Final_Project_Sam_Carey_MiniBooNE_Classifier_RF_XGBoost_NN.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "a42dc1cc",
   "metadata": {},
   "outputs": [],
   "source": [
    "%reload_ext autoreload\n",
    "%autoreload 2\n",
    "%matplotlib inline\n",
    "import matplotlib\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import seaborn as sns\n",
    "from sklearn.model_selection import train_test_split\n",
    "import feature_engineering as fe\n",
    "import tuning as tm\n",
    "import visualize as vz"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "e228992c",
   "metadata": {},
   "outputs": [],
   "source": [
    "from pathlib import Path"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "d9337c22",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Requirement already satisfied: xgboost in c:\\users\\samph\\anaconda3\\lib\\site-packages (1.7.2)\n",
      "Requirement already satisfied: scipy in c:\\users\\samph\\anaconda3\\lib\\site-packages (from xgboost) (1.7.1)\n",
      "Requirement already satisfied: numpy in c:\\users\\samph\\anaconda3\\lib\\site-packages (from xgboost) (1.20.3)\n"
     ]
    }
   ],
   "source": [
    "#install XGBoost\n",
    "import sys\n",
    "!{sys.executable} -m pip install xgboost"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "807958c3",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.base import BaseEstimator, TransformerMixin\n",
    "\n",
    "# Define Dictifier class to turn df into dictionary as part of pipeline\n",
    "class Dictifier(BaseEstimator, TransformerMixin):       \n",
    "    def fit(self, X, y=None):\n",
    "        return self\n",
    "\n",
    "    def transform(self, X):\n",
    "        if type(X) == pd.core.frame.DataFrame:\n",
    "            return X.to_dict(\"records\")\n",
    "        else:\n",
    "            return pd.DataFrame(X).to_dict(\"records\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "060cb2f2",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.pipeline import Pipeline\n",
    "from sklearn.feature_extraction import DictVectorizer\n",
    "import xgboost as xgb\n",
    "from sklearn.preprocessing import MinMaxScaler\n",
    "pipeline = Pipeline([\n",
    "    #(\"featureunion\", numeric_categorical_union),\n",
    "    (\"dictifier\", Dictifier()),\n",
    "    ('ohe_onestep', DictVectorizer(sparse=False)),\n",
    "    (\"clf\", xgb.XGBClassifier(max_depth=3))\n",
    "])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9d2392ee",
   "metadata": {},
   "source": [
    "pre-process dataset for training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "6aa51370",
   "metadata": {},
   "outputs": [],
   "source": [
    " #data from UCI: https://archive.ics.uci.edu/ml/datasets/MiniBooNE+particle+identification#\n",
    "df = pd.read_csv(\"MiniBooNE.csv\") \n",
    "df.rename(columns = {'51':'signal'}, inplace = True) #1: electron neutrino, 0: muon neutrino (background)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "2403eec0",
   "metadata": {
    "scrolled": true
   },
   "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>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>10</th>\n",
       "      <th>...</th>\n",
       "      <th>42</th>\n",
       "      <th>43</th>\n",
       "      <th>44</th>\n",
       "      <th>45</th>\n",
       "      <th>46</th>\n",
       "      <th>47</th>\n",
       "      <th>48</th>\n",
       "      <th>49</th>\n",
       "      <th>50</th>\n",
       "      <th>signal</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2.59413</td>\n",
       "      <td>0.468803</td>\n",
       "      <td>20.6916</td>\n",
       "      <td>0.322648</td>\n",
       "      <td>0.009682</td>\n",
       "      <td>0.374393</td>\n",
       "      <td>0.803479</td>\n",
       "      <td>0.896592</td>\n",
       "      <td>3.59665</td>\n",
       "      <td>0.249282</td>\n",
       "      <td>...</td>\n",
       "      <td>-31.3730</td>\n",
       "      <td>0.442259</td>\n",
       "      <td>5.86453</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.090519</td>\n",
       "      <td>0.176909</td>\n",
       "      <td>0.457585</td>\n",
       "      <td>0.071769</td>\n",
       "      <td>0.245996</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>3.86388</td>\n",
       "      <td>0.645781</td>\n",
       "      <td>18.1375</td>\n",
       "      <td>0.233529</td>\n",
       "      <td>0.030733</td>\n",
       "      <td>0.361239</td>\n",
       "      <td>1.069740</td>\n",
       "      <td>0.878714</td>\n",
       "      <td>3.59243</td>\n",
       "      <td>0.200793</td>\n",
       "      <td>...</td>\n",
       "      <td>45.9597</td>\n",
       "      <td>-0.478507</td>\n",
       "      <td>6.11126</td>\n",
       "      <td>0.001182</td>\n",
       "      <td>0.091800</td>\n",
       "      <td>-0.465572</td>\n",
       "      <td>0.935523</td>\n",
       "      <td>0.333613</td>\n",
       "      <td>0.230621</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3.38584</td>\n",
       "      <td>1.197140</td>\n",
       "      <td>36.0807</td>\n",
       "      <td>0.200866</td>\n",
       "      <td>0.017341</td>\n",
       "      <td>0.260841</td>\n",
       "      <td>1.108950</td>\n",
       "      <td>0.884405</td>\n",
       "      <td>3.43159</td>\n",
       "      <td>0.177167</td>\n",
       "      <td>...</td>\n",
       "      <td>-11.5608</td>\n",
       "      <td>-0.297008</td>\n",
       "      <td>8.27204</td>\n",
       "      <td>0.003854</td>\n",
       "      <td>0.141721</td>\n",
       "      <td>-0.210559</td>\n",
       "      <td>1.013450</td>\n",
       "      <td>0.255512</td>\n",
       "      <td>0.180901</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4.28524</td>\n",
       "      <td>0.510155</td>\n",
       "      <td>674.2010</td>\n",
       "      <td>0.281923</td>\n",
       "      <td>0.009174</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.998822</td>\n",
       "      <td>0.823390</td>\n",
       "      <td>3.16382</td>\n",
       "      <td>0.171678</td>\n",
       "      <td>...</td>\n",
       "      <td>-18.4586</td>\n",
       "      <td>0.453886</td>\n",
       "      <td>2.48112</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.180938</td>\n",
       "      <td>0.407968</td>\n",
       "      <td>4.341270</td>\n",
       "      <td>0.473081</td>\n",
       "      <td>0.258990</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5.93662</td>\n",
       "      <td>0.832993</td>\n",
       "      <td>59.8796</td>\n",
       "      <td>0.232853</td>\n",
       "      <td>0.025066</td>\n",
       "      <td>0.233556</td>\n",
       "      <td>1.370040</td>\n",
       "      <td>0.787424</td>\n",
       "      <td>3.66546</td>\n",
       "      <td>0.174862</td>\n",
       "      <td>...</td>\n",
       "      <td>42.9600</td>\n",
       "      <td>-0.975752</td>\n",
       "      <td>2.66109</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.170836</td>\n",
       "      <td>-0.814403</td>\n",
       "      <td>4.679490</td>\n",
       "      <td>1.924990</td>\n",
       "      <td>0.253893</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 51 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "         1         2         3         4         5         6         7  \\\n",
       "0  2.59413  0.468803   20.6916  0.322648  0.009682  0.374393  0.803479   \n",
       "1  3.86388  0.645781   18.1375  0.233529  0.030733  0.361239  1.069740   \n",
       "2  3.38584  1.197140   36.0807  0.200866  0.017341  0.260841  1.108950   \n",
       "3  4.28524  0.510155  674.2010  0.281923  0.009174  0.000000  0.998822   \n",
       "4  5.93662  0.832993   59.8796  0.232853  0.025066  0.233556  1.370040   \n",
       "\n",
       "          8        9        10  ...       42        43       44        45  \\\n",
       "0  0.896592  3.59665  0.249282  ... -31.3730  0.442259  5.86453  0.000000   \n",
       "1  0.878714  3.59243  0.200793  ...  45.9597 -0.478507  6.11126  0.001182   \n",
       "2  0.884405  3.43159  0.177167  ... -11.5608 -0.297008  8.27204  0.003854   \n",
       "3  0.823390  3.16382  0.171678  ... -18.4586  0.453886  2.48112  0.000000   \n",
       "4  0.787424  3.66546  0.174862  ...  42.9600 -0.975752  2.66109  0.000000   \n",
       "\n",
       "         46        47        48        49        50  signal  \n",
       "0  0.090519  0.176909  0.457585  0.071769  0.245996       1  \n",
       "1  0.091800 -0.465572  0.935523  0.333613  0.230621       1  \n",
       "2  0.141721 -0.210559  1.013450  0.255512  0.180901       1  \n",
       "3  0.180938  0.407968  4.341270  0.473081  0.258990       1  \n",
       "4  0.170836 -0.814403  4.679490  1.924990  0.253893       1  \n",
       "\n",
       "[5 rows x 51 columns]"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "372c0f76",
   "metadata": {
    "scrolled": true
   },
   "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>count</th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "      <th>min</th>\n",
       "      <th>10%</th>\n",
       "      <th>25%</th>\n",
       "      <th>50%</th>\n",
       "      <th>75%</th>\n",
       "      <th>90%</th>\n",
       "      <th>max</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>130,064.000</td>\n",
       "      <td>1.252</td>\n",
       "      <td>60.120</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>3.358</td>\n",
       "      <td>4.013</td>\n",
       "      <td>4.788</td>\n",
       "      <td>5.628</td>\n",
       "      <td>6.386</td>\n",
       "      <td>17.057</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>130,064.000</td>\n",
       "      <td>-2.108</td>\n",
       "      <td>59.913</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>0.489</td>\n",
       "      <td>0.876</td>\n",
       "      <td>1.382</td>\n",
       "      <td>1.948</td>\n",
       "      <td>2.511</td>\n",
       "      <td>8.803</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>130,064.000</td>\n",
       "      <td>123.795</td>\n",
       "      <td>196.011</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>16.770</td>\n",
       "      <td>49.295</td>\n",
       "      <td>106.136</td>\n",
