+OBJpksPwT4X0YviEgwEAxQocK9Nb9zF1cTkxm7MJI1USfpZSnHe91r39V0pdevXyc1NZWffvoJPz8/B1aqlMrpHBkUGQ2QmwwXFOkPWIDWGb1ujJkGTAOwWCwZrsOTxJ+/RtDsMPaduMSbT/ow2Ne+6UpXrVpFVFQUL730Eu3atWPPnj3kzZvXCRUrpXIyR57MjgfSXpdZDjiWfiERaQ+8DnQ1xiQ6sB63EPrHObpP2kT8+WvMHNSYIS0q3zEkzpw5Q//+/enSpQtz58691cRPQ0Ip5QyODIrtQFURqSwi+YA+wIq0C4hIfWAq1pA45cBa3MLC0Dj6Tt9Ckfx5+G6EL60fy3y6UmMMCxYsoEaNGixcuJC33nqLbdu2aRM/pZRTOWzoyRiTLCKjgDVAbuArY0yUiLwDhBpjVgAfAUWARbZv1bHGmK6OqslVUlINH6zaw4yNh2nxaCm+6FefEoXu/GEfGxtLQEAAdevWJSQkhNq1azuhWqWU+isxxrOG/C0WiwkNDXV1GXa7lJDEP+ZFsH7/aQY2r8QbXWpkOl2pMYaff/751ixzW7ZsoVGjRuTOfXd3ZyulVFoiEmaMsWTlvdrryYEOn7nK05M2sengGd5/ujbjutbMNCQOHTqEn58fjz/++K0mfk2bNtWQUEq5lLbwcJBNB88wYm44uQTmDGlCsyr333bZm0383njjDfLmzcvUqVO1iZ9Sym1oUDjA7M1/8PbKaKo8UJgZAxpR4f7MZ6J76qmn+N///seTTz7J5MmTKVeunHMKVUopO2hQZKOklFTGrYhi7tZY2td4kAm961G0QMaXsN64cYM8efKQK1cuBg4ciL+/P3369NH+TEopt6PnKLLJ+as38A/ZytytsQxvU4Wp/pbbhsS2bdto2LAhX375JQC9evWib9++GhJKKbekQZEN9p+8TLdJmwiPvcCE3nV5uWN1cmcwp/W1a9d44YUXaNasGefPn6dKlSouqFYppe6ODj3do5/3nGTMgkgK5svNt8FNqV+hZIbLbdy4kYCAAGJiYhg6dCjjx4+nePHiTq5WKaXungZFFhljmLYhhv+s3kvNMsWYPsDCw8UL3nb5mxMLrVu3jjZt2jivUKWUukcaFFmQkJTCa0t3sTTiKF3qPMzHPetSMN/f73VYuXIle/bs4V//+hdt27YlOjqaPHl0lyulPIueo7hLpy4n0Hf6FpZGHGXs44/xRd/6fwuJ06dP069fP7p27cr8+fNvNfHTkFBKeSINiruw++hFun2xib3HLzP5uQaM9qv6lyuVjDHMmzePGjVqsHjxYt555x22bt2qTfyUUh5Nv+La6Yedx3lhUST3FcrH4uHNqFnm7yeiY2NjGTRoEPXr1yckJISaNWu6oFKllMpeekRxBzenKx05Lxyfh4uxfFSLv4REamoqa9asAaBixYr89ttvbNq0SUNCKeU1NCgyce1GMqPmh/PZTwfo2bAc84Ob8kDRP+e0PnDgAO3ataNjx45s2LABgMaNG2sTP6WUV9Ghp9s4duE6QbNDiT5+idc71yCw5Z8z0SUnJzNhwgT+7//+j/z58xMSEqJN/JRSXkuDIgNhR84zdE4YiUkpfBXQiLbVH/zL608++SRr1qyhW7dufPnll5QpU8ZFlSqllONpUKSzJCyeV5fu4uESBZgf1ISqDxUFIDExkbx585IrVy4CAwMZPHgwzz77rPZnUkp5PT1HYXNzutIXFu3AUqkky0b43gqJLVu20KBBAyZNmgRAz5496dWrl4aEUipH0KAALickETQ7lKkbYvBvWpFZgxtTsnA+rl69yvPPP0/z5s25fPkyVatWdXWpSinldDl+6OnI2asEzgol5sxV3u1eC/+mFQH47bffCAgI4PDhw4wYMYIPPviAYsWKubhapZRyvhwdFL8fsk5XCjBnSGOaVyl167Xk5GTy5s3L+vXradWqlatKVEopl8uxQfHNliOMWxFFpVKFCQmwUPH+wixbtow9e/bw6quv0rZtW6KiorQ/k1Iqx8tx5yiSUlJ5c9lu3li2m5ZVS7F0RHMKJF+hV69ePP300yxevFib+CmlVBo5KiguXLtBwFfbmLPlCMGtHmH6AAvLFy3Ax8eH5cuX8+9//5stW7ZoEz+llEojx3xlPnjqMoGzQjl2IYGPn61Lz4blOHLkCIGBgVgsFkJCQqhevbqry1RKKbeTI4Ji3b5TjJ4XQf68uZgb2Igze7cD5ahYsSKbNm2ifv362p9JKaVuw6uHnowxzPgthiFfb6f8fYWY0PEhXgh4hs6dO7N+/XoALBaLhoRSSmXCa48oEpNTeP273SwOi+eJGqUod/QXHm/xNgULFmTmzJl6yatSStnJK4Pi9OVEhn0TRtiR84z2q8rqj/7BtB/X8swzzzBp0iRKly7t6hKVUspjeF1QRB+7RNDsUE5fvMzE3g3pWr8cFS8FM3RoMD169HB1eUop5XG8KihW7z7B899Gkvv0PpJ+msQfD4+C+qM1IJRS6h54RVAYY/j8l4N8/MMO8oR/yx8bllChQgVq1Kjh6tKUUsrjeXxQXL+RwkuLd7D4+7Vc/3Eil8+eYNSoUbz//vsUKVLE1eUppZTH8+igOHExgaDZoew+dpF+TSqwOrIE05ctwtfX19WlKaWU1/DYoIiIPc+zr3zGpRNHmD9pPH41HuKjUb31ngillMpmDr3hTkQ6isg+ETkoIq9k8Hp+EfnW9vpWEalkz3pn/hhBqyee4tD8d3jo3E5aVikJoCGhlFIO4LCgEJHcwCSgE+AD9BURn3SLDQHOG2MeBSYA4++03pj4EwQ+1ZJrB7fz5tvvEb59qzbxU0opB3Lk0FNj4KAxJgZARBYA3YDoNMt0A8bZfl4MfCEiYowxt1vp+ZPHKFejHqsWzaV2Tb2qSSmlHM2RQVEWiEvzOB5ocrtljDHJInIRuB84k3YhEQkGgm0PE+P3ROyuUyv9wUmOVIp0+yoH033xJ90Xf9J98adqWX2jI4NCMngu/ZGCPctgjJkGTAMQkVBjjOXey/N8ui/+pPviT7ov/qT74k8iEprV9zryZHY8UD7N43LAsdstIyJ5gOLAOQfWpJRS6i45Mii2A1VFpLKI5AP6ACvSLbMCCLD93BP4JbPzE0oppZzPYUNPtnMOo4A1QG7gK2NMlIi8A4QaY1YAIcAcETmI9Uiijx2rnuaomj2Q7os/6b74k+6LP+m++FOW94XoF3illFKZ8eoZ7pRSSt07DQqllFKZctugcFT7D09kx74YKyLRIrJTRH4WkYquqNMZ7rQv0izXU0SMiHjtpZH27AsR6WX7txElIvOcXaOz2PE7UkFE1olIhO33pLMr6nQ0EflKRE6JyO7bvC4iMtG2n3aKSAO7VmyMcbv/sJ78PgQ8AuQDdgA+6ZYZAUyx/dwH+NbVdbtwX7QFCtl+Hp6T94VtuaLABmALYHF13S78d1EViABK2h4/6Oq6XbgvpgHDbT/7AH+4um4H7YtWQANg921e7wz8D+s9bE2Brfas112PKG61/zDG3AButv9Iqxswy/bzYsBPRDK6gc/T3XFfGGPWGWOu2R5uwXrPijey598FwLvAh0CCM4tzMnv2RRAwyRhzHsAYc8rJNTqLPfvCAMVsPxfn7/d0eQVjzAYyvxetGzDbWG0BSojIw3dar7sGRUbtP8rebhljTDJws/2Ht7FnX6Q1BOs3Bm90x30hIvWB8saY751ZmAvY8+/iMeAxEdkkIltEpKPTqnMue/bFOKC/iMQDq4B/OKc0t3O3nyeA+85HkW3tP7yA3X9OEekPWIDWDq3IdTLdFyKSC2sX4oHOKsiF7Pl3kQfr8FMbrEeZv4lILWPMBQfX5mz27Iu+wNfGmE9EpBnW+7dqGWNSHV+eW8nS56a7HlFo+48/2bMvEJH2wOtAV2NMopNqc7Y77YuiQC3gVxH5A+sY7AovPaFt7+/IcmNMkjHmMLAPa3B4G3v2xRBgIYAxZjNQAGvDwJzGrs+T9Nw1KLT9x5/uuC9swy1TsYaEt45Dwx32hTHmojGmlDGmkjGmEtbzNV2NMVluhubG7PkdWYb1QgdEpBTWoagYp1bpHPbsi1jAD0BEamANitNOrdI9rAAG2K5+agpcNMYcv9Ob3HLoyTiu/YfHsXNffAQUARbZzufHGmO6uqxoB7FzX+QIdu6LNUAHEYkGUoCXjDFnXVe1Y9i5L14ApovI81iHWgZ64xdLEZmPdaixlO18zFtAXgBjzBSs52c6AweBa8Agu9brhftKKaVUNnLXoSellFJuQoNCKaVUpjQolFJKZUqDQimlVKY0KJRSSmVKg0JliYikiEhkmv8qiUgbEblo69C5R0Tesi2b9vm9IvLxHdYdZrsePu1zo2wdL43tnoC7qfX1NHWmrXu0iIwTkWsi8mCa5a9k8OfcLSIrRaREJtuZKiK+6Z5rJSLhIpIsIj3vpm5HE5EyIrI4zeP5to6iz4vIO7abOG/3XouITLT93EZEmjujZuUaenmsyhIRuWKMKZLuuTbAi8aYJ0WkMBCJ9f6WommeL4i1o+kQY8ymDNZbCZiY/j4Q202F54FfsXaEPZMddYvIOGAwMN8Y83L6ZdL9PAvYb4z5923WHQk0NMakpPvzFANeBFYYYxZn9F5XE5HSWDuJ3nWLets+vGKMyfQLgPJcekShHMIYcxUIA6qke/461gC5XSOyTsDqDNYXYYz5I5vLvOkroLeI3HeH5TZzm7ptd/vuTxsSAMaYP4wxO4Es9RQSkZoiss12VLNTRKrajt72isgs23OLRaSQbfmGIrLedlS2RmydQUXkURH5SUR22I5wqtjWc3PegrXAg7bttBSRr28eAYlIIxH53fbebSJS1HYU8b0tCIcBz6d572ERyWt7bzER+ePmY+WZNChUVhVMM4TzXfoXReR+rL2WotI9XxJrv6ENt1lvRzIICge7gjUsxtxuARHJjbUFxO3u/s4w4LLBMOC/xph6WBs+xtuerwZMM8bUAS4BI2wfxp8DPY0xDbH+mW4e/czF2nK8LtAcSN+2oStwyBhTzxjz280nbUOA3wJjbO9tD1y/+botvKcAE9K891egi22RPsASY0zSPe8J5TIaFCqrrts+GOoZY55O83xLEYnA+g31P8aYqDTP7wROAN8bY06kX6HtQ6mcMcYV/YgmAgEiUizd8wVtQ0pngfuAH2/z/idwTFBsBl4TkZeBirYjMoC4NEN33wAtsIZHLeBHW81vAOVEpChQ1hjzHYAxJiHN/CV3Ug04bozZbnvvJVtb/8zM4M/WEIOAmXZuS7kpDQqV3X4zxtQ3xjS09ZZJ+3wdoDYwXETqZfDelsDGrG5YRGbajnBW3e17ba2352GdOTGt67Zv8xWxzp42MoPtFgJKGGOyNBmOiIxMc3RWJl1d87B+278OrBGRdjdfSv9HwNpCOipNgNc2xnQg49bSdpeXwbYyZQuwSiLSGshtjMlwWk7lOTQolFMZY/YDHwAvZ/ByR+5h0iVjzCDbB2RW50P+FBhKBs0yjTEXgdHAixmMt7cF1mVxmxhjJqX5cP9L2IjII0CMMWYi1mGvOraXKoh1XgWwzrWwEWsb8QduPi8ieUWkpjHmEhAvIt1tz+e/eU7DDnuBMiLSyPbeomJt65/WZawXLKQ1G5iPHk14BQ0K5QpTgFYiUjnd822A9Rm9wXYpazzW/vk7RWRGdhdlu5LqOyD/bV6PwDofc/pOxbc9P2E7ERwPPAtMFZGojJbLRG9gt20oqTrWD2CAPViHynZiHRKbbJsGtCcwXkR2YL1o4OZlq/7AaNvyvwOl7dm4bZ29gc9t6/wRa4vutFYCT988mW17bi5QEmtYKA+nl8cqtyAi5YDpxphOrq7lbolIONDEWSdsbVcafW+MqeWM7WWF7YqpbsYYf1fXou6dW85HoXIeY0w81m/mHscY08DVNbgTEfkc699lVocAlZvRIwqllFKZ0nMUSimlMqVBoZRSKlMaFEoppTKlQaGUUipTGhRKKaUy9f84Tcyd7cHqdgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "show_roc_curve(fpr, tpr)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.6071428571428572"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics import auc\n",
    "\n",
    "roc_auc = auc(fpr, tpr)\n",
    "roc_auc"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Try a `Random Forest Classifier`:**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "RandomForestClassifier(bootstrap=True, ccp_alpha=0.0, class_weight=None,\n",
       "                       criterion='gini', max_depth=None, max_features='auto',\n",
       "                       max_leaf_nodes=None, max_samples=None,\n",
       "                       min_impurity_decrease=0.0, min_impurity_split=None,\n",
       "                       min_samples_leaf=1, min_samples_split=2,\n",
       "                       min_weight_fraction_leaf=0.0, n_estimators=100,\n",
       "                       n_jobs=None, oob_score=False, random_state=42, verbose=0,\n",