       "      <td>138.849</td>\n",
       "      <td>196.829</td>\n",
       "      <td>4,747.670</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>130,064.000</td>\n",
       "      <td>-3.313</td>\n",
       "      <td>59.835</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>0.219</td>\n",
       "      <td>0.242</td>\n",
       "      <td>0.272</td>\n",
       "      <td>0.315</td>\n",
       "      <td>0.362</td>\n",
       "      <td>0.737</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>130,064.000</td>\n",
       "      <td>-3.585</td>\n",
       "      <td>59.818</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.005</td>\n",
       "      <td>0.013</td>\n",
       "      <td>0.025</td>\n",
       "      <td>0.179</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>130,064.000</td>\n",
       "      <td>-3.434</td>\n",
       "      <td>59.827</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.085</td>\n",
       "      <td>0.141</td>\n",
       "      <td>0.213</td>\n",
       "      <td>0.348</td>\n",
       "      <td>0.704</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>130,064.000</td>\n",
       "      <td>-2.628</td>\n",
       "      <td>59.877</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>0.569</td>\n",
       "      <td>0.715</td>\n",
       "      <td>0.923</td>\n",
       "      <td>1.180</td>\n",
       "      <td>1.424</td>\n",
       "      <td>6.241</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>130,064.000</td>\n",
       "      <td>-2.774</td>\n",
       "      <td>59.867</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>0.735</td>\n",
       "      <td>0.783</td>\n",
       "      <td>0.834</td>\n",
       "      <td>0.874</td>\n",
       "      <td>0.901</td>\n",
       "      <td>0.990</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>130,064.000</td>\n",
       "      <td>-0.156</td>\n",
       "      <td>60.025</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>3.161</td>\n",
       "      <td>3.285</td>\n",
       "      <td>3.427</td>\n",
       "      <td>3.585</td>\n",
       "      <td>3.755</td>\n",
       "      <td>7.170</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>130,064.000</td>\n",
       "      <td>-3.421</td>\n",
       "      <td>59.828</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>0.120</td>\n",
       "      <td>0.143</td>\n",
       "      <td>0.171</td>\n",
       "      <td>0.200</td>\n",
       "      <td>0.230</td>\n",
       "      <td>0.525</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>130,064.000</td>\n",
       "      <td>0.710</td>\n",
       "      <td>60.077</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>4.013</td>\n",
       "      <td>4.152</td>\n",
       "      <td>4.289</td>\n",
       "      <td>4.435</td>\n",
       "      <td>4.603</td>\n",
       "      <td>9.559</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>130,064.000</td>\n",
       "      <td>161.447</td>\n",
       "      <td>134.250</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>4.720</td>\n",
       "      <td>73.392</td>\n",
       "      <td>155.760</td>\n",
       "      <td>244.561</td>\n",
       "      <td>326.050</td>\n",
       "      <td>537.262</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>130,064.000</td>\n",
       "      <td>-3.145</td>\n",
       "      <td>59.847</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>-0.417</td>\n",
       "      <td>0.149</td>\n",
       "      <td>0.633</td>\n",
       "      <td>0.873</td>\n",
       "      <td>0.960</td>\n",
       "      <td>1.000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>130,064.000</td>\n",
       "      <td>-3.285</td>\n",
       "      <td>59.837</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>0.044</td>\n",
       "      <td>0.112</td>\n",
       "      <td>0.244</td>\n",
       "      <td>0.469</td>\n",
       "      <td>0.700</td>\n",
       "      <td>1.000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>130,064.000</td>\n",
       "      <td>-1.353</td>\n",
       "      <td>59.956</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>1.572</td>\n",
       "      <td>1.778</td>\n",
       "      <td>2.096</td>\n",
       "      <td>2.524</td>\n",
       "      <td>3.062</td>\n",
       "      <td>11.296</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>130,064.000</td>\n",
       "      <td>933.035</td>\n",
       "      <td>652.826</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>458.752</td>\n",
       "      <td>562.799</td>\n",
       "      <td>732.302</td>\n",
       "      <td>1,048.533</td>\n",
       "      <td>1,684.897</td>\n",
       "      <td>6,907.690</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>130,064.000</td>\n",
       "      <td>-3.141</td>\n",
       "      <td>59.846</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>0.157</td>\n",
       "      <td>0.214</td>\n",
       "      <td>0.315</td>\n",
       "      <td>0.508</td>\n",
       "      <td>0.943</td>\n",
       "      <td>3.902</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>130,064.000</td>\n",
       "      <td>23.311</td>\n",
       "      <td>68.403</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>5.959</td>\n",
       "      <td>9.894</td>\n",
       "      <td>18.268</td>\n",
       "      <td>33.131</td>\n",
       "      <td>55.719</td>\n",
       "      <td>1,097.130</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>130,064.000</td>\n",
       "      <td>-3.327</td>\n",
       "      <td>59.834</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>0.211</td>\n",
       "      <td>0.233</td>\n",
       "      <td>0.260</td>\n",
       "      <td>0.294</td>\n",
       "      <td>0.335</td>\n",
       "      <td>0.759</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>130,064.000</td>\n",
       "      <td>119.908</td>\n",
       "      <td>44,367.689</td>\n",
       "      <td>-17,256.700</td>\n",
       "      <td>0.193</td>\n",
       "      <td>0.467</td>\n",
       "      <td>0.785</td>\n",
       "      <td>1.077</td>\n",
       "      <td>1.356</td>\n",
       "      <td>16,000,900.000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>130,064.000</td>\n",
       "      <td>-3.695</td>\n",
       "      <td>59.841</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>-2.412</td>\n",
       "      <td>-0.858</td>\n",
       "      <td>0.216</td>\n",
       "      <td>0.946</td>\n",
       "      <td>1.558</td>\n",
       "      <td>22.404</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>130,064.000</td>\n",
       "      <td>-3.287</td>\n",
       "      <td>59.836</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>0.249</td>\n",
       "      <td>0.277</td>\n",
       "      <td>0.307</td>\n",
       "      <td>0.338</td>\n",
       "      <td>0.368</td>\n",
       "      <td>0.597</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>130,064.000</td>\n",
       "      <td>80.653</td>\n",
       "      <td>106.956</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>22.798</td>\n",
       "      <td>35.937</td>\n",
       "      <td>57.870</td>\n",
       "      <td>98.040</td>\n",
       "      <td>175.287</td>\n",
       "      <td>1,428.590</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>130,064.000</td>\n",
       "      <td>0.238</td>\n",
       "      <td>60.075</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>2.127</td>\n",
       "      <td>2.686</td>\n",
       "      <td>3.505</td>\n",
       "      <td>4.565</td>\n",
       "      <td>5.798</td>\n",
       "      <td>78.363</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>130,064.000</td>\n",
       "      <td>-3.156</td>\n",
       "      <td>59.844</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>0.358</td>\n",
       "      <td>0.414</td>\n",
       "      <td>0.459</td>\n",
       "      <td>0.483</td>\n",
       "      <td>0.494</td>\n",
       "      <td>0.500</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>130,064.000</td>\n",
       "      <td>-4.194</td>\n",
       "      <td>59.818</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>-2.773</td>\n",
       "      <td>-0.936</td>\n",
       "      <td>0.001</td>\n",
       "      <td>0.399</td>\n",
       "      <td>0.823</td>\n",
       "      <td>4.717</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>130,064.000</td>\n",
       "      <td>97.404</td>\n",
       "      <td>98.222</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>7.165</td>\n",
       "      <td>33.047</td>\n",
       "      <td>109.063</td>\n",
       "      <td>145.078</td>\n",
       "      <td>177.879</td>\n",
       "      <td>1,024.210</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>130,064.000</td>\n",
       "      <td>-2.034</td>\n",
       "      <td>59.912</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>1.244</td>\n",
       "      <td>1.392</td>\n",
       "      <td>1.559</td>\n",
       "      <td>1.716</td>\n",
       "      <td>1.859</td>\n",
       "      <td>3.647</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>130,064.000</td>\n",
       "      <td>-3.422</td>\n",
       "      <td>59.828</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>0.113</td>\n",
       "      <td>0.136</td>\n",
       "      <td>0.164</td>\n",
       "      <td>0.203</td>\n",
       "      <td>0.246</td>\n",
       "      <td>0.653</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30</th>\n",
       "      <td>130,064.000</td>\n",
       "      <td>-3.899</td>\n",
       "      <td>59.802</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>-1.098</td>\n",
       "      <td>-0.687</td>\n",
       "      <td>-0.252</td>\n",
       "      <td>0.091</td>\n",
       "      <td>0.344</td>\n",
       "      <td>5.870</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>31</th>\n",
       "      <td>130,064.000</td>\n",
       "      <td>3.952</td>\n",
       "      <td>60.315</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>4.995</td>\n",
       "      <td>5.895</td>\n",
       "      <td>7.208</td>\n",
       "      <td>8.824</td>\n",
       "      <td>10.532</td>\n",
       "      <td>23.933</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32</th>\n",
       "      <td>130,064.000</td>\n",
       "      <td>-2.583</td>\n",
       "      <td>59.880</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>0.505</td>\n",
       "      <td>0.749</td>\n",
       "      <td>1.032</td>\n",
       "      <td>1.294</td>\n",
       "      <td>1.485</td>\n",
       "      <td>4.127</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>33</th>\n",
       "      <td>130,064.000</td>\n",
       "      <td>-2.382</td>\n",
       "      <td>59.895</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>0.423</td>\n",
       "      <td>0.644</td>\n",
       "      <td>1.043</td>\n",
       "      <td>1.601</td>\n",
       "      <td>2.231</td>\n",
       "      <td>7.994</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>34</th>\n",
       "      <td>130,064.000</td>\n",
       "      <td>361.939</td>\n",
       "      <td>126.438</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>222.847</td>\n",
       "      <td>305.134</td>\n",
       "      <td>385.884</td>\n",
       "      <td>444.887</td>\n",
       "      <td>477.029</td>\n",
       "      <td>499.999</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>35</th>\n",
       "      <td>130,064.000</td>\n",
       "      <td>-3.522</td>\n",
       "      <td>59.822</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>0.022</td>\n",
       "      <td>0.040</td>\n",
       "      <td>0.067</td>\n",
       "      <td>0.101</td>\n",
       "      <td>0.130</td>\n",
       "      <td>0.403</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>36</th>\n",
       "      <td>130,064.000</td>\n",
       "      <td>-3.300</td>\n",
       "      <td>59.835</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>0.209</td>\n",
       "      <td>0.248</td>\n",
       "      <td>0.292</td>\n",
       "      <td>0.339</td>\n",
       "      <td>0.385</td>\n",
       "      <td>0.897</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>37</th>\n",
       "      <td>130,064.000</td>\n",
       "      <td>-3.205</td>\n",
       "      <td>59.847</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>-0.487</td>\n",
       "      <td>-0.202</td>\n",
       "      <td>0.202</td>\n",
       "      <td>0.778</td>\n",
       "      <td>1.509</td>\n",
       "      <td>8.192</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>38</th>\n",
       "      <td>130,064.000</td>\n",
       "      <td>-6.365</td>\n",
       "      <td>59.659</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>-4.143</td>\n",
       "      <td>-3.486</td>\n",
       "      <td>-2.734</td>\n",
       "      <td>-2.065</td>\n",
       "      <td>-1.523</td>\n",
       "      <td>0.670</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>39</th>\n",
       "      <td>130,064.000</td>\n",
       "      <td>-2.318</td>\n",
       "      <td>59.895</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>0.933</td>\n",
       "      <td>1.091</td>\n",
       "      <td>1.270</td>\n",
       "      <td>1.451</td>\n",
       "      <td>1.621</td>\n",
       "      <td>6.059</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>40</th>\n",
       "      <td>130,064.000</td>\n",
       "      <td>-3.503</td>\n",
       "      <td>59.823</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>0.050</td>\n",
       "      <td>0.065</td>\n",
       "      <td>0.084</td>\n",
       "      <td>0.107</td>\n",
       "      <td>0.141</td>\n",
       "      <td>0.430</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>41</th>\n",
       "      <td>130,064.000</td>\n",
       "      <td>141.180</td>\n",
       "      <td>75.499</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>103.003</td>\n",
       "      <td>122.902</td>\n",
       "      <td>145.414</td>\n",
       "      <td>167.088</td>\n",
       "      <td>185.858</td>\n",
       "      <td>331.925</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>42</th>\n",
       "      <td>130,064.000</td>\n",
       "      <td>-25.226</td>\n",
       "      <td>67.576</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>-57.146</td>\n",
       "      <td>-40.068</td>\n",
       "      <td>-21.592</td>\n",
       "      <td>-3.330</td>\n",
       "      <td>16.147</td>\n",
       "      <td>161.298</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>43</th>\n",
       "      <td>130,064.000</td>\n",
       "      <td>-2.864</td>\n",
       "      <td>59.874</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>-0.361</td>\n",
       "      <td>-0.028</td>\n",
       "      <td>0.459</td>\n",
       "      <td>1.150</td>\n",
       "      <td>2.072</td>\n",
       "      <td>43.651</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>44</th>\n",
       "      <td>130,064.000</td>\n",
       "      <td>2.393</td>\n",
       "      <td>60.349</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>2.480</td>\n",
       "      <td>3.413</td>\n",
       "      <td>4.762</td>\n",
       "      <td>6.857</td>\n",
       "      <td>10.650</td>\n",
       "      <td>85.823</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>45</th>\n",
       "      <td>130,064.000</td>\n",
       "      <td>-3.586</td>\n",
       "      <td>59.818</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.002</td>\n",
       "      <td>0.023</td>\n",
       "      <td>0.447</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>46</th>\n",
       "      <td>130,064.000</td>\n",
       "      <td>-3.416</td>\n",
       "      <td>59.828</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>0.086</td>\n",
       "      <td>0.122</td>\n",
       "      <td>0.169</td>\n",
       "      <td>0.225</td>\n",
       "      <td>0.286</td>\n",
       "      <td>0.606</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>47</th>\n",
       "      <td>130,064.000</td>\n",
       "      <td>-3.753</td>\n",
       "      <td>59.816</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>-0.958</td>\n",
       "      <td>-0.343</td>\n",
       "      <td>0.042</td>\n",
       "      <td>0.295</td>\n",
       "      <td>0.507</td>\n",
       "      <td>1.933</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>48</th>\n",
       "      <td>130,064.000</td>\n",
       "      <td>-0.745</td>\n",
       "      <td>60.019</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>0.907</td>\n",
       "      <td>1.420</td>\n",
       "      <td>2.364</td>\n",
       "      <td>3.778</td>\n",
       "      <td>5.503</td>\n",
       "      <td>21.768</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>49</th>\n",
       "      <td>130,064.000</td>\n",
       "      <td>-1.987</td>\n",
       "      <td>59.946</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>-0.036</td>\n",
       "      <td>0.274</td>\n",
       "      <td>1.124</td>\n",
       "      <td>2.281</td>\n",
       "      <td>3.772</td>\n",
       "      <td>25.423</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50</th>\n",
       "      <td>130,064.000</td>\n",
       "      <td>-3.365</td>\n",
       "      <td>59.831</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>0.163</td>\n",
       "      <td>0.197</td>\n",
       "      <td>0.231</td>\n",
       "      <td>0.263</td>\n",
       "      <td>0.294</td>\n",
       "      <td>0.625</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>signal</th>\n",
       "      <td>130,064.000</td>\n",
       "      <td>0.281</td>\n",
       "      <td>0.449</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              count     mean         std          min      10%      25%  \\\n",
       "1       130,064.000    1.252      60.120     -999.000    3.358    4.013   \n",
       "2       130,064.000   -2.108      59.913     -999.000    0.489    0.876   \n",
       "3       130,064.000  123.795     196.011     -999.000   16.770   49.295   \n",
       "4       130,064.000   -3.313      59.835     -999.000    0.219    0.242   \n",
       "5       130,064.000   -3.585      59.818     -999.000    0.000    0.000   \n",
       "6       130,064.000   -3.434      59.827     -999.000    0.000    0.085   \n",
       "7       130,064.000   -2.628      59.877     -999.000    0.569    0.715   \n",
       "8       130,064.000   -2.774      59.867     -999.000    0.735    0.783   \n",
       "9       130,064.000   -0.156      60.025     -999.000    3.161    3.285   \n",
       "10      130,064.000   -3.421      59.828     -999.000    0.120    0.143   \n",
       "11      130,064.000    0.710      60.077     -999.000    4.013    4.152   \n",
       "12      130,064.000  161.447     134.250     -999.000    4.720   73.392   \n",
       "13      130,064.000   -3.145      59.847     -999.000   -0.417    0.149   \n",
       "14      130,064.000   -3.285      59.837     -999.000    0.044    0.112   \n",
       "15      130,064.000   -1.353      59.956     -999.000    1.572    1.778   \n",
       "16      130,064.000  933.035     652.826     -999.000  458.752  562.799   \n",
       "17      130,064.000   -3.141      59.846     -999.000    0.157    0.214   \n",
       "18      130,064.000   23.311      68.403     -999.000    5.959    9.894   \n",
       "19      130,064.000   -3.327      59.834     -999.000    0.211    0.233   \n",
       "20      130,064.000  119.908  44,367.689  -17,256.700    0.193    0.467   \n",
       "21      130,064.000   -3.695      59.841     -999.000   -2.412   -0.858   \n",
       "22      130,064.000   -3.287      59.836     -999.000    0.249    0.277   \n",
       "23      130,064.000   80.653     106.956     -999.000   22.798   35.937   \n",
       "24      130,064.000    0.238      60.075     -999.000    2.127    2.686   \n",
       "25      130,064.000   -3.156      59.844     -999.000    0.358    0.414   \n",
       "26      130,064.000   -4.194      59.818     -999.000   -2.773   -0.936   \n",
       "27      130,064.000   97.404      98.222     -999.000    7.165   33.047   \n",
       "28      130,064.000   -2.034      59.912     -999.000    1.244    1.392   \n",
       "29      130,064.000   -3.422      59.828     -999.000    0.113    0.136   \n",
       "30      130,064.000   -3.899      59.802     -999.000   -1.098   -0.687   \n",