       "                       warm_start=False)"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.ensemble import RandomForestClassifier\n",
    "\n",
    "forest_clf = RandomForestClassifier(random_state=42)\n",
    "forest_clf.fit(X_train_prepared, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_valid_predict = forest_clf.predict(X_valid)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[45, 11],\n",
       "       [35,  9]], dtype=int64)"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "confusion_matrix(y_valid, y_valid_predict)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Precision score:\t 0.65\n",
      "Recall score:\t\t 0.29545454545454547\n",
      "F1 score:\t\t 0.40625\n"
     ]
    }
   ],
   "source": [
    "performance_scores(y_train, y_valid_predict)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The performance of this model is worst!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [],
   "source": [
    "fpr, tpr, thresholds = roc_curve(y_valid, y_valid_predict)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.5040584415584416"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "roc_auc = auc(fpr, tpr)\n",
    "roc_auc"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Try a `Linear SVC Classifier`:**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.svm import LinearSVC\n",
    "\n",
    "lin_svm_clf = Pipeline((\n",
    "    ('std_scaler', StandardScaler()),\n",
    "    (\"linear_svc\", LinearSVC(C=1, loss=\"hinge\", random_state=42))\n",
    "    ))\n",
    "\n",
    "lin_svm_clf.fit(X_train_prepared, y_train)\n",
    "y_valid_predict = lin_svm_clf.predict(X_valid)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[24, 32],\n",
       "       [ 1, 43]], dtype=int64)"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "confusion_matrix(y_valid, y_valid_predict)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Precision score:\t 0.4\n",
      "Recall score:\t\t 0.6818181818181818\n",
      "F1 score:\t\t 0.5042016806722689\n"
     ]
    }
   ],
   "source": [
    "performance_scores(y_train, y_valid_predict)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [],
   "source": [
    "precisions, recalls, thresholds = precision_recall_curve(y_valid, y_valid_predict)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [],
   "source": [
    "fpr, tpr, thresholds = roc_curve(y_valid, y_valid_predict)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "show_roc_curve(fpr, tpr)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.702922077922078"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "roc_auc = auc(fpr, tpr)\n",
    "roc_auc"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The performance of this model is a bit worse than `Stochastic Gradient Descent (SGD) Classifier`."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Try a `Nonlinear SVC Classifier`:**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Pipeline(memory=None,\n",
       "         steps=[('std_scaler',\n",
       "                 StandardScaler(copy=True, with_mean=True, with_std=True)),\n",
       "                ('svm_clf',\n",
       "                 SVC(C=1, break_ties=False, cache_size=200, class_weight=None,\n",
       "                     coef0=1, decision_function_shape='ovr', degree=2,\n",
       "                     gamma='scale', kernel='poly', max_iter=-1,\n",
       "                     probability=False, random_state=42, shrinking=True,\n",
       "                     tol=0.001, verbose=False))],\n",
       "         verbose=False)"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.svm import SVC\n",
    "\n",
    "poly_svm_clf = Pipeline((\n",
    "    ('std_scaler', StandardScaler()),\n",
    "    (\"svm_clf\", SVC(kernel=\"poly\", C=1, coef0=1, degree=2, random_state=42))\n",
    "    ))\n",
    "\n",
    "poly_svm_clf.fit(X_train_prepared, y_train)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_valid_predict = poly_svm_clf.predict(X_valid)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[41, 15],\n",
       "       [18, 26]], dtype=int64)"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "confusion_matrix(y_valid, y_valid_predict)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Precision score:\t 0.6341463414634146\n",
      "Recall score:\t\t 0.5909090909090909\n",
      "F1 score:\t\t 0.611764705882353\n"
     ]
    }
   ],
   "source": [
    "performance_scores(y_valid, y_valid_predict)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [],
   "source": [
    "fpr, tpr, thresholds = roc_curve(y_valid, y_valid_predict)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.661525974025974"
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "roc_auc = auc(fpr, tpr)\n",
    "roc_auc"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Optimal hyperparameteres** for `SVC` model through a `Random Search`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 30 candidates, totalling 150 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=4)]: Using backend LokyBackend with 4 concurrent workers.\n",
      "[Parallel(n_jobs=4)]: Done  33 tasks      | elapsed:    3.1s\n",
      "[Parallel(n_jobs=4)]: Done 150 out of 150 | elapsed:    9.5s finished\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "RandomizedSearchCV(cv=5, error_score=nan,\n",
       "                   estimator=SVC(C=1.0, break_ties=False, cache_size=200,\n",
       "                                 class_weight=None, coef0=0.0,\n",
       "                                 decision_function_shape='ovr', degree=3,\n",
       "                                 gamma='scale', kernel='rbf', max_iter=-1,\n",
       "                                 probability=False, random_state=None,\n",
       "                                 shrinking=True, tol=0.001, verbose=False),\n",
       "                   iid='deprecated', n_iter=30, n_jobs=4,\n",
       "                   param_distributions={'C': <scipy.stats._distn_infrastructure.rv_frozen object at 0x000002926331BE08>,\n",
       "                                        'gamma': <scipy.stats._distn_infrastructure.rv_frozen object at 0x000002926331B5C8>,\n",
       "                                        'kernel': ['linear', 'poly', 'rbf',\n",
       "                                                   'sigmoid']},\n",
       "                   pre_dispatch='2*n_jobs', random_state=42, refit=True,\n",
       "                   return_train_score=False, scoring='f1', verbose=2)"
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.model_selection import RandomizedSearchCV\n",
    "from scipy.stats import expon, reciprocal\n",
    "\n",
    "# Note: gamma is ignored when kernel is \"linear\"\n",
    "param_distribs = {\n",
    "        'kernel': ['linear', 'poly', 'rbf', 'sigmoid'],\n",
    "        'C': reciprocal(10, 10000),\n",
    "        'gamma': expon(scale=1.0),\n",
    "    }\n",
    "\n",
    "svm_clf = SVC()\n",