       "31      130,064.000    3.952      60.315     -999.000    4.995    5.895   \n",
       "32      130,064.000   -2.583      59.880     -999.000    0.505    0.749   \n",
       "33      130,064.000   -2.382      59.895     -999.000    0.423    0.644   \n",
       "34      130,064.000  361.939     126.438     -999.000  222.847  305.134   \n",
       "35      130,064.000   -3.522      59.822     -999.000    0.022    0.040   \n",
       "36      130,064.000   -3.300      59.835     -999.000    0.209    0.248   \n",
       "37      130,064.000   -3.205      59.847     -999.000   -0.487   -0.202   \n",
       "38      130,064.000   -6.365      59.659     -999.000   -4.143   -3.486   \n",
       "39      130,064.000   -2.318      59.895     -999.000    0.933    1.091   \n",
       "40      130,064.000   -3.503      59.823     -999.000    0.050    0.065   \n",
       "41      130,064.000  141.180      75.499     -999.000  103.003  122.902   \n",
       "42      130,064.000  -25.226      67.576     -999.000  -57.146  -40.068   \n",
       "43      130,064.000   -2.864      59.874     -999.000   -0.361   -0.028   \n",
       "44      130,064.000    2.393      60.349     -999.000    2.480    3.413   \n",
       "45      130,064.000   -3.586      59.818     -999.000    0.000    0.000   \n",
       "46      130,064.000   -3.416      59.828     -999.000    0.086    0.122   \n",
       "47      130,064.000   -3.753      59.816     -999.000   -0.958   -0.343   \n",
       "48      130,064.000   -0.745      60.019     -999.000    0.907    1.420   \n",
       "49      130,064.000   -1.987      59.946     -999.000   -0.036    0.274   \n",
       "50      130,064.000   -3.365      59.831     -999.000    0.163    0.197   \n",
       "signal  130,064.000    0.281       0.449        0.000    0.000    0.000   \n",
       "\n",
       "            50%        75%        90%             max  \n",
       "1         4.788      5.628      6.386          17.057  \n",
       "2         1.382      1.948      2.511           8.803  \n",
       "3       106.136    138.849    196.829       4,747.670  \n",
       "4         0.272      0.315      0.362           0.737  \n",
       "5         0.005      0.013      0.025           0.179  \n",
       "6         0.141      0.213      0.348           0.704  \n",
       "7         0.923      1.180      1.424           6.241  \n",
       "8         0.834      0.874      0.901           0.990  \n",
       "9         3.427      3.585      3.755           7.170  \n",
       "10        0.171      0.200      0.230           0.525  \n",
       "11        4.289      4.435      4.603           9.559  \n",
       "12      155.760    244.561    326.050         537.262  \n",
       "13        0.633      0.873      0.960           1.000  \n",
       "14        0.244      0.469      0.700           1.000  \n",
       "15        2.096      2.524      3.062          11.296  \n",
       "16      732.302  1,048.533  1,684.897       6,907.690  \n",
       "17        0.315      0.508      0.943           3.902  \n",
       "18       18.268     33.131     55.719       1,097.130  \n",
       "19        0.260      0.294      0.335           0.759  \n",
       "20        0.785      1.077      1.356  16,000,900.000  \n",
       "21        0.216      0.946      1.558          22.404  \n",
       "22        0.307      0.338      0.368           0.597  \n",
       "23       57.870     98.040    175.287       1,428.590  \n",
       "24        3.505      4.565      5.798          78.363  \n",
       "25        0.459      0.483      0.494           0.500  \n",
       "26        0.001      0.399      0.823           4.717  \n",
       "27      109.063    145.078    177.879       1,024.210  \n",
       "28        1.559      1.716      1.859           3.647  \n",
       "29        0.164      0.203      0.246           0.653  \n",
       "30       -0.252      0.091      0.344           5.870  \n",
       "31        7.208      8.824     10.532          23.933  \n",
       "32        1.032      1.294      1.485           4.127  \n",
       "33        1.043      1.601      2.231           7.994  \n",
       "34      385.884    444.887    477.029         499.999  \n",
       "35        0.067      0.101      0.130           0.403  \n",
       "36        0.292      0.339      0.385           0.897  \n",
       "37        0.202      0.778      1.509           8.192  \n",
       "38       -2.734     -2.065     -1.523           0.670  \n",
       "39        1.270      1.451      1.621           6.059  \n",
       "40        0.084      0.107      0.141           0.430  \n",
       "41      145.414    167.088    185.858         331.925  \n",
       "42      -21.592     -3.330     16.147         161.298  \n",
       "43        0.459      1.150      2.072          43.651  \n",
       "44        4.762      6.857     10.650          85.823  \n",
       "45        0.000      0.002      0.023           0.447  \n",
       "46        0.169      0.225      0.286           0.606  \n",
       "47        0.042      0.295      0.507           1.933  \n",
       "48        2.364      3.778      5.503          21.768  \n",
       "49        1.124      2.281      3.772          25.423  \n",
       "50        0.231      0.263      0.294           0.625  \n",
       "signal    0.000      1.000      1.000           1.000  "
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "vz.pretty_describe(df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "5fb78262",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training set size: 104,051\n",
      "Test set size: 26,013\n"
     ]
    }
   ],
   "source": [
    "# Train/Test split\n",
    "train, test = fe.train_test_stratified_split(df=df,\n",
    "                                        target='signal',\n",
    "                                        test_size=0.2,\n",
    "                                        random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "b1397e96",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Class ratio: 0.28062\n"
     ]
    }
   ],
   "source": [
    "\n",
    "train_X, train_y = fe.x_y_split(df=train,\n",
    "                          y_col='signal')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "e5c1c666",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Class ratio: 0.28063\n"
     ]
    }
   ],
   "source": [
    "test_X, test_y = fe.x_y_split(df=test,\n",
    "                          y_col='signal')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0992c5d2",
   "metadata": {},
   "source": [
    "# PCA"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "506f56a9",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.decomposition import PCA"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "3d701083",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "5.00% among 104,051 sample points are identified and removed as outliers\n"
     ]
    }
   ],
   "source": [
    "#remove outliers using k-NN\n",
    "train_X_removed_outliers, train_y_removed_outliers = fe.remove_outliers_knn(x=train_X,\n",
    "                                                                      y=train_y,\n",
    "                                                                      contamination=0.05)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "c4cfb99c",
   "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>count</th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "      <th>min</th>\n",
       "      <th>10%</th>\n",
       "      <th>25%</th>\n",
       "      <th>50%</th>\n",
       "      <th>75%</th>\n",
       "      <th>90%</th>\n",
       "      <th>max</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>98,848.000</td>\n",
       "      <td>1.053</td>\n",
       "      <td>61.972</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>3.409</td>\n",
       "      <td>4.062</td>\n",
       "      <td>4.816</td>\n",
       "      <td>5.646</td>\n",
       "      <td>6.392</td>\n",
       "      <td>17.057</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>98,848.000</td>\n",
       "      <td>-2.283</td>\n",
       "      <td>61.760</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>0.576</td>\n",
       "      <td>0.947</td>\n",
       "      <td>1.427</td>\n",
       "      <td>1.981</td>\n",
       "      <td>2.538</td>\n",
       "      <td>8.606</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>98,848.000</td>\n",
       "      <td>104.047</td>\n",
       "      <td>120.060</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>16.222</td>\n",
       "      <td>47.204</td>\n",
       "      <td>103.797</td>\n",
       "      <td>135.201</td>\n",
       "      <td>170.475</td>\n",
       "      <td>1,587.080</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>98,848.000</td>\n",
       "      <td>-3.538</td>\n",
       "      <td>61.677</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>0.219</td>\n",
       "      <td>0.242</td>\n",
       "      <td>0.273</td>\n",
       "      <td>0.316</td>\n",
       "      <td>0.362</td>\n",
       "      <td>0.713</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>98,848.000</td>\n",
       "      <td>-3.812</td>\n",
       "      <td>61.660</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.005</td>\n",
       "      <td>0.012</td>\n",
       "      <td>0.023</td>\n",
       "      <td>0.179</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>98,848.000</td>\n",
       "      <td>-3.659</td>\n",
       "      <td>61.669</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.089</td>\n",
       "      <td>0.142</td>\n",
       "      <td>0.213</td>\n",
       "      <td>0.350</td>\n",
       "      <td>0.704</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>98,848.000</td>\n",
       "      <td>-2.860</td>\n",
       "      <td>61.720</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>0.564</td>\n",
       "      <td>0.708</td>\n",
       "      <td>0.915</td>\n",
       "      <td>1.174</td>\n",
       "      <td>1.423</td>\n",
       "      <td>5.979</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>98,848.000</td>\n",
       "      <td>-3.002</td>\n",
       "      <td>61.710</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>0.734</td>\n",
       "      <td>0.781</td>\n",
       "      <td>0.831</td>\n",
       "      <td>0.871</td>\n",
       "      <td>0.897</td>\n",
       "      <td>0.967</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>98,848.000</td>\n",
       "      <td>-0.377</td>\n",
       "      <td>61.873</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>3.171</td>\n",
       "      <td>3.290</td>\n",
       "      <td>3.430</td>\n",
       "      <td>3.587</td>\n",
       "      <td>3.755</td>\n",
       "      <td>7.170</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>98,848.000</td>\n",
       "      <td>-3.647</td>\n",
       "      <td>61.670</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>0.120</td>\n",
       "      <td>0.143</td>\n",
       "      <td>0.171</td>\n",
       "      <td>0.201</td>\n",
       "      <td>0.231</td>\n",
       "      <td>0.525</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>98,848.000</td>\n",
       "      <td>0.487</td>\n",
       "      <td>61.927</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>4.020</td>\n",
       "      <td>4.156</td>\n",
       "      <td>4.292</td>\n",
       "      <td>4.436</td>\n",
       "      <td>4.602</td>\n",
       "      <td>9.559</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>98,848.000</td>\n",
       "      <td>162.591</td>\n",
       "      <td>134.486</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>7.436</td>\n",
       "      <td>76.677</td>\n",
       "      <td>157.356</td>\n",
       "      <td>245.162</td>\n",
       "      <td>325.683</td>\n",
       "      <td>537.262</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>98,848.000</td>\n",
       "      <td>-3.381</td>\n",
       "      <td>61.689</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>-0.430</td>\n",
       "      <td>0.133</td>\n",
       "      <td>0.619</td>\n",
       "      <td>0.865</td>\n",
       "      <td>0.956</td>\n",
       "      <td>1.000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>98,848.000</td>\n",
       "      <td>-3.516</td>\n",
       "      <td>61.679</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>0.043</td>\n",
       "      <td>0.109</td>\n",
       "      <td>0.237</td>\n",
       "      <td>0.460</td>\n",
       "      <td>0.698</td>\n",
       "      <td>1.000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>98,848.000</td>\n",
       "      <td>-1.651</td>\n",
       "      <td>61.796</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>1.564</td>\n",
       "      <td>1.763</td>\n",
       "      <td>2.067</td>\n",
       "      <td>2.459</td>\n",
       "      <td>2.897</td>\n",
       "      <td>7.712</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>98,848.000</td>\n",
       "      <td>857.645</td>\n",
       "      <td>520.960</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>455.099</td>\n",
       "      <td>555.354</td>\n",
       "      <td>714.846</td>\n",
       "      <td>982.721</td>\n",
       "      <td>1,454.210</td>\n",
       "      <td>5,198.120</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>98,848.000</td>\n",
       "      <td>-3.404</td>\n",
       "      <td>61.686</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>0.155</td>\n",
       "      <td>0.211</td>\n",
       "      <td>0.305</td>\n",
       "      <td>0.473</td>\n",
       "      <td>0.820</td>\n",
       "      <td>3.362</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>98,848.000</td>\n",
       "      <td>21.088</td>\n",
       "      <td>67.289</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>5.869</td>\n",
       "      <td>9.714</td>\n",
       "      <td>17.879</td>\n",
       "      <td>31.922</td>\n",
       "      <td>52.311</td>\n",
       "      <td>288.930</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>98,848.000</td>\n",
       "      <td>-3.552</td>\n",
       "      <td>61.676</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>0.213</td>\n",
       "      <td>0.235</td>\n",
       "      <td>0.261</td>\n",
       "      <td>0.294</td>\n",
       "      <td>0.335</td>\n",
       "      <td>0.759</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>98,848.000</td>\n",
       "      <td>-3.030</td>\n",
       "      <td>61.710</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>0.196</td>\n",
       "      <td>0.468</td>\n",
       "      <td>0.788</td>\n",
       "      <td>1.085</td>\n",
       "      <td>1.370</td>\n",
       "      <td>7.537</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>98,848.000</td>\n",
       "      <td>-3.906</td>\n",
       "      <td>61.681</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>-2.369</td>\n",
       "      <td>-0.824</td>\n",
       "      <td>0.234</td>\n",
       "      <td>0.955</td>\n",
       "      <td>1.553</td>\n",
       "      <td>22.404</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>98,848.000</td>\n",
       "      <td>-3.514</td>\n",
       "      <td>61.678</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>0.248</td>\n",
       "      <td>0.276</td>\n",
       "      <td>0.306</td>\n",
       "      <td>0.337</td>\n",
       "      <td>0.366</td>\n",
       "      <td>0.597</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>98,848.000</td>\n",
       "      <td>70.738</td>\n",
       "      <td>91.727</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>22.582</td>\n",
       "      <td>35.302</td>\n",
       "      <td>55.898</td>\n",
       "      <td>90.547</td>\n",
       "      <td>148.906</td>\n",
       "      <td>638.484</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>98,848.000</td>\n",
       "      <td>0.049</td>\n",
       "      <td>61.925</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>2.138</td>\n",
       "      <td>2.702</td>\n",
       "      <td>3.544</td>\n",
       "      <td>4.617</td>\n",
       "      <td>5.857</td>\n",
       "      <td>78.363</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>98,848.000</td>\n",
       "      <td>-3.382</td>\n",
       "      <td>61.686</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>0.357</td>\n",
       "      <td>0.413</td>\n",
       "      <td>0.458</td>\n",
       "      <td>0.483</td>\n",
       "      <td>0.493</td>\n",
       "      <td>0.500</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>98,848.000</td>\n",
       "      <td>-4.419</td>\n",
       "      <td>61.658</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>-2.781</td>\n",
       "      <td>-0.942</td>\n",
       "      <td>0.002</td>\n",
       "      <td>0.410</td>\n",
       "      <td>0.835</td>\n",
       "      <td>4.621</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>98,848.000</td>\n",
       "      <td>92.928</td>\n",
       "      <td>92.725</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>7.303</td>\n",
       "      <td>32.997</td>\n",
       "      <td>107.757</td>\n",
       "      <td>142.887</td>\n",
       "      <td>171.231</td>\n",
       "      <td>382.727</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>98,848.000</td>\n",
       "      <td>-2.258</td>\n",
       "      <td>61.756</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>1.251</td>\n",
       "      <td>1.398</td>\n",
       "      <td>1.563</td>\n",
       "      <td>1.718</td>\n",
       "      <td>1.858</td>\n",
       "      <td>3.647</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>98,848.000</td>\n",
       "      <td>-3.647</td>\n",
       "      <td>61.670</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>0.113</td>\n",
       "      <td>0.136</td>\n",
       "      <td>0.165</td>\n",
       "      <td>0.203</td>\n",
       "      <td>0.246</td>\n",
       "      <td>0.647</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30</th>\n",
       "      <td>98,848.000</td>\n",
       "      <td>-4.139</td>\n",
       "      <td>61.642</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>-1.112</td>\n",
       "      <td>-0.702</td>\n",
       "      <td>-0.266</td>\n",
       "      <td>0.079</td>\n",
       "      <td>0.335</td>\n",
       "      <td>5.870</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>31</th>\n",
       "      <td>98,848.000</td>\n",
       "      <td>3.700</td>\n",
       "      <td>62.166</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>4.989</td>\n",
       "      <td>5.895</td>\n",
       "      <td>7.197</td>\n",
       "      <td>8.807</td>\n",
       "      <td>10.478</td>\n",
       "      <td>23.933</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32</th>\n",
       "      <td>98,848.000</td>\n",
       "      <td>-2.804</td>\n",
       "      <td>61.723</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>0.514</td>\n",
       "      <td>0.753</td>\n",
       "      <td>1.032</td>\n",
       "      <td>1.298</td>\n",
       "      <td>1.492</td>\n",
       "      <td>4.127</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>33</th>\n",
       "      <td>98,848.000</td>\n",
       "      <td>-2.610</td>\n",
       "      <td>61.739</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>0.429</td>\n",
       "      <td>0.649</td>\n",
       "      <td>1.043</td>\n",
       "      <td>1.597</td>\n",
       "      <td>2.220</td>\n",
       "      <td>7.994</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>34</th>\n",
       "      <td>98,848.000</td>\n",
       "      <td>361.795</td>\n",
       "      <td>127.907</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>223.241</td>\n",
       "      <td>305.249</td>\n",
       "      <td>385.986</td>\n",
       "      <td>444.805</td>\n",
       "      <td>476.947</td>\n",
       "      <td>499.999</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>35</th>\n",
       "      <td>98,848.000</td>\n",
       "      <td>-3.748</td>\n",
       "      <td>61.664</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>0.022</td>\n",
       "      <td>0.039</td>\n",
       "      <td>0.066</td>\n",
       "      <td>0.100</td>\n",
       "      <td>0.130</td>\n",
       "      <td>0.344</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>36</th>\n",
       "      <td>98,848.000</td>\n",