    "rnd_search = RandomizedSearchCV(svm_clf, param_distributions=param_distribs,\n",
    "                                n_iter=30, cv=5, scoring='f1',\n",
    "                                verbose=2, n_jobs=4, random_state=42)\n",
    "rnd_search.fit(X_train_prepared, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'C': 8906.20438616168, 'gamma': 0.628789100540856, 'kernel': 'linear'}"
      ]
     },
     "execution_count": 57,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rnd_search.best_params_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_valid_predict = rnd_search.predict(X_valid)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[36, 20],\n",
       "       [ 4, 40]], dtype=int64)"
      ]
     },
     "execution_count": 59,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "confusion_matrix(y_valid, y_valid_predict)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Precision score:\t 0.6666666666666666\n",
      "Recall score:\t\t 0.9090909090909091\n",
      "F1 score:\t\t 0.7692307692307692\n"
     ]
    }
   ],
   "source": [
    "performance_scores(y_valid, y_valid_predict)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [],
   "source": [
    "fpr, tpr, thresholds = roc_curve(y_valid, y_valid_predict)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "show_roc_curve(fpr, tpr)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.775974025974026"
      ]
     },
     "execution_count": 63,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "roc_auc = auc(fpr, tpr)\n",
    "roc_auc"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As 3000 out of 10,000 attributes have no predictive power, use only maximum **7000 top important features**.\n",
    "Use a regressor to select these important attributes."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "RandomizedSearchCV(cv=5, error_score=nan,\n",
       "                   estimator=RandomForestRegressor(bootstrap=True,\n",
       "                                                   ccp_alpha=0.0,\n",
       "                                                   criterion='mse',\n",
       "                                                   max_depth=None,\n",
       "                                                   max_features='auto',\n",
       "                                                   max_leaf_nodes=None,\n",
       "                                                   max_samples=None,\n",
       "                                                   min_impurity_decrease=0.0,\n",
       "                                                   min_impurity_split=None,\n",
       "                                                   min_samples_leaf=1,\n",
       "                                                   min_samples_split=2,\n",
       "                                                   min_weight_fraction_leaf=0.0,\n",
       "                                                   n_estimators=100,\n",
       "                                                   n_jobs=None, oob_score=Fals...\n",
       "                   iid='deprecated', n_iter=10, n_jobs=None,\n",
       "                   param_distributions={'max_features': <scipy.stats._distn_infrastructure.rv_frozen object at 0x000002926331B088>,\n",
       "                                        'n_estimators': <scipy.stats._distn_infrastructure.rv_frozen object at 0x000002926328C348>},\n",
       "                   pre_dispatch='2*n_jobs', random_state=42, refit=True,\n",
       "                   return_train_score=False, scoring='neg_mean_squared_error',\n",
       "                   verbose=0)"
      ]
     },
     "execution_count": 64,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.ensemble import RandomForestRegressor\n",
    "from scipy.stats import randint\n",
    "\n",
    "param_distribs = {\n",
    "        'n_estimators': randint(low=1, high=200),\n",
    "        'max_features': randint(low=1, high=7000),\n",
    "    }\n",
    "\n",
    "forest_reg = RandomForestRegressor(random_state=42)\n",
    "rnd_search = RandomizedSearchCV(forest_reg, param_distributions=param_distribs,\n",
    "                                n_iter=10, cv=5, scoring='neg_mean_squared_error', random_state=42)\n",
    "rnd_search.fit(X_train_prepared, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [],
   "source": [
    "feature_importances = rnd_search.best_estimator_.feature_importances_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(0.009764291971167202, 6480) (0.008043831213868355, 4191) (0.007451403735105302, 1882) (0.007070771465149308, 5472) (0.006928581609909464, 7563)\n"
     ]
    }
   ],
   "source": [
    "features_sorted = sorted(zip(feature_importances, list(X_train)), reverse=True)\n",
    "print(features_sorted[0], features_sorted[1], features_sorted[2], features_sorted[3], features_sorted[4])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [],
   "source": [
    "features_non_zero = feature_importances.nonzero()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "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>6</th>\n",
       "      <th>14</th>\n",
       "      <th>26</th>\n",
       "      <th>38</th>\n",
       "      <th>68</th>\n",
       "      <th>73</th>\n",
       "      <th>97</th>\n",
       "      <th>100</th>\n",
       "      <th>130</th>\n",
       "      <th>133</th>\n",
       "      <th>...</th>\n",
       "      <th>9953</th>\n",
       "      <th>9961</th>\n",
       "      <th>9968</th>\n",
       "      <th>9969</th>\n",
       "      <th>9971</th>\n",
       "      <th>9979</th>\n",
       "      <th>9987</th>\n",
       "      <th>9994</th>\n",
       "      <th>9995</th>\n",
       "      <th>9999</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>83</th>\n",
       "      <td>467</td>\n",
       "      <td>287</td>\n",
       "      <td>0</td>\n",
       "      <td>33</td>\n",
       "      <td>396</td>\n",
       "      <td>141</td>\n",
       "      <td>141</td>\n",
       "      <td>119</td>\n",
       "      <td>103</td>\n",
       "      <td>5</td>\n",
       "      <td>...</td>\n",
       "      <td>311</td>\n",
       "      <td>228</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>57</td>\n",
       "      <td>240</td>\n",
       "      <td>0</td>\n",
       "      <td>458</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>53</th>\n",
       "      <td>100</td>\n",
       "      <td>297</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>556</td>\n",
       "      <td>173</td>\n",
       "      <td>0</td>\n",
       "      <td>68</td>\n",
       "      <td>124</td>\n",
       "      <td>19</td>\n",
       "      <td>...</td>\n",
       "      <td>806</td>\n",
       "      <td>123</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>221</td>\n",
       "      <td>364</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>70</th>\n",
       "      <td>426</td>\n",
       "      <td>340</td>\n",
       "      <td>0</td>\n",
       "      <td>24</td>\n",
       "      <td>415</td>\n",
       "      <td>219</td>\n",
       "      <td>306</td>\n",
       "      <td>155</td>\n",
       "      <td>131</td>\n",
       "      <td>4</td>\n",
       "      <td>...</td>\n",
       "      <td>321</td>\n",
       "      <td>236</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>83</td>\n",
       "      <td>129</td>\n",
       "      <td>0</td>\n",
       "      <td>409</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>45</th>\n",
       "      <td>65</td>\n",
       "      <td>297</td>\n",
       "      <td>0</td>\n",
       "      <td>13</td>\n",
       "      <td>669</td>\n",
       "      <td>120</td>\n",
       "      <td>0</td>\n",
       "      <td>43</td>\n",
       "      <td>117</td>\n",
       "      <td>11</td>\n",
       "      <td>...</td>\n",
       "      <td>816</td>\n",
       "      <td>133</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>7</td>\n",
       "      <td>2</td>\n",
       "      <td>31</td>\n",
       "      <td>0</td>\n",
       "      <td>195</td>\n",
       "      <td>292</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>44</th>\n",
       "      <td>382</td>\n",
       "      <td>217</td>\n",
       "      <td>0</td>\n",
       "      <td>56</td>\n",
       "      <td>456</td>\n",
       "      <td>137</td>\n",
       "      <td>46</td>\n",
       "      <td>39</td>\n",
       "      <td>97</td>\n",
       "      <td>9</td>\n",
       "      <td>...</td>\n",
       "      <td>404</td>\n",
       "      <td>291</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>8</td>\n",
       "      <td>0</td>\n",
       "      <td>37</td>\n",
       "      <td>153</td>\n",
       "      <td>0</td>\n",
       "      <td>531</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>60</th>\n",
       "      <td>0</td>\n",
       "      <td>199</td>\n",
       "      <td>12</td>\n",
       "      <td>23</td>\n",
       "      <td>473</td>\n",
       "      <td>68</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>197</td>\n",
       "      <td>28</td>\n",
       "      <td>...</td>\n",
       "      <td>412</td>\n",
       "      <td>166</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>96</td>\n",
       "      <td>37</td>\n",
       "      <td>3</td>\n",
       "      <td>169</td>\n",
       "      <td>190</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>71</th>\n",
       "      <td>447</td>\n",
       "      <td>312</td>\n",
       "      <td>0</td>\n",
       "      <td>25</td>\n",
       "      <td>404</td>\n",
       "      <td>151</td>\n",
       "      <td>0</td>\n",
       "      <td>132</td>\n",
       "      <td>112</td>\n",
       "      <td>25</td>\n",
       "      <td>...</td>\n",
       "      <td>296</td>\n",
       "      <td>240</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>10</td>\n",
       "      <td>248</td>\n",
       "      <td>0</td>\n",
       "      <td>400</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>0</td>\n",
       "      <td>497</td>\n",
       "      <td>0</td>\n",
       "      <td>35</td>\n",
       "      <td>378</td>\n",
       "      <td>54</td>\n",
       "      <td>0</td>\n",
       "      <td>109</td>\n",
       "      <td>321</td>\n",
       "      <td>42</td>\n",
       "      <td>...</td>\n",
       "      <td>579</td>\n",
       "      <td>369</td>\n",
       "      <td>0</td>\n",
       "      <td>50</td>\n",
       "      <td>0</td>\n",
       "      <td>39</td>\n",
       "      <td>9</td>\n",
       "      <td>71</td>\n",
       "      <td>409</td>\n",
       "      <td>254</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>92</th>\n",
       "      <td>397</td>\n",
       "      <td>228</td>\n",
       "      <td>0</td>\n",
       "      <td>72</td>\n",
       "      <td>426</td>\n",
       "      <td>186</td>\n",
       "      <td>7</td>\n",
       "      <td>0</td>\n",
       "      <td>104</td>\n",
       "      <td>31</td>\n",
       "      <td>...</td>\n",
       "      <td>442</td>\n",
       "      <td>310</td>\n",
       "      <td>0</td>\n",
       "      <td>19</td>\n",
       "      <td>62</td>\n",
       "      <td>0</td>\n",
       "      <td>20</td>\n",
       "      <td>262</td>\n",
       "      <td>0</td>\n",
       "      <td>518</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>51</th>\n",
       "      <td>440</td>\n",
       "      <td>325</td>\n",
       "      <td>0</td>\n",
       "      <td>57</td>\n",
       "      <td>432</td>\n",
       "      <td>212</td>\n",
       "      <td>58</td>\n",
       "      <td>76</td>\n",
       "      <td>126</td>\n",
       "      <td>12</td>\n",
       "      <td>...</td>\n",
       "      <td>305</td>\n",
       "      <td>316</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>55</td>\n",
       "      <td>6</td>\n",
       "      <td>23</td>\n",
       "      <td>118</td>\n",
       "      <td>0</td>\n",
       "      <td>435</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>100 rows × 1093 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "    6     14    26    38    68    73    97    100   130   133   ...  9953  \\\n",
       "83   467   287     0    33   396   141   141   119   103     5  ...   311   \n",
       "53   100   297     0     1   556   173     0    68   124    19  ...   806   \n",
       "70   426   340     0    24   415   219   306   155   131     4  ...   321   \n",
       "45    65   297     0    13   669   120     0    43   117    11  ...   816   \n",
       "44   382   217     0    56   456   137    46    39    97     9  ...   404   \n",
       "..   ...   ...   ...   ...   ...   ...   ...   ...   ...   ...  ...   ...   \n",
       "60     0   199    12    23   473    68     0     0   197    28  ...   412   \n",
       "71   447   312     0    25   404   151     0   132   112    25  ...   296   \n",
       "14     0   497     0    35   378    54     0   109   321    42  ...   579   \n",
       "92   397   228     0    72   426   186     7     0   104    31  ...   442   \n",
       "51   440   325     0    57   432   212    58    76   126    12  ...   305   \n",
       "\n",
       "    9961  9968  9969  9971  9979  9987  9994  9995  9999  \n",
       "83   228     0     0     0     0    57   240     0   458  \n",
       "53   123     0     0     3     0     0     0   221   364  \n",
       "70   236     0     0     0     0    83   129     0   409  \n",
       "45   133     0     0     7     2    31     0   195   292  \n",
       "44   291     0     0     8     0    37   153     0   531  \n",
       "..   ...   ...   ...   ...   ...   ...   ...   ...   ...  \n",
       "60   166     0     0     0    96    37     3   169   190  \n",
       "71   240     0     0     0     0    10   248     0   400  \n",
       "14   369     0    50     0    39     9    71   409   254  \n",
       "92   310     0    19    62     0    20   262     0   518  \n",