       "      <td>-3.524</td>\n",
       "      <td>61.678</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>0.211</td>\n",
       "      <td>0.250</td>\n",
       "      <td>0.294</td>\n",
       "      <td>0.341</td>\n",
       "      <td>0.387</td>\n",
       "      <td>0.771</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>37</th>\n",
       "      <td>98,848.000</td>\n",
       "      <td>-3.441</td>\n",
       "      <td>61.688</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>-0.485</td>\n",
       "      <td>-0.203</td>\n",
       "      <td>0.196</td>\n",
       "      <td>0.759</td>\n",
       "      <td>1.474</td>\n",
       "      <td>8.192</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>38</th>\n",
       "      <td>98,848.000</td>\n",
       "      <td>-6.582</td>\n",
       "      <td>61.496</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>-4.119</td>\n",
       "      <td>-3.473</td>\n",
       "      <td>-2.730</td>\n",
       "      <td>-2.066</td>\n",
       "      <td>-1.529</td>\n",
       "      <td>0.670</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>39</th>\n",
       "      <td>98,848.000</td>\n",
       "      <td>-2.555</td>\n",
       "      <td>61.738</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>0.927</td>\n",
       "      <td>1.085</td>\n",
       "      <td>1.260</td>\n",
       "      <td>1.439</td>\n",
       "      <td>1.608</td>\n",
       "      <td>6.059</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>40</th>\n",
       "      <td>98,848.000</td>\n",
       "      <td>-3.730</td>\n",
       "      <td>61.665</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>0.050</td>\n",
       "      <td>0.065</td>\n",
       "      <td>0.083</td>\n",
       "      <td>0.107</td>\n",
       "      <td>0.140</td>\n",
       "      <td>0.430</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>41</th>\n",
       "      <td>98,848.000</td>\n",
       "      <td>139.850</td>\n",
       "      <td>77.198</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>102.313</td>\n",
       "      <td>121.993</td>\n",
       "      <td>144.317</td>\n",
       "      <td>165.857</td>\n",
       "      <td>184.423</td>\n",
       "      <td>321.234</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>42</th>\n",
       "      <td>98,848.000</td>\n",
       "      <td>-25.357</td>\n",
       "      <td>68.459</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>-56.701</td>\n",
       "      <td>-39.986</td>\n",
       "      <td>-21.744</td>\n",
       "      <td>-3.618</td>\n",
       "      <td>15.364</td>\n",
       "      <td>149.219</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>43</th>\n",
       "      <td>98,848.000</td>\n",
       "      <td>-3.062</td>\n",
       "      <td>61.718</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>-0.337</td>\n",
       "      <td>0.002</td>\n",
       "      <td>0.489</td>\n",
       "      <td>1.184</td>\n",
       "      <td>2.104</td>\n",
       "      <td>43.651</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>44</th>\n",
       "      <td>98,848.000</td>\n",
       "      <td>2.214</td>\n",
       "      <td>62.200</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>2.547</td>\n",
       "      <td>3.463</td>\n",
       "      <td>4.802</td>\n",
       "      <td>6.884</td>\n",
       "      <td>10.700</td>\n",
       "      <td>85.823</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>45</th>\n",
       "      <td>98,848.000</td>\n",
       "      <td>-3.812</td>\n",
       "      <td>61.660</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.002</td>\n",
       "      <td>0.021</td>\n",
       "      <td>0.417</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>46</th>\n",
       "      <td>98,848.000</td>\n",
       "      <td>-3.644</td>\n",
       "      <td>61.670</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>0.085</td>\n",
       "      <td>0.120</td>\n",
       "      <td>0.166</td>\n",
       "      <td>0.222</td>\n",
       "      <td>0.283</td>\n",
       "      <td>0.606</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>47</th>\n",
       "      <td>98,848.000</td>\n",
       "      <td>-3.951</td>\n",
       "      <td>61.658</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>-0.901</td>\n",
       "      <td>-0.308</td>\n",
       "      <td>0.056</td>\n",
       "      <td>0.302</td>\n",
       "      <td>0.511</td>\n",
       "      <td>1.893</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>48</th>\n",
       "      <td>98,848.000</td>\n",
       "      <td>-0.933</td>\n",
       "      <td>61.868</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>0.926</td>\n",
       "      <td>1.455</td>\n",
       "      <td>2.404</td>\n",
       "      <td>3.827</td>\n",
       "      <td>5.569</td>\n",
       "      <td>21.768</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>49</th>\n",
       "      <td>98,848.000</td>\n",
       "      <td>-2.178</td>\n",
       "      <td>61.791</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>-0.022</td>\n",
       "      <td>0.314</td>\n",
       "      <td>1.176</td>\n",
       "      <td>2.321</td>\n",
       "      <td>3.802</td>\n",
       "      <td>24.273</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50</th>\n",
       "      <td>98,848.000</td>\n",
       "      <td>-3.590</td>\n",
       "      <td>61.673</td>\n",
       "      <td>-999.000</td>\n",
       "      <td>0.164</td>\n",
       "      <td>0.198</td>\n",
       "      <td>0.232</td>\n",
       "      <td>0.264</td>\n",
       "      <td>0.295</td>\n",
       "      <td>0.625</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         count     mean      std       min      10%      25%      50%  \\\n",
       "1   98,848.000    1.053   61.972  -999.000    3.409    4.062    4.816   \n",
       "2   98,848.000   -2.283   61.760  -999.000    0.576    0.947    1.427   \n",
       "3   98,848.000  104.047  120.060  -999.000   16.222   47.204  103.797   \n",
       "4   98,848.000   -3.538   61.677  -999.000    0.219    0.242    0.273   \n",
       "5   98,848.000   -3.812   61.660  -999.000    0.000    0.000    0.005   \n",
       "6   98,848.000   -3.659   61.669  -999.000    0.000    0.089    0.142   \n",
       "7   98,848.000   -2.860   61.720  -999.000    0.564    0.708    0.915   \n",
       "8   98,848.000   -3.002   61.710  -999.000    0.734    0.781    0.831   \n",
       "9   98,848.000   -0.377   61.873  -999.000    3.171    3.290    3.430   \n",
       "10  98,848.000   -3.647   61.670  -999.000    0.120    0.143    0.171   \n",
       "11  98,848.000    0.487   61.927  -999.000    4.020    4.156    4.292   \n",
       "12  98,848.000  162.591  134.486  -999.000    7.436   76.677  157.356   \n",
       "13  98,848.000   -3.381   61.689  -999.000   -0.430    0.133    0.619   \n",
       "14  98,848.000   -3.516   61.679  -999.000    0.043    0.109    0.237   \n",
       "15  98,848.000   -1.651   61.796  -999.000    1.564    1.763    2.067   \n",
       "16  98,848.000  857.645  520.960  -999.000  455.099  555.354  714.846   \n",
       "17  98,848.000   -3.404   61.686  -999.000    0.155    0.211    0.305   \n",
       "18  98,848.000   21.088   67.289  -999.000    5.869    9.714   17.879   \n",
       "19  98,848.000   -3.552   61.676  -999.000    0.213    0.235    0.261   \n",
       "20  98,848.000   -3.030   61.710  -999.000    0.196    0.468    0.788   \n",
       "21  98,848.000   -3.906   61.681  -999.000   -2.369   -0.824    0.234   \n",
       "22  98,848.000   -3.514   61.678  -999.000    0.248    0.276    0.306   \n",
       "23  98,848.000   70.738   91.727  -999.000   22.582   35.302   55.898   \n",
       "24  98,848.000    0.049   61.925  -999.000    2.138    2.702    3.544   \n",
       "25  98,848.000   -3.382   61.686  -999.000    0.357    0.413    0.458   \n",
       "26  98,848.000   -4.419   61.658  -999.000   -2.781   -0.942    0.002   \n",
       "27  98,848.000   92.928   92.725  -999.000    7.303   32.997  107.757   \n",
       "28  98,848.000   -2.258   61.756  -999.000    1.251    1.398    1.563   \n",
       "29  98,848.000   -3.647   61.670  -999.000    0.113    0.136    0.165   \n",
       "30  98,848.000   -4.139   61.642  -999.000   -1.112   -0.702   -0.266   \n",
       "31  98,848.000    3.700   62.166  -999.000    4.989    5.895    7.197   \n",
       "32  98,848.000   -2.804   61.723  -999.000    0.514    0.753    1.032   \n",
       "33  98,848.000   -2.610   61.739  -999.000    0.429    0.649    1.043   \n",
       "34  98,848.000  361.795  127.907  -999.000  223.241  305.249  385.986   \n",
       "35  98,848.000   -3.748   61.664  -999.000    0.022    0.039    0.066   \n",
       "36  98,848.000   -3.524   61.678  -999.000    0.211    0.250    0.294   \n",
       "37  98,848.000   -3.441   61.688  -999.000   -0.485   -0.203    0.196   \n",
       "38  98,848.000   -6.582   61.496  -999.000   -4.119   -3.473   -2.730   \n",
       "39  98,848.000   -2.555   61.738  -999.000    0.927    1.085    1.260   \n",
       "40  98,848.000   -3.730   61.665  -999.000    0.050    0.065    0.083   \n",
       "41  98,848.000  139.850   77.198  -999.000  102.313  121.993  144.317   \n",
       "42  98,848.000  -25.357   68.459  -999.000  -56.701  -39.986  -21.744   \n",
       "43  98,848.000   -3.062   61.718  -999.000   -0.337    0.002    0.489   \n",
       "44  98,848.000    2.214   62.200  -999.000    2.547    3.463    4.802   \n",
       "45  98,848.000   -3.812   61.660  -999.000    0.000    0.000    0.000   \n",
       "46  98,848.000   -3.644   61.670  -999.000    0.085    0.120    0.166   \n",
       "47  98,848.000   -3.951   61.658  -999.000   -0.901   -0.308    0.056   \n",
       "48  98,848.000   -0.933   61.868  -999.000    0.926    1.455    2.404   \n",
       "49  98,848.000   -2.178   61.791  -999.000   -0.022    0.314    1.176   \n",
       "50  98,848.000   -3.590   61.673  -999.000    0.164    0.198    0.232   \n",
       "\n",
       "        75%        90%        max  \n",
       "1     5.646      6.392     17.057  \n",
       "2     1.981      2.538      8.606  \n",
       "3   135.201    170.475  1,587.080  \n",
       "4     0.316      0.362      0.713  \n",
       "5     0.012      0.023      0.179  \n",
       "6     0.213      0.350      0.704  \n",
       "7     1.174      1.423      5.979  \n",
       "8     0.871      0.897      0.967  \n",
       "9     3.587      3.755      7.170  \n",
       "10    0.201      0.231      0.525  \n",
       "11    4.436      4.602      9.559  \n",
       "12  245.162    325.683    537.262  \n",
       "13    0.865      0.956      1.000  \n",
       "14    0.460      0.698      1.000  \n",
       "15    2.459      2.897      7.712  \n",
       "16  982.721  1,454.210  5,198.120  \n",
       "17    0.473      0.820      3.362  \n",
       "18   31.922     52.311    288.930  \n",
       "19    0.294      0.335      0.759  \n",
       "20    1.085      1.370      7.537  \n",
       "21    0.955      1.553     22.404  \n",
       "22    0.337      0.366      0.597  \n",
       "23   90.547    148.906    638.484  \n",
       "24    4.617      5.857     78.363  \n",
       "25    0.483      0.493      0.500  \n",
       "26    0.410      0.835      4.621  \n",
       "27  142.887    171.231    382.727  \n",
       "28    1.718      1.858      3.647  \n",
       "29    0.203      0.246      0.647  \n",
       "30    0.079      0.335      5.870  \n",
       "31    8.807     10.478     23.933  \n",
       "32    1.298      1.492      4.127  \n",
       "33    1.597      2.220      7.994  \n",
       "34  444.805    476.947    499.999  \n",
       "35    0.100      0.130      0.344  \n",
       "36    0.341      0.387      0.771  \n",
       "37    0.759      1.474      8.192  \n",
       "38   -2.066     -1.529      0.670  \n",
       "39    1.439      1.608      6.059  \n",
       "40    0.107      0.140      0.430  \n",
       "41  165.857    184.423    321.234  \n",
       "42   -3.618     15.364    149.219  \n",
       "43    1.184      2.104     43.651  \n",
       "44    6.884     10.700     85.823  \n",
       "45    0.002      0.021      0.417  \n",
       "46    0.222      0.283      0.606  \n",
       "47    0.302      0.511      1.893  \n",
       "48    3.827      5.569     21.768  \n",
       "49    2.321      3.802     24.273  \n",
       "50    0.264      0.295      0.625  "
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "vz.pretty_describe(train_X_removed_outliers)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "15835f1f",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.pipeline import Pipeline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "9f7dfe0c",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import RobustScaler"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "8f9953dc",
   "metadata": {},
   "outputs": [],
   "source": [
    "pca_pipeline = Pipeline(\n",
    "    [('standardize', RobustScaler(quantile_range = (0.1, 0.9))),\n",
    "     ('pca', PCA(n_components=0.96, random_state=42))]\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "17b24a81",
   "metadata": {},
   "outputs": [],
   "source": [
    "train_X_pca = pca_pipeline.fit_transform(train_X_removed_outliers)\n",
    "train_y_pca = train_y_removed_outliers.values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "a73befad",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9794930638873893"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.sum(pca_pipeline.named_steps['pca'].explained_variance_ratio_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "b8d0dca7",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['z1', 'z2', 'z3', 'z4']\n"
     ]
    }
   ],
   "source": [
    "column_names_pca = [\"z{0}\".format(i+1) for i in range(train_X_pca.shape[1])]\n",
    "print(column_names_pca)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "7cfd3f6e",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Dimensionally reduced to 4 parameters\n",
    "train_pca = pd.DataFrame(train_X_pca, columns=column_names_pca)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "222dcee4",
   "metadata": {},
   "outputs": [],
   "source": [
    "train_pca['y'] = train_y_pca"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "f57e468b",
   "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>count</th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "      <th>min</th>\n",
       "      <th>10%</th>\n",
       "      <th>25%</th>\n",
       "      <th>50%</th>\n",
       "      <th>75%</th>\n",
       "      <th>90%</th>\n",
       "      <th>max</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>z1</th>\n",
       "      <td>98,848.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.483</td>\n",
       "      <td>-2.378</td>\n",
       "      <td>-0.344</td>\n",
       "      <td>-0.091</td>\n",
       "      <td>0.054</td>\n",
       "      <td>0.143</td>\n",
       "      <td>0.201</td>\n",
       "      <td>6.399</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>z2</th>\n",
       "      <td>98,848.000</td>\n",
       "      <td>-0.000</td>\n",
       "      <td>0.379</td>\n",
       "      <td>-0.461</td>\n",
       "      <td>-0.260</td>\n",
       "      <td>-0.201</td>\n",
       "      <td>-0.106</td>\n",
       "      <td>0.061</td>\n",
       "      <td>0.362</td>\n",
       "      <td>3.474</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>z3</th>\n",
       "      <td>98,848.000</td>\n",
       "      <td>-0.000</td>\n",
       "      <td>0.142</td>\n",
       "      <td>-0.494</td>\n",
       "      <td>-0.198</td>\n",
       "      <td>-0.095</td>\n",
       "      <td>0.014</td>\n",
       "      <td>0.106</td>\n",
       "      <td>0.174</td>\n",
       "      <td>0.535</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>z4</th>\n",
       "      <td>98,848.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.095</td>\n",
       "      <td>-0.175</td>\n",
       "      <td>-0.101</td>\n",
       "      <td>-0.059</td>\n",
       "      <td>-0.005</td>\n",
       "      <td>0.041</td>\n",
       "      <td>0.086</td>\n",
       "      <td>1.325</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>y</th>\n",
       "      <td>98,848.000</td>\n",
       "      <td>0.258</td>\n",
       "      <td>0.438</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         count    mean    std     min     10%     25%     50%    75%    90%  \\\n",
       "z1  98,848.000   0.000  0.483  -2.378  -0.344  -0.091   0.054  0.143  0.201   \n",
       "z2  98,848.000  -0.000  0.379  -0.461  -0.260  -0.201  -0.106  0.061  0.362   \n",
       "z3  98,848.000  -0.000  0.142  -0.494  -0.198  -0.095   0.014  0.106  0.174   \n",
       "z4  98,848.000   0.000  0.095  -0.175  -0.101  -0.059  -0.005  0.041  0.086   \n",
       "y   98,848.000   0.258  0.438   0.000   0.000   0.000   0.000  1.000  1.000   \n",
       "\n",
       "      max  \n",
       "z1  6.399  \n",
       "z2  3.474  \n",
       "z3  0.535  \n",
       "z4  1.325  \n",
       "y   1.000  "
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "vz.pretty_describe(train_pca)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "e153e5b8",
   "metadata": {},
   "outputs": [],
   "source": [
    "test_X_pca = pca_pipeline.transform(test_X)\n",
    "test_y_pca = test_y.values"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a1b902b1",
   "metadata": {},
   "source": [
    "# Trees"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "24295136",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.2580831175137585"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_y_pca.sum() / len(train_y_pca)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d8a627f2",
   "metadata": {},
   "source": [
    "Since class ratio is not too imblanced (~25%), AUC-ROC score will be used as the key performace metric"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "72977405",
   "metadata": {},
   "outputs": [],
   "source": [
    "import time"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "1ce1f83e",
   "metadata": {},
   "outputs": [],
   "source": [
    "import itertools\n",
    "from scipy import stats\n",
    "from scipy.stats import randint as sp_randint\n",
    "from sklearn.model_selection import GridSearchCV, RandomizedSearchCV, cross_val_score\n",
    "from sklearn.tree import DecisionTreeClassifier\n",
    "from sklearn.ensemble import RandomForestClassifier, AdaBoostClassifier\n",
    "from sklearn.metrics import confusion_matrix, roc_curve, roc_auc_score, classification_report, accuracy_score, precision_score, recall_score, f1_score, precision_recall_curve"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "36715b1d",
   "metadata": {},
   "source": [
    "Preliminary Tuning using Single Decision Tree"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "16225850",
   "metadata": {},
   "source": [
    "1. max_depth"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "74a27072",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 3 folds for each of 8 candidates, totalling 24 fits\n",
      "The best parameters: {'max_depth': 8}\n",
      "The best score: 0.9063719347190323\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x360 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "GridSearchCV(cv=3, estimator=DecisionTreeClassifier(), n_jobs=6,\n",
       "             param_grid={'max_depth': [1, 2, 3, 4, 5, 6, 7, 8]},\n",
       "             scoring='roc_auc', verbose=2)"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tm.evaluate_params(estimator=DecisionTreeClassifier(),\n",
    "                   cv_method='grid',\n",
    "                   param_grid={'max_depth' : [i+1 for i in range(8)]},\n",
    "                   x=train_X_pca,\n",
    "                   y=train_y_pca)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cef37a69",
   "metadata": {},
   "source": [
    "2. min_samples_split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "c9e16f20",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 3 folds for each of 100 candidates, totalling 300 fits\n",
      "The best parameters: {'min_samples_split': 0.006061000000000001}\n",
      "The best score: 0.9154355951398175\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x360 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "GridSearchCV(cv=3, estimator=DecisionTreeClassifier(), n_jobs=6,\n",