       "51   316     0     0    55     6    23   118     0   435  \n",
       "\n",
       "[100 rows x 1093 columns]"
      ]
     },
     "execution_count": 68,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train_important = pd.DataFrame(X_train, columns=X_train.columns[features_non_zero], index=X_train.index)\n",
    "X_train_important"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Hyperparameters optimzation**, but use only the **attributes with predictive power**:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 30 candidates, totalling 150 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=4)]: Using backend LokyBackend with 4 concurrent workers.\n",
      "[Parallel(n_jobs=4)]: Done  44 tasks      | elapsed:    0.8s\n",
      "[Parallel(n_jobs=4)]: Done 150 out of 150 | elapsed:    1.7s finished\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "RandomizedSearchCV(cv=5, error_score=nan,\n",
       "                   estimator=SVC(C=1.0, break_ties=False, cache_size=200,\n",
       "                                 class_weight=None, coef0=0.0,\n",
       "                                 decision_function_shape='ovr', degree=3,\n",
       "                                 gamma='scale', kernel='rbf', max_iter=-1,\n",
       "                                 probability=False, random_state=None,\n",
       "                                 shrinking=True, tol=0.001, verbose=False),\n",
       "                   iid='deprecated', n_iter=30, n_jobs=4,\n",
       "                   param_distributions={'C': <scipy.stats._distn_infrastructure.rv_frozen object at 0x00000292630C7908>,\n",
       "                                        'gamma': <scipy.stats._distn_infrastructure.rv_frozen object at 0x0000029262CDB848>,\n",
       "                                        'kernel': ['linear', 'poly', 'rbf',\n",
       "                                                   'sigmoid']},\n",
       "                   pre_dispatch='2*n_jobs', random_state=42, refit=True,\n",
       "                   return_train_score=False, scoring='f1', verbose=2)"
      ]
     },
     "execution_count": 69,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.model_selection import RandomizedSearchCV\n",
    "from scipy.stats import expon, reciprocal\n",
    "\n",
    "# Note: gamma is ignored when kernel is \"linear\"\n",
    "param_distribs = {\n",
    "        'kernel': ['linear', 'poly', 'rbf', 'sigmoid'],\n",
    "        'C': reciprocal(10, 10000),\n",
    "        'gamma': expon(scale=1.0),\n",
    "    }\n",
    "\n",
    "svm_clf = SVC()\n",
    "rnd_search = RandomizedSearchCV(svm_clf, param_distributions=param_distribs,\n",
    "                                n_iter=30, cv=5, scoring='f1',\n",
    "                                verbose=2, n_jobs=4, random_state=42)\n",
    "rnd_search.fit(X_train_important, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'C': 2183.096839052459, 'gamma': 0.9084469696321253, 'kernel': 'poly'}"
      ]
     },
     "execution_count": 70,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rnd_search.best_params_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_valid_important = pd.DataFrame(X_valid, columns=X_valid.columns[features_non_zero], \n",
    "                                 index=X_valid.index)\n",
    "\n",
    "y_valid_predict = rnd_search.predict(X_valid_important)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[51,  5],\n",
       "       [12, 32]], dtype=int64)"
      ]
     },
     "execution_count": 72,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "confusion_matrix(y_valid, y_valid_predict)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Precision score:\t 0.8648648648648649\n",
      "Recall score:\t\t 0.7272727272727273\n",
      "F1 score:\t\t 0.7901234567901235\n"
     ]
    }
   ],
   "source": [
    "performance_scores(y_valid, y_valid_predict)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {},
   "outputs": [],
   "source": [
    "precisions, recalls, thresholds = precision_recall_curve(y_valid, y_valid_predict)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "show_recall_vs_precision_curve(recalls, precisions)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXQAAAEKCAYAAAACS67iAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjMsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+AADFEAAAgAElEQVR4nO3deXyU5bn/8c9FIICALGEHWUSUnSREBdzwIAoiqBQFj/ZQl3qsB6vHHltb7a9qtbWVX61W+qPWKrV6FKWnx6CoCIq0FBXIwr4LEgkQVtlDkvv3xz0ZZrJOyCSTDN/365UXM8/zzDNXnoSLm2vu57rNOYeIiNR/DWIdgIiIRIcSuohInFBCFxGJE0roIiJxQgldRCROKKGLiMSJShO6mb1sZrvNbFU5+83MnjezTWa2wsxSox+miIhUJpIR+kxgdAX7xwC9A193A/+v+mGJiEhVVZrQnXOLgH0VHHI98KrzPgNamVmnaAUoIiKRaRiFc3QBtoc8zwlsyy15oJndjR/F06xZsyF9+vSp8pvtOLSD3EO5xefjrEZnhX01bdgUMzuNb0NEpO5bvnz5Hudcu7L2RSOhl5U9y+wn4Jx7EXgRIC0tzS1btqzKb1ZYVMj6vevJyM0gIzeDzJ2ZZORmkHciD4BGDRoxoP0AUjqmkNopldROqQzqMIhmic2q/F4iInWNmW0rb180EnoOcE7I867Ajiict0wJDRLo164f/dr147ZBtwFQ5Ir4cv+XwSSfsTOD9A3pvJz1MgANrAF92vbxCb6jT/LJHZNp2aRlTYUpIlLropHQ04GpZvYmcDFw0DlXqtxSkxpYA3q16UWvNr24qf9NADjnyPkmJziCz8jN4JMvP+G1Fa8FX9erda/gKD61UyopHVNo16zM/8mIiNR5Vlm3RTN7AxgBtAV2AT8DGgE452aYL1i/gJ8JcxS43TlXaS3ldEsu1bXr8K6wJJ+5M5Mt+7cE959z9jnB5F6c6Du36Ky6vIjUCWa23DmXVua+WLXPjVVCL8v+Y/vJ2pkVLNdk5Gawfs96XOCjgPbN2oeVa1I7pdKjVQ8leRGpdUrop+Fw/mGyd2YHk3xmbiar81ZTUFQAQKsmrcJG8amdUundpjcJDRJiHLmIxDMl9Cg5XnCcVbtXnfrwNTeDFbtWcKLwBADNGjUjuWNyWE2+X7t+NEpoFOPIRSReKKHXoJOFJ1m7Zy2ZuZlho/kjJ48A0DihMQM7DAwr1wzsMJAmDZvEOHIRqY+U0GtZYVEhm/ZtCptGmZmbyf7j+wFIsAT6t+8fVrIZ3GEwLRq3iHHkIlLXKaHXAc45th3cFlauWZ67nN1HdgNgGOcnnV9qGmXrpq1jHLmI1CVK6HWUc47cw7l++mRuZnCGzVcHvwoe06NVj1IzbDo07xDDqEUklpTQ65k9R/eE1eQzcjPYtG9TcH/nFp1LzZU/5+xzNI1S5AyghB4HDh4/SPau7LCSzdo9aylyRQAkNU0KK9ekdkrl3Nbn0sC0holIPFFCj1NHTx5lxa4VYXe9rty1kpNFJwFokdiClE4pYeWaC9peQMMG0ej4ICKxoIR+BskvzGf17tVhM2yyd2ZzrOAYAE0bNmVwx8FhSb5/+/4kJiTGOHIRiYQS+hmuoKiA9XvWl+ph882Jb4BTLYdDyzWDOgzirEZnxThyESlJCV1KKXJFbNm/Jawmn5Gbwd5jewHfwbJv275hH76q5bBI7CmhS0SKWw6HlmsycjPYcehUe/vz2pwXNo0ypVMKbc9qG8OoRc4sSuhSLTsP7yQzNzOsZPPlgS+D+4tbDod+dWreSdMoRWqAErpE3b5j+061HA58bdi7IdhyuEOzDmF3vKrlsEh0KKFLrSjZcjgjN4PVu1dT6AoB33K45F2vvZN6a668SBUooUvMHC84zspdK8Pq8it3rQy2HG6e2JzBHQaHlWv6tu2rlsMi5VBClzqluOVwaLkma2dWWMvhQR0GhSX5Ae0HqOWwCEroUg8UFhWycd/GUo3KDhw/AEDDBg3p165fWMlmcMfBNE9sHuPIRWqXErrUS845th7YWmoaZWjL4QvaXhD2wataDku8U0KXuBHacjj0a/s324PH9GzVs1RfebUclnihhC5xL+9IXnCefPGfZbUcDp1h0/XsrppGKfWOErqckQ4eP3hqrnygXLNuz7pgy+G2Z7UNJvmUTilqOSz1ghK6SEDJlsMZuRms2r0q2HL47MZnhy0cktoplQuSLiChQUKMI5f6bOdOaN0aGjeGvXshKen0z6WELlKBEwUnWJ13quVw5s5MtRyWanEOVq2COXMgPR0+/9w/vu46OHoUzqpGI1MldJEqKm45HDrDJjM3k0P5hwDfcnhgh4FhSX5Qh0E0bdQ0xpFLrO3YAZdcAlu3+ucXXgjjx8Ntt0GPHtU/vxK6SBQUuSI279scNpJfnrucfcf2AZBgCfRp2yesXJPcMZmzG58d48ilpuzfD++/70fh55wDzzzjR+dTpsBll/kReadO0X1PJXSRGuKcY/s320tNo8w9nBs8pneb3qUalSWdVY0iqsTcK6/Aq6/C3/8OhYXQoQPcfjv88pc1/94VJXQtLilSDWZGt5bd6NayGzf0uSG4vbjlcHG55vOvP2fW6lnB/d1adis1jbJTiygP5SQqCgt9DXz+fPjpT8HMP9+zB370I19OufBCaFAHJkdphC5SS/Yd2xeW5DNzM8NaDnds3rHUDJvuLbtrrnwMHDkC8+b5Usp770FeHjRsCKtXw/nnw8mT0ChG/eNUchGpow6dOET2ruywcs2avDXBlsOtm7QudderWg7XjJwcn6Q7dIB334Vx46BVK7j2Wj8KHz0aWtaBFRiV0EXqkWMnj7Fy98qw0fyKXSvIL8wHfMvh5I7JYeWavu360rCBKqhV4RxkZflReHo6ZGTAz34Gjz0Gx475ssoll8RuJF4eJXSReu5k4UnW5K0Jm0aZtTOLoyePAtCkYRMGdRgUVrJRy+HSnPM18KIi6NsXNmzwz4cP9yPyCROgd+9YR1kxJXSROBTacjj06+CJg4BvOdy/Xf+wks3gDoNpltgsxpHXrrw8mDvXj8L37oWFC/32X/7STym89lpo3z6mIVaJErrIGcI5x5cHvgyfK79jOXlH8wDfcrhP2z6+d02gZJPSKYVWTVrFOPLo+9vf4De/gX/+04/IO3f2tfAXXoCEetzJodrTFs1sNPAckAC85Jx7usT+bsCfgVaBYx52zs2tVtQiUmVmxrmtz+Xc1ucysd9EwCf5HYd2hJVrFm1bxH+v/O/g685tfW7YNMqUTim0b1Z/hq0FBbB4sR+F/+AHPnnv2eNnq/z0p76ckprqyyvxrNIRupklABuAUUAOsBS4xTm3JuSYF4FM59z/M7N+wFznXI+KzqsRukhshbYcLv7avH9zcH+XFl3CyjWpnVLp0qJLnZlGefSon42Snu5LKvv3Q2Ii/M//wNixp+rl8aa6I/SLgE3OuS2Bk70JXA+sCTnGAcX3N7cEdpx+uCJSG9o1a8fVva7m6l5XB7cdOH7gVMvhQMnmvY3vBVsOtzurXdgdr8Uth2sryW/dCocPw4ABPoFPmuQ7F44f779GjYIWLfyx8ZjMKxPJCH0iMNo5d1fg+beBi51zU0OO6QTMA1oDzYCrnHPLyzjX3cDdAN26dRuybdu2aH0fIlJDjuQfCW85vNO3HC4oKgCgZeOWYTX51E6pnJ90flRaDhcVwbJlp6YWrlzp+6PMmeP3Z2TA4MH1uyZeVdX6UNTMbgKuKZHQL3LO3RdyzIOBc/1fMxsG/AkY4Fzgn/UyqOQiUn+dKDjBqt2rwko22buyOV5wHICzGp3F4A6Dw8o1/dr1i6jlcOhdmKNHw4cf+oR96aV+FD5uXN2fWliTqpvQhwGPOeeuCTz/MYBz7pchx6zGj+K3B55vAYY653aXd14ldJH4UlBUwLo968Jq8lk7s4IthxMTEhnYfmBYyaa45XBurq+Hz5njP9zcvt33DP/rX+H4cRgzBtq0ifE3WEdUN6E3xH8oOhL4Gv+h6L8651aHHPM+MMs5N9PM+gILgC6ugpMroYvEv5Ith4uXAgxtOZz4TV+ObU6F3FQ6uFRuuDiZn/+0Be3axTj4Oqra89DN7Frgt/gpiS87554ysyeAZc659MDMlj8CzfEfkP7QOTevonMqoYucWfLzYdEieCfdMXzMVzTpkcF7mRnMXpxJUfsMDnGq5fD5SeeHffCa0jFFLYcDdGORiMREQQHMmuU/0PzgA/jmG2jSxC8EMXVq+LG5h3JLTaPcdvDUxInuLbuXalR2JrYcVkIXkVqzaZOfXnjVVX4ueNeuvqf4uHH+Q82RIyNfU3Pv0b1k7sz0jcoC5ZoNezcE93ds3rFUX/luLbvVmbnyNUEJXURqTGEhfPbZqamF69b5OzVzcvxc8G3b/PJs0VoA4psT35C9MzusJr82b22w5XCbpm1KzZU/r815cdNyWAldRKLq8GFo1swn7Pvvh+ef9wtAjBhxamphNBZEjlRxy+HQcs3K3SvDWg6XXDykT9s+9bLlsBK6iFRbTo6fVpieDh9/DEuW+P4oGRmwcWPdWQCiWH5hfrDlcHHJpqyWw6HlmgHtB9C4YeMYR14xJXQROW0bN/pb7DMz/fNeveD66+Hee/3j+qSwqJANezeETaPMzM0Mazk8oP0AUjv6BmV1seWwErqIROT4cfjkEz8K79/fz0Q5etTfbj96tC+l9OkTX31SSrYczsjNYHnucvYc3QNAA2vABUkXhJVrkjsmx6zlsBK6iFTojTdg9mx/m/2RI74+PnUqPP105a+NR845vj70dViSz9yZSc43OcFjerXuVerD13bNav5uKCV0EQlyzs9EWbIE7rjDbxs3zpdUirsWjhjh54tLuN1Hdoet9ZqRm8GW/VuC+7ue3bXUNMrOLTpHdRqlErrIGa6gAP7xD19KmTPHzxU3g507/fJrBw74DzTjqZRSW/Yf2x9sOVx8Y9S6Petw+Nxa3HI49Ktnq56nneSV0EXOQAcP+i6FzZvDSy/Bd7/rF4AYOdKPwq+7zt/0I9F3OP9weMvh3AxW560Othx+9ppneWDoA6d17movQSci9cOXX/oR+Jw5fjHk6dPh7rt9Ak9K8gtANG8e6yjjX/PE5gw/ZzjDzxke3Ha84Dird68mIzeDS7tdWiPvq4QuEgeOHYOhQ2HFCv+8Tx948EEYNsw/b98ebrwxdvGJn/c+pPMQhnQeUmPvoYQuUs8cPQrz5/tReFER/OlP0LSpT95TpmgBiDOZErpIPZGe7mvhH33k54uffTZMmHBqMeQZM2IdocRafHSrEYkzzvnyyVNP+RE5+GmFK1f6mvhHH0FeHrzyimamyCma5SJSR+Tnw6efnuqXUryG+ief+HnhJ074WSpK4Gc2zXIRqaP27fMj8K5dYdUquPpqXw8fNQoefRTGjoVOgTUcGtftnlFSByihi9SyjRtP9Q5fvNjfrfnii5CSAnPnwhVXRL4AhEgoJXSRWjRypG89CzBwIDz8MHzrW/65mV/dXuR0KaGL1IBDh2DePF8Pz8ryPcMbNIAbbvBftb0AhJwZlNBFoujTT32Hwo8/9h9ytm4N117rE3zLlnDffbGOUOKZErrIaXLOj7znzIGbbvL9ww8d8o2vpk71t9tfcolfmk2kNuhXTaQKTp70c8CL+6V8/bUvpXTp4hP62LH+S1MLJRaU0EUqkZcHO3bA4ME+oU+c6JP4Ndf4Ufi110K7wLoGSuQSS0roIiU4B2vXnuodvmSJn1K4fLmfTvj3v/vRuBaAkLpGCV0EKCz0vcPB31r/0kv+8ZAh8LOf+ZF4sSE11yxPpFqU0OWMdfAgfPCBH4m//77vk9Kli58XPmSIFoCQ+kcJXc44q1bBAw/4KYYFBb7+fcMNfpoh+NXtReojJXSJa0VFsHSpH4WnpflFHtq0gdxc+MEPfCnl4otPlVtE6jMldIlLc+bAO+/Au+/Crl0+Yf/whz6hd+4Mq1fHOkKR6FNCl7iQmwvZ2afKJY8/7ptgjRnjR+Fjxvi7NkXimRK61EvO+Q8xi7sWLl3qpxHu3eunFs6e7UfiiYmxjlSk9mjFIqk3TpzwN/YATJvmb/T56U/9TT5PPQVffOF7iYNvfKVkLmcajdClTtu7108pTE/3Uwz/+7/9dMLrr/cfbo4dCx07xjpKkbohooRuZqOB54AE4CXn3NNlHHMz8BjggGzn3L9GMU45w+zZ4+eD/+MffqZKx44wefKpeeHnn++/ROSUShO6mSUA04FRQA6w1MzSnXNrQo7pDfwYuMQ5t9/M2tdUwBJ/Cgv97fXp6dCihS+jJCX5Jdd+8hP/oeaQIb60IiLli2SEfhGwyTm3BcDM3gSuB9aEHPNdYLpzbj+Ac253tAOV+PPBB/DGG/Dee7600qiRH4WDb3I1b15s4xOpbyIZ83QBtoc8zwlsC3U+cL6ZLTazzwIlmlLM7G4zW2Zmy/Ly8k4vYqm3vvoK/vhHP0MF4G9/8/PEx4yBt97yZZZXX41tjCL1WSQj9LIagroyztMbGAF0Bf5uZgOccwfCXuTci8CLAGlpaSXPIXGmeAGI4qmFWVl++4UXQnKyX9ln+nQtACESLZH8VcoBzgl53hXYUcYxnznnTgJfmtl6fIJfGpUopd44ftxPL2zZ0s9OGTvW176HD4df/9rXwy+4wB+rG31EoiuSkstSoLeZ9TSzRGAykF7imP8FrgQws7b4EsyWaAYqddfu3fDKK/62+rZt4be/9dtHjICZM2HnTt9D/KGHTiVzEYm+SkfozrkCM5sKfIiftviyc261mT0BLHPOpQf2XW1ma4BC4CHn3N6aDFxizzkYNcoviOycn1I4ZYrfBv6OzSlTYhujyJnEnItNKTstLc0tW7YsJu8tVXfypJ8Tnp4O27f7W+vBj7pbtIBx43xdXEuwidQsM1vunEsra58+jpIKLVoEM2b4eviBA35u+FVX+d7hiYnwzDOxjlBEiulWDQmzZQs895yvewOsWwfz5/v6+N/+5qcWvvuu+qSI1EUaoZ/hCgt9U6s5c3w5pbhPeLt28K//6mvgd96pBSBE6gMl9DPQkSP+zsxu3fziD8OH+4R9+eVw112+Ht6rlz+2cePYxioikVNCP0Pk5vpSSXq6L6FcdZUflXfu7LcPH6554SL1nRL6GeA734E//9k/7tkT/v3ffU282NixMQlLRKJMCT2OnDjhV7JPT4cFC2D5cj8X/Ior/A0948dDv36aWigSr5TQ40B2tl+x54MP4NAhv2rP1VefWo7t9ttjHaGI1AYl9HpowwY/Ch8+3H8VFPibfm65xY/C/+VfTi3FJiJnDiX0eqCoCBYvPjW1cP16v/1nP/MJPTUVcnK0AITImU4JvY46dAg2bvTJGvzCD3l5cOWVMHWqn1rYvbvfZ6a6uIgoodcpX311ahS+cKHvXLh9ux95z5kD550HZ58d6yhFpK5SQo+hoqJTo+snnvAlFIDeveG++3w9vFjxSF1EpDxK6LXs2DHfbjY93Y+658zxCyBfdZWfkTJunHqGi8jpUUKvJTk5ftQ9bx4cPQrNm8M115zqkVI8Y0VE5HQpodcA52DNGj8K79TJ36nZpo3vXHj77X4UPmKE+qSISHQpoUfRokXwP//jyyhbAgvwffvbPqGfdRasXRvT8EQkzmnmcjUcOOAXfij2zDN+MYg+ffyfOTnw6quxi09EziwaoVfRli2nPtBctMjfpZmTA126wAsv+KmGzZrFOkoRORMpoVeiqMgn7cREeOMNv+gDQP/+8F//5acWduzotxXf6CMiEgsquZThyBH43//1K/V06gQzZ/rtV14Jzz4LmzbBqlXwy1/CsGFazUdE6gaN0EOcPAkTJsBHH/lWtC1bwpgxp+aFd+wIDzwQ2xhFRMpzxiZ053zb2Tlz4Jtv/AeajRr50so99/hSymWX+W0iIvXBGZfQlyyB117ziXz7dn/b/ZVX+gRvBn/9a6wjFBE5PXFfQ9+7F/7yF393Jvg7NWfO9Lfb/+lPfq3NBQvUrVBE6r+4HKEXLwCRnu77iBcVQVISXHutr4H/8IdaAEJE4k9cJPSCAt8/vHVrWLECBg/225OT4ZFHfD28uFthy5axi1NEpCbV24T+zTe+fJKeDu+9BxMnwh/+AAMHwosv+sZX3brFOkoRkdpTLxP6lCn+Jp+TJ30p5brr4IYb/D4z+O53YxufiEgs1MuE3rs33H+/L6UMGwYN6+V3ISISXfUyFT76aKwjEBGpe+J+2qKIyJlCCV1EJE4ooYuIxAkldBGROBFRQjez0Wa23sw2mdnDFRw30cycmaVFL0QREYlEpQndzBKA6cAYoB9wi5n1K+O4FsD3gc+jHaSIiFQukhH6RcAm59wW51w+8CZwfRnH/Rz4NXA8ivGJiEiEIknoXYDtIc9zAtuCzCwFOMc5925FJzKzu81smZkty8vLq3KwIiJSvkgSelmNZV1wp1kD4FngB5WdyDn3onMuzTmX1q5du8ijFBGRSkWS0HOAc0KedwV2hDxvAQwAFprZVmAokK4PRkVEalckCX0p0NvMeppZIjAZSC/e6Zw76Jxr65zr4ZzrAXwGjHfOLauRiEVEpEyVJnTnXAEwFfgQWAu85ZxbbWZPmNn4mg5QREQiE1FzLufcXGBuiW3/p5xjR1Q/LBERqSrdKSoiEieU0EVE4oQSuohInFBCFxGJE0roIiJxQgldRCROKKGLiMQJJXQRkTihhC4iEieU0EVE4oQSuohInFBCFxGJE0roIiJxQgldRCROKKGLiMQJJXQRkTihhC4iEieU0EVE4oQSuohInFBCFxGJE0roIiJxQgldRCROKKGLiMQJJXQRkTihhC4iEieU0EVE4oQSuohInFBCFxGJE0roIiJxQgldRCROKKGLiMQJJXQRkTihhC4iEieU0EVE4kRECd3MRpvZejPbZGYPl7H/QTNbY2YrzGyBmXWPfqgiIlKRShO6mSUA04ExQD/gFjPrV+KwTCDNOTcImA38OtqBiohIxSIZoV8EbHLObXHO5QNvAteHHuCc+8Q5dzTw9DOga3TDFBGRykSS0LsA20Oe5wS2ledO4P2ydpjZ3Wa2zMyW5eXlRR6liIhUKpKEbmVsc2UeaHYbkAY8U9Z+59yLzrk051xau3btIo9SREQq1TCCY3KAc0KedwV2lDzIzK4CHgGucM6diE54IiISqUhG6EuB3mbW08wSgclAeugBZpYC/AEY75zbHf0wRUSkMpUmdOdcATAV+BBYC7zlnFttZk+Y2fjAYc8AzYG3zSzLzNLLOZ2IiNSQSEouOOfmAnNLbPs/IY+vinJcIiJSRREl9Npy8uRJcnJyOH78eKxDkRrSpEkTunbtSqNGjWIdikjcqVMJPScnhxYtWtCjRw/MyppcI/WZc469e/eSk5NDz549Yx2OSNypU71cjh8/TlJSkpJ5nDIzkpKS9D8wkRpSpxI6oGQe5/TzFak5dS6hi4jI6VFCLyEhIYHk5GQGDBjATTfdxNGjRyt/USWWLVvG97///XL379ixg4kTJ1b7fUTkzKaEXkLTpk3Jyspi1apVJCYmMmPGjLD9zjmKioqqdM60tDSef/75cvd37tyZ2bNnn1a8IiLF6nRCHzGi9Nfvf+/3HT1a9v6ZM/3+PXtK76uqyy67jE2bNrF161b69u3LvffeS2pqKtu3b2fevHkMGzaM1NRUbrrpJg4fPgzA0qVLGT58OIMHD+aiiy7i0KFDLFy4kOuuuw6ATz/9lOTkZJKTk0lJSeHQoUNs3bqVAQMGAP6D4dtvv52BAweSkpLCJ598AsDMmTOZMGECo0ePpnfv3vzwhz+s+jckInGtTif0WCooKOD9999n4MCBAKxfv55/+7d/IzMzk2bNmvHkk08yf/58MjIySEtL4ze/+Q35+flMmjSJ5557juzsbObPn0/Tpk3Dzjtt2jSmT59OVlYWf//730vtnz59OgArV67kjTfeYMqUKcFZIVlZWcyaNYuVK1cya9Ystm/fjohIsTo1D72khQvL33fWWRXvb9u24v3lOXbsGMnJyYAfod95553s2LGD7t27M3ToUAA+++wz1qxZwyWXXAJAfn4+w4YNY/369XTq1IkLL7wQgLPPPrvU+S+55BIefPBBbr31ViZMmEDXruGt4//xj39w3333AdCnTx+6d+/Ohg0bABg5ciQtW7YEoF+/fmzbto1zzjkHERGo4wk9Fopr6CU1a9Ys+Ng5x6hRo3jjjTfCjlmxYkWl0/Iefvhhxo4dy9y5cxk6dCjz58+nSZMmYecuT+PGjYOPExISKCgoqPT7EZEzh0oup2Ho0KEsXryYTZs2AXD06FE2bNhAnz592LFjB0uXLgXg0KFDpZLu5s2bGThwID/60Y9IS0tj3bp1Yfsvv/xyXn/9dQA2bNjAV199xQUXXFAL35WI1HdK6KehXbt2zJw5k1tuuYVBgwYxdOhQ1q1bR2JiIrNmzeK+++5j8ODBjBo1qtRdkb/97W8ZMGAAgwcPpmnTpowZMyZs/7333kthYSEDBw5k0qRJzJw5M2xkLiJSHqvov/g1KS0tzS1btixs29q1a+nbt29M4pHao5+zyOkzs+XOubSy9mmELiISJ5TQRUTihBK6iEicUEIXEYkTSugiInFCCV1EJE4ooZcQ2j533LhxHDhwIKrnnzlzJlOnTgXgscceY9q0acF9S5Ys4bvf/S5Hjx7l1ltvZeDAgQwYMIBLL7002PwrGkLb+S5cuJB//vOfwX0zZszg1VdfrfD1d911F2vWrAHgF7/4RdTiEpHq0a3/JYTe+j9lyhSmT5/OI488Uivv/cEHHzB69Giee+45OnTowMqVKwHfGCyaiyqnpaWRluansS5cuJDmzZszfPhwAO65555KX//SSy8FH//iF7/gJz/5SdRiE5HTV2cT+gMfPEDWztI9VaojuWMyvx3924iPHzZsGCtWrAg+f+aZZ3jrrbc4ceIEN954I48//jgAr776KtOmTcPMGDRoEH/5y1+YM2cOTz75JPn5+SQlJfH666/ToUOHCt9vwYIFPPjgg3z66ad07949uAi3cCUAAA4qSURBVD301v/XXnuN559/nvz8fC6++GJ+//vfk5CQQPPmzbn//vt59913adq0Ke+88w4dOnTg7bff5vHHHychIYGWLVuyaNEiFi5cyLRp03jhhReYMWMGCQkJvPbaa/zud79jwYIFNG/enLFjxzJlyhS++OILALZu3cr48eNZsWIFI0aMYNq0acyePTvYzKx///6ce+65tG3blvvvvx+ARx55hA4dOlS4uIeIRI9KLuUoLCxkwYIFjB8/HoB58+axceNGvvjiC7Kysli+fDmLFi1i9erVPPXUU3z88cdkZ2fz3HPPAXDppZfy2WefkZmZyeTJk/n1r39d4fvt2bOHRo0a0bJlS+644w5+9atfMWzYMB599FE2btwI+DssZ82axeLFi8nKyiIhISHY9+XIkSMMHTqU7OxsLr/8cv74xz8C8MQTT/Dhhx+SnZ1Nenp62Hv26NGDe+65h//8z/8kKyuLyy67LLivb9++5Ofns2XLFgBmzZrFzTffHPb6p59+Ovg/mtdff50777yTP//5zwAUFRXx5ptvcuutt57W9ReRqquzI/SqjKSjqXjEuXXrVoYMGcKoUaMAn9DnzZtHSkoKAIcPH2bjxo1kZ2czceJE2rZtC0CbNm0AyMnJYdKkSeTm5pKfn0/Pnj0rfN958+Zx9dVXA5CcnMyWLVuYN28e8+fP58ILL2TJkiUsWLCA5cuXB9vzHjt2jPbt2wOQmJgYXERjyJAhfPTRR4Bv1/ud73yHm2++mQkTJlTpWtx888289dZbPPzww8yaNYtZs2ZVeHyPHj1ISkoiMzOTXbt2kZKSQlJSUpXeU0ROn0boJRSPOLdt20Z+fn5wwQnnHD/+8Y/JysoiKyuLTZs2ceedd+KcK7Nl7n333cfUqVNZuXIlf/jDH0o16Srp/fffZ/To0cHnzZs3Z8KECfz+97/ntttuY+7cuTjnmDJlSjCG9evX89hjjwHQqFGjYByhrXVnzJjBk08+yfbt20lOTmbv3r0RX4tJkybx1ltvsWHDBsyM3r17V/qau+66i5kzZ/LKK69wxx13RPxeIlJ9SujlaNmyJc8//zzTpk3j5MmTXHPNNbz88svB2SZff/01u3fvZuTIkbz11lvBRLlv3z4ADh48SJcuXQCCZYjyOOdYsWJFcGGNxYsXs3//fsAvnrFmzRq6d+/OyJEjmT17Nrt37w6+17Zt2yo89+bNm7n44ot54oknaNu2balVjlq0aMGhQ4fKfG2vXr1ISEjg5z//OZMmTSrzmEaNGnHy5Mng8xtvvJEPPviApUuXcs0111QYm4hEV50tudQFKSkpDB48mDfffJNvf/vbrF27lmHDhgF+BP3aa6/Rv39/HnnkEa644goSEhJISUlh5syZPPbYY9x000106dKFoUOH8uWXX5b7PsuXLyclJSU4wt68eTPf+973ggtSjx07lm9961uYGU8++SRXX301RUVFNGrUiOnTp4d9gFrSQw89xMaNG3HOMXLkSAYPHsynn34a3D9u3DgmTpzIO++8w+9+97tSr580aRIPPfRQufHffffdDBo0iNTUVF5//XUSExO58soradWqFQkJCRFdZxGJDrXPrQOefPJJzjvvPCZPnhzrUKqtqKiI1NRU3n777XJLNGfqz1kkGipqn6sReh3w6KOPxjqEqFizZg3XXXcdN954Y0T1dhGJLiV0iZp+/foFpzmKSO2rcx+KxqoEJLVDP1+RmlOnEnqTJk3Yu3ev/tLHKecce/fupUmTJrEORSQu1amSS9euXcnJySEvLy/WoUgNadKkCV27do11GCJxqU4l9EaNGlV6R6WIiJQtopKLmY02s/VmtsnMHi5jf2MzmxXY/7mZ9Yh2oCIiUrFKE7qZJQDTgTFAP+AWM+tX4rA7gf3OufOAZ4FfRTtQERGpWCQj9IuATc65Lc65fOBN4PoSx1wPFN/fPhsYaWU1OBERkRoTSQ29CxDaACQHuLi8Y5xzBWZ2EEgC9oQeZGZ3A3cHnh42s/WnEzTQtuS56wjFVTWKq+rqamyKq2qqE1e5vT4iSehljbRLziuM5Biccy8CL0bwnhUHZLasvFtfY0lxVY3iqrq6GpviqpqaiiuSkksOcE7I867AjvKOMbOGQEtgXzQCFBGRyESS0JcCvc2sp5klApOB9BLHpANTAo8nAh873R0kIlKrKi25BGriU4EPgQTgZefcajN7AljmnEsH/gT8xcw24UfmNd02sNplmxqiuKpGcVVdXY1NcVVNjcQVs/a5IiISXXWql4uIiJw+JXQRkThRZxO6md1kZqvNrMjMyp3eU15bgsCHuJ+b2cZAW4LEKMXVxsw+Cpz3IzNrXcYxV5pZVsjXcTO7IbBvppl9GbIvubbiChxXGPLe6SHbY3m9ks1sSeDnvcLMJoXsi+r1qk4bCzP7cWD7ejOL6oKpEcT1oJmtCVyfBWbWPWRfmT/TWorrO2aWF/L+d4XsmxL4uW80syklX1vDcT0bEtMGMzsQsq8mr9fLZrbbzFaVs9/M7PlA3CvMLDVkX/Wvl3OuTn4BfYELgIVAWjnHJACbgXOBRCAb6BfY9xYwOfB4BvC9KMX1a+DhwOOHgV9Vcnwb/AfFZwWezwQm1sD1iigu4HA522N2vYDzgd6Bx52BXKBVtK9XRb8vIcfcC8wIPJ4MzAo87hc4vjHQM3CehFqM68qQ36HvFcdV0c+0luL6DvBCGa9tA2wJ/Nk68Lh1bcVV4vj78JM5avR6Bc59OZAKrCpn/7XA+/h7d4YCn0fzetXZEbpzbq1zrrI7SctsS2BmBvwLvg0B+LYEN0QptNA2B5GcdyLwvnPuaJTevzxVjSso1tfLObfBObcx8HgHsBtoF6X3D1WdNhbXA2865044574ENgXOVytxOec+Cfkd+gx/P0hNi+R6leca4CPn3D7n3H7gI2B0jOK6BXgjSu9dIefcIiq+B+d64FXnfQa0MrNOROl61dmEHqGy2hJ0wbcdOOCcKyixPRo6OOdyAQJ/tq/k+MmU/mV6KvDfrWfNrHEtx9XEzJaZ2WfFZSDq0PUys4vwo67NIZujdb3K+30p85jA9ShuYxHJa2syrlB34kd5xcr6mdZmXN8K/Hxmm1nxTYh14noFSlM9gY9DNtfU9YpEebFH5XrFtB+6mc0HOpax6xHn3DuRnKKMba6C7dWOK9JzBM7TCRiIn8Nf7MfATnzSehH4EfBELcbVzTm3w8zOBT42s5XAN2UcF6vr9RdginOuKLD5tK9XWW9RxrZI21hU63eqEhGf28xuA9KAK0I2l/qZOuc2l/X6GohrDvCGc+6Emd2D/9/Nv0T42pqMq9hkYLZzrjBkW01dr0jU6O9XTBO6c+6qap6ivLYEe/D/lWkYGGWV1a7gtOIys11m1sk5lxtIQLsrONXNwN+ccydDzp0beHjCzF4B/qs24wqUNHDObTGzhUAK8FdifL3M7GzgPeDRwH9Fi8992terDFVpY5Fj4W0sInltTcaFmV2F/0fyCufcieLt5fxMo5GgKo3LObc35OkfOdU6OwcYUeK1C6MQU0RxhZgM/Efohhq8XpEoL/aoXK/6XnIpsy2B858yfIKvX4NvSxDJiD8SoW0OKjtvqdpdIKkV161vAMr8NLwm4jKz1sUlCzNrC1wCrIn19Qr87P6Gry2+XWJfNK9XddpYpAOTzc+C6Qn0Br6oRixVisvMUoA/AOOdc7tDtpf5M63FuDqFPB0PrA08/hC4OhBfa+Bqwv+nWqNxBWK7AP8B45KQbTV5vSKRDvxbYLbLUOBgYNASnetVU5/2VvcLuBH/r9YJYBfwYWB7Z2BuyHHXAhvw/8I+ErL9XPxfuE3A20DjKMWVBCwANgb+bBPYnga8FHJcD+BroEGJ138MrMQnpteA5rUVFzA88N7ZgT/vrAvXC7gNOAlkhXwl18T1Kuv3BV/CGR943CTw/W8KXI9zQ177SOB164ExUf59ryyu+YG/B8XXJ72yn2ktxfVLYHXg/T8B+oS89o7AddwE3F6bcQWePwY8XeJ1NX293sDP0jqJz193AvcA9wT2G37BoM2B908LeW21r5du/RcRiRP1veQiIiIBSugiInFCCV1EJE4ooYuIxAkldBGROKGELvWOmSWFdMvbaWZfBx4fMLOozyk2sxFm9m4VX7PQyugSar474QvRi07kFCV0qXecc3udc8nOuWR8Z8hnA4+TgaKKXx1cyFwk7iihS7xJMLM/mu+tPs/MmkJwxPwLM/sUuN/M2pnZX81saeDrksBxV4SM/jPNrEXgvM0DzafWmdnrgTtXMbORgeNWmu+FXap5mJndbr4n96f4OxNFaoQSusSb3sB051x/4ADwrZB9rZxzVzjn/i/wHH5kf2HgmJcCx/wX8B+BEf9lwLHA9hTgAXxf9HOBS8ysCb5f+yTn3EB8b6TvhQYTuDX+cXwiHxV4vUiNUEKXePOlcy4r8Hg5vgVDsVkhj68CXjCzLHx/jbMDo/HFwG/M7Pv4fwCKWwp/4ZzLcb4LZFbgvBcE3m9D4Jg/4xc4CHUxsNA5l+d87+5ZiNQQ1RIl3pwIeVwINA15fiTkcQNgmHPuGOGeNrP38L1CPgt0OCzrvA0pu+VpWdRfQ2qFRuhyppoHTC1+YoG1Ss2sl3NupXPuV8AyoE8F51gH9DCz8wLPvw18WuKYz4ERgZk5jYCbovUNiJSkhC5nqu8DaeZX2lmD74gH8ICZrTKzbHz9/P3yTuCcOw7cDrxtfqGQIvysm9BjcvFd/5bgOyZmRPsbESmmbosiInFCI3QRkTihhC4iEieU0EVE4oQSuohInFBCFxGJE0roIiJxQgldRCRO/H//Kr5iKOSzzwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "show_threshold_vs_recall_precision_curve(thresholds, recalls, precisions)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {},
   "outputs": [],
   "source": [
    "fpr, tpr, thresholds = roc_curve(y_valid, y_valid_predict)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "show_roc_curve(fpr, tpr)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8189935064935064"
      ]
     },
     "execution_count": 79,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "roc_auc = auc(fpr, tpr)\n",
    "roc_auc"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Try again `Stochastic Gradient Descent (SGD) Classifier`, but using only the important attributes and find the optimal hyperparameters using a `Random Search`:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 30 candidates, totalling 150 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=4)]: Using backend LokyBackend with 4 concurrent workers.\n",
      "[Parallel(n_jobs=4)]: Done  92 tasks      | elapsed:    0.9s\n",
      "[Parallel(n_jobs=4)]: Done 150 out of 150 | elapsed:    1.5s finished\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "RandomizedSearchCV(cv=5, error_score=nan,\n",
       "                   estimator=SGDClassifier(alpha=0.0001, average=False,\n",
       "                                           class_weight=None,\n",
       "                                           early_stopping=False, epsilon=0.1,\n",
       "                                           eta0=0.0, fit_intercept=True,\n",
       "                                           l1_ratio=0.15,\n",
       "                                           learning_rate='optimal',\n",
       "                                           loss='hinge', max_iter=1000,\n",
       "                                           n_iter_no_change=5, n_jobs=None,\n",
       "                                           penalty='l2', power_t=0.5,\n",
       "                                           random_state=42, shuffle=True,\n",
       "                                           tol=0.001, validation_fraction=0.1,\n",
       "                                           verbose=0, warm_start=False),\n",
       "                   iid='deprecated', n_iter=30, n_jobs=4,\n",
       "                   param_distributions={'alpha': <scipy.stats._distn_infrastructure.rv_frozen object at 0x0000029262F68E08>,\n",
       "                                        'loss': ['hinge', 'log', 'huber']},\n",
       "                   pre_dispatch='2*n_jobs', random_state=42, refit=True,\n",
       "                   return_train_score=False, scoring='f1', verbose=2)"
      ]
     },
     "execution_count": 80,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Note: gamma is ignored when kernel is \"linear\"\n",
    "param_distribs = {\n",
    "        'loss': ['hinge', 'log', 'huber'],\n",
    "        'alpha': reciprocal(100, 100000),\n",
    "    }\n",
    "\n",
    "\n",
    "sgd_clf = SGDClassifier(random_state=42)\n",
    "\n",
    "rnd_search = RandomizedSearchCV(sgd_clf, param_distributions=param_distribs,\n",
    "                                n_iter=30, cv=5, scoring='f1',\n",
    "                                verbose=2, n_jobs=4, random_state=42)\n",
    "rnd_search.fit(X_train_important, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'alpha': 3752.0558551242802, 'loss': 'hinge'}"
      ]
     },
     "execution_count": 81,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rnd_search.best_params_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_valid_predict = rnd_search.predict(X_valid_important)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[41, 15],\n",
       "       [ 1, 43]], dtype=int64)"
      ]
     },
     "execution_count": 83,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "confusion_matrix(y_valid, y_valid_predict)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Precision score:\t 0.7413793103448276\n",
      "Recall score:\t\t 0.9772727272727273\n",
      "F1 score:\t\t 0.8431372549019609\n"
     ]
    }
   ],
   "source": [
    "performance_scores(y_valid, y_valid_predict)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [],
   "source": [
    "precisions, recalls, thresholds = precision_recall_curve(y_valid, y_valid_predict)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "show_recall_vs_precision_curve(recalls, precisions)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "show_threshold_vs_recall_precision_curve(thresholds, recalls, precisions)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "metadata": {},
   "outputs": [],
   "source": [
    "fpr, tpr, thresholds = roc_curve(y_valid, y_valid_predict)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "show_roc_curve(fpr, tpr)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8547077922077922"
      ]
     },
     "execution_count": 90,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "roc_auc = auc(fpr, tpr)\n",
    "roc_auc"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<table>\n",
    "    <thead>\n",
    "        <tr>\n",
    "            <th colspan=\"5\">Performance of models</th>\n",
    "        </tr>\n",
    "    </thead>\n",
    "    <tbody>\n",
    "        <tr>\n",
    "            <td><b>Model</b></td>\n",
    "            <td><b>Precision</b></td>\n",
    "            <td><b>Recall</b></td>\n",
    "            <td><b>F1</b></td>\n",
    "            <td><b>ROC_AUC</b></td>\n",
    "        </tr>\n",
    "         <tr>\n",
    "            <td><b>SGD Classifier</b></td>\n",
    "            <td>0.5</td>\n",
    "            <td>1.0</td>\n",
    "            <td>0.6667</td>\n",
    "            <td>0.6071</td> \n",
    "        </tr>\n",
    "        <tr>\n",
    "            <td><b>Random Forest Classifier</b></td>\n",
    "            <td>0.65</td>\n",
    "            <td>0.2954</td>\n",
    "            <td>0.4062</td>\n",
    "            <td>0.5040</td>\n",
    "        </tr>\n",
    "         <tr>\n",
    "            <td><b>Linear SVC</b></td>\n",
    "            <td>0.4</td>\n",
    "            <td>0.6818</td>\n",
    "            <td>0.5042</td>\n",
    "            <td>0.7029</td> \n",
    "        </tr>\n",
    "         <tr>\n",
    "            <td><b>Nonlinear SVC</b></td>\n",
    "            <td>0.6341</td>\n",
    "            <td>0.5909</td>\n",
    "            <td>0.6117</td>\n",
    "            <td>0.6615</td> \n",
    "        </tr>\n",
    "        <tr>\n",
    "            <td><b>Optimal SVC</b></td>\n",
    "            <td>0.6667</td>\n",
    "            <td>0.9090</td>\n",
    "            <td>0.7692</td>\n",
    "            <td>0.7759</td> \n",
    "        </tr>\n",
    "        <tr>\n",
    "            <td><b>Optimal SVC selected attribs</b></td>\n",
    "            <td>0.8648</td>\n",
    "            <td>0.7272</td>\n",
    "            <td>0.7901</td>\n",
    "            <td>0.8189</td> \n",
    "        </tr>\n",
    "        <tr>\n",
    "            <td><b>Optimal SGD Classifier selected attribs</b></td>\n",
    "            <td>0.7413</td>\n",
    "            <td>0.9772</td>\n",
    "            <td><b>0.8431</b></td>\n",
    "            <td><b>0.8547</b></td> \n",
    "        </tr>\n",
    "    </tbody>\n",
    "</table>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    " **Best model** is the  optimized SGD Classifier with top important features."
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.6"
  },
  "nav_menu": {
   "height": "279px",
   "width": "309px"
  },
  "toc": {
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": false,
   "toc_cell": false,
   "toc_position": {},
   "toc_section_display": "block",
   "toc_window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}