       "             param_grid={'min_samples_split': array([1.000e-06, 1.020e-04, 2.030e-04, 3.040e-04, 4.050e-04, 5.060e-04,\n",
       "       6.070e-04, 7.080e-04, 8.090e-04, 9.100e-04, 1.011e-03, 1.112e-03,\n",
       "       1.213e-03, 1.314e-03, 1.415e-03, 1.516e-03, 1.617e-03, 1.718e-03,\n",
       "       1.819e-03, 1.920e-03, 2.021e-03, 2.122e-03, 2.223e-03, 2.324e-03,\n",
       "       2.425e-03...\n",
       "       6.667e-03, 6.768e-03, 6.869e-03, 6.970e-03, 7.071e-03, 7.172e-03,\n",
       "       7.273e-03, 7.374e-03, 7.475e-03, 7.576e-03, 7.677e-03, 7.778e-03,\n",
       "       7.879e-03, 7.980e-03, 8.081e-03, 8.182e-03, 8.283e-03, 8.384e-03,\n",
       "       8.485e-03, 8.586e-03, 8.687e-03, 8.788e-03, 8.889e-03, 8.990e-03,\n",
       "       9.091e-03, 9.192e-03, 9.293e-03, 9.394e-03, 9.495e-03, 9.596e-03,\n",
       "       9.697e-03, 9.798e-03, 9.899e-03, 1.000e-02])},\n",
       "             scoring='roc_auc', verbose=2)"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tm.evaluate_params(estimator=DecisionTreeClassifier(),\n",
    "                   cv_method='grid',\n",
    "                   param_grid={'min_samples_split' : np.linspace(0.000001, 0.01, 100)},\n",
    "                   x=train_X_pca,\n",
    "                   y=train_y_pca)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1a0825f6",
   "metadata": {},
   "source": [
    "3. min_samples_leaf"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "id": "827f8909",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 3 folds for each of 100 candidates, totalling 300 fits\n",
      "The best parameters: {'min_samples_leaf': 0.0015160000000000002}\n",
      "The best score: 0.9177498919808692\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x360 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "GridSearchCV(cv=3, estimator=DecisionTreeClassifier(), n_jobs=6,\n",
       "             param_grid={'min_samples_leaf': array([1.000e-06, 1.020e-04, 2.030e-04, 3.040e-04, 4.050e-04, 5.060e-04,\n",
       "       6.070e-04, 7.080e-04, 8.090e-04, 9.100e-04, 1.011e-03, 1.112e-03,\n",
       "       1.213e-03, 1.314e-03, 1.415e-03, 1.516e-03, 1.617e-03, 1.718e-03,\n",
       "       1.819e-03, 1.920e-03, 2.021e-03, 2.122e-03, 2.223e-03, 2.324e-03,\n",
       "       2.425e-03,...\n",
       "       6.667e-03, 6.768e-03, 6.869e-03, 6.970e-03, 7.071e-03, 7.172e-03,\n",
       "       7.273e-03, 7.374e-03, 7.475e-03, 7.576e-03, 7.677e-03, 7.778e-03,\n",
       "       7.879e-03, 7.980e-03, 8.081e-03, 8.182e-03, 8.283e-03, 8.384e-03,\n",
       "       8.485e-03, 8.586e-03, 8.687e-03, 8.788e-03, 8.889e-03, 8.990e-03,\n",
       "       9.091e-03, 9.192e-03, 9.293e-03, 9.394e-03, 9.495e-03, 9.596e-03,\n",
       "       9.697e-03, 9.798e-03, 9.899e-03, 1.000e-02])},\n",
       "             scoring='roc_auc', verbose=2)"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tm.evaluate_params(estimator=DecisionTreeClassifier(),\n",
    "                   cv_method='grid',\n",
    "                   param_grid={'min_samples_leaf' : np.linspace(0.000001, 0.01, 100)},\n",
    "                   x=train_X_pca,\n",
    "                   y=train_y_pca)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d7d42bf5",
   "metadata": {},
   "source": [
    "4. max_leaf_nodes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "id": "7fb99da2",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 3 folds for each of 20 candidates, totalling 60 fits\n",
      "The best parameters: {'max_leaf_nodes': 152}\n",
      "The best score: 0.9129450324448068\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x360 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "GridSearchCV(cv=3, estimator=DecisionTreeClassifier(), n_jobs=6,\n",
       "             param_grid={'max_leaf_nodes': [2, 52, 102, 152, 202, 252, 302, 352,\n",
       "                                            402, 452, 502, 552, 602, 652, 702,\n",
       "                                            752, 802, 852, 902, 952]},\n",
       "             scoring='roc_auc', verbose=2)"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tm.evaluate_params(estimator=DecisionTreeClassifier(),\n",
    "                   cv_method='grid',\n",
    "                   param_grid={'max_leaf_nodes' : [i for i in range(2, 1000, 50)]},\n",
    "                   x=train_X_pca,\n",
    "                   y=train_y_pca)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ce6dac26",
   "metadata": {},
   "source": [
    "# Random Forest Classifier - PCA Reduced"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "a02dbe01",
   "metadata": {},
   "outputs": [],
   "source": [
    "rfc = RandomForestClassifier()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "31fa6464",
   "metadata": {},
   "outputs": [],
   "source": [
    "param_grid_rfc = {'n_estimators' : stats.randint(600, 1000),\n",
    "                  'max_depth' : [i for i in range(6, 9)],\n",
    "                  'min_samples_split' : stats.uniform(0.004, 0.009),\n",
    "                  'min_samples_leaf' : stats.uniform(0.0001, 0.005),\n",
    "                  'max_leaf_nodes' : stats.randint(150, 250),\n",
    "                  'max_features' : stats.randint(3, train_X_pca.shape[1] + 1)}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "id": "fa29fcb3",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 3 folds for each of 25 candidates, totalling 75 fits\n",
      "The best parameters: {'max_depth': 8, 'max_features': 3, 'max_leaf_nodes': 224, 'min_samples_leaf': 0.0010658516513744824, 'min_samples_split': 0.0066432325316914615, 'n_estimators': 723}\n",
      "The best score: 0.9207178169703708\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA+UAAAFpCAYAAAD6LokAAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjQuMywgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/MnkTPAAAACXBIWXMAAAsTAAALEwEAmpwYAAAvCklEQVR4nO3df3Rd5X3n+/f32CIKsg1GyHLABuHGAeyUOqmvS2nIpJBkicyacsPtGEjbJFxyga6AST3TKZOZ23Zm1pqyulr30nvTAk1oknVTwG3ClDulJllMZzwJNOAEQXAwxQEFzA/ZiDYySk+QOd/7h7bUY0Wyj83Z2rL8fq11ls/e+9lb3/Osw0Yf7Wc/OzITSZIkSZI0+2pVFyBJkiRJ0vHKUC5JkiRJUkUM5ZIkSZIkVcRQLkmSJElSRQzlkiRJkiRVZGHVBcwSp5iXJEmSjm9RdQHSdLxSLkmSJElSRQzlkiRJkiRV5HgZvi5JOs6dvvIMXtzzfNVlqI1OW7GSF55/ruoy1Eb+dzr/+N+pdHiReVzcbn1cfEhJ0swigstve7DqMtRGd197AcfJ7zHHDf87nX/m2H+n3lOuOcnh65IkSZIkVcRQLkmSJElSRQzlkiRJkiRVxFAuSZIkSVJFDOWSJEmSJFXEUC5JkiRJUkUM5ZIkSZIkVWRh1QUcb05feQYv7nm+6jLURqetWMkLzz9XdRmSJEmSjkGG8ln24p7nufy2B6suQ21097UXVF2CJEmSpGOUw9clSZIkSaqIoVySJEmSpIoYyiVJkiRJqoihXJIkSZKkihjKJUmSJEmqiKFckiRJkqSKGMolSZIkSaqIoVySJEmSpIoYyiVJkiRJqoihXJIkSZKkihjKJUmSJEmqiKFckiRJkqSKGMolSZIkSaqIoVySJEmSpIoYyiVJkiRJqoihXJIkSZKkihjKJUmSJEmqiKFckiRJkqSKGMolSZIkSaqIoVySJEmSpIoYyiVJkiRJqoihXJIkSZKkihjKJUmSJEmqiKFckiRJkqSKGMolSZIkSaqIoVySJEmSpIoYyiVJkiRJqoihXJIkSZKkihjKJUmSJEmqiKFckiRJkqSKGMolSZIkSaqIoVySJEmSpIpEZlZdQ+kiYhtwatV1FE4FXqm6iHnM/i2X/Vsu+7dc9m+57N/y2Lflsn/LNZf695XM7K+6CGmq4yKUzyURsSMz11ddx3xl/5bL/i2X/Vsu+7dc9m957Nty2b/lsn+lw3P4uiRJkiRJFTGUS5IkSZJUEUP57Lu96gLmOfu3XPZvuezfctm/5bJ/y2Pflsv+LZf9Kx2G95RLkiRJklQRr5RLkiRJklSRUkN5RPRHxFMRsTsibppm+9KIuCciHo+IhyPincX6lRHxNxHxZETsjIgbm/b5l8W6RkQ4k6MkSZIk6ZhVWiiPiAXAZ4BLgDXAlRGxZkqzTwMDmXke8FHglmL9AeBfZea5wPnAJ5v2fQK4DNheVu2SJEmSJM2GhSUeewOwOzOfAYiIu4BLge82tVkD/A5AZu6KiL6I6M3Ml4CXivX7I+JJ4HTgu5n5ZHG8lgvp7+/Pbdu2teEjSZIkSTpGtR4g3iTzh6Yx4/evzOHrpwPPNy3vKdY1e4zxq95ExAbgTGBFc4OI6APeBXzzSH54RFwTETsiYseTTz55ZJVLkiRJ0hEwf+holRnKp/tLwNSp3m8GlkbEAHAD8CjjQ9fHDxCxCPgy8KnMHDmSH56Zt2fm+sxc39PTc0SFS5IkSdKRMH/oaJU5fH0PsLJpeQXwYnODImhfBRDj49GfLV5ERAfjgfxLmfmVEuuUJEmSJKkSZV4pfwRYHRFnRcQJwBXAvc0NIuLkYhvAJ4DtmTlSBPTPAU9m5pYSa5QkSZIkqTKlhfLMPABcD9wPPAlszcydEXFdRFxXNDsX2BkRuxifpX3i0Wc/B/wKcFFEDBSvDwFExIcjYg/ws8BfRcT9ZX0GSZIkSZLKVObwdTLzPuC+KetubXr/ELB6mv2+zgyz02XmPcA97a1UkiRJkqTZV+bwdUmSJEmSdAilXimXJEmSdPxpNJLB4VGGRur0Lumkr7uLWm3WHhMuHVMM5ZIkSZLaptFItu18mc1bB6iPNejsqLFl4zr61y43mEvTcPi6JEmSpLYZHB6dDOQA9bEGm7cOMDg8WnFl0txkKJckSZLUNkMj9clAPqE+1mDv/npFFc2+xx9/nIg44tfpK8+ounRVwOHrkiRJktqmd0knnR21g4J5Z0eNZYs7K6xqdo2NjXH5bQ8e8X53X3tBCdVorvNKuSRJkqS26evuYsvGdXR2jEeNiXvK+7q7Kq5Mmpu8Ui5JkiSpbWq1oH/tcs7ZdCF799dZttjZ16VDMZRLkiRJaqtaLVjVs4hVPYuqLkWa8xy+LkmSJElSRQzlkiRJkiRVxFAuSZIkSVJFDOWSJEmSJFXEUC5JkiRJUkUM5ZIkSZIkVcRQLkmSJElSRQzlkiRJkiRVxFAuSZIkSVJFDOWSJEmSJFXEUC5JkiRJUkUM5ZIkSZIkVcRQLkmSJElSRQzlkiRJkiRVxFAuSZIkSVJFDOWSJEmSJFXEUC5JkiRJUkUM5ZIkSZIkVaTUUB4R/RHxVETsjoibptm+NCLuiYjHI+LhiHhnsX5lRPxNRDwZETsj4samfU6JiK9FxNPFv0vL/AySJEmSJJWltFAeEQuAzwCXAGuAKyNizZRmnwYGMvM84KPALcX6A8C/ysxzgfOBTzbtexPwQGauBh4oliVJkiRJOuaUeaV8A7A7M5/JzNeBu4BLp7RZw3iwJjN3AX0R0ZuZL2Xmt4v1+4EngdOLfS4FvlC8/wLwv5b4GSRJkiRJKk2Zofx04Pmm5T38U7Ce8BhwGUBEbADOBFY0N4iIPuBdwDeLVb2Z+RJA8e+ydhcuSZIkSdJsKDOUxzTrcsryzcDSiBgAbgAeZXzo+vgBIhYBXwY+lZkjR/TDI66JiB0RsWPfvn1HVLgkSZIkHYnm/FF1LTq2LCzx2HuAlU3LK4AXmxsUQfsqgIgI4NniRUR0MB7Iv5SZX2nabSgi3paZL0XE24C90/3wzLwduB1g/fr1U/8YIEmSJElt05w/IsL8oZaVeaX8EWB1RJwVEScAVwD3NjeIiJOLbQCfALZn5kgR0D8HPJmZW6Yc917gY8X7jwF/WdonkCRJkiSpRKWF8sw8AFwP3M/4RG1bM3NnRFwXEdcVzc4FdkbELsZnaZ949NnPAb8CXBQRA8XrQ8W2m4EPRMTTwAeKZUmSJEmSjjllDl8nM+8D7puy7tam9w8Bq6fZ7+tMf086mTkMXNzeSiVJkiRJmn1lDl+XJEmSJEmHYCiXJEmSJKkihnJJkiRJkipiKJckSZIkqSKGckmSJEmSKmIolyRJkiSpIoZySZIkSZIqYiiXJEmSJKkiC6suQJIkSZLmlQjuvvaCI9+vtpCIaH89MzhtxUpeeP65Wft5mp6hXJIkSZLaKZPLb3uw6ioO66j+cKC2c/i6JEmSJEkVMZRLkiRJklQRQ7kkSZIkSRUxlEuSJEmSVBFDuSRJkiRJFTGUS5IkSZJUER+JJkmSJKmtGo1kcHiUoZE6vUs66evuolabvedvS8cSQ7kkSZKktmk0km07X2bz1gHqYw06O2ps2biO/rXLDebSNBy+LkmSJKltBodHJwM5QH2sweatAwwOj1ZcmTQ3GcolSZIktc3QSH0ykE+ojzXYu79eUUXS3GYolyRJktQ2vUs66ew4OGZ0dtRYtrizooqkuc1QLkmSJKlt+rq72LJx3WQwn7invK+7q+LKpLnJid4kSZIktU2tFvSvXc45my5k7/46yxY7+7p0KIZySZIkSW1VqwWrehaxqmdR1aVIc57D1yVJkiRJqohXyiVJ81qjkQwOjzI0Uqd3iUMoJUnS3GIolyTNW41Gsm3ny5PPy52YbKh/7XKDuSRJmhMcvi5JmrcGh0cnAzmMPyd389YBBodHK65MkiRpXKmhPCL6I+KpiNgdETdNs31pRNwTEY9HxMMR8c6mbXdExN6IeGLKPj8VEQ9FxHci4v+LiCVlfgZJ0rFraKQ+Gcgn1Mca7N1fr6giSZKkg5UWyiNiAfAZ4BJgDXBlRKyZ0uzTwEBmngd8FLiladvngf5pDv1Z4KbM/EngHuDX21y6JGme6F3SOfmc3AmdHTWWLe6sqCJJkqSDlXmlfAOwOzOfyczXgbuAS6e0WQM8AJCZu4C+iOgtlrcDr05z3LOB7cX7rwH/Wwm1S5Lmgb7uLrZsXDcZzCfuKe/r7qq4MkmSpHFlTvR2OvB80/Ie4GemtHkMuAz4ekRsAM4EVgBDhzjuE8AvAH8J/Etg5XSNIuIa4BqAM8444yjKlyQd62q1oH/tcs7ZdCF799dZttjZ1yVJ5WjOH9KRKPNK+XS/8eSU5ZuBpRExANwAPAocOMxx/3fgkxHxLWAx8Pp0jTLz9sxcn5nre3p6jqhwSdL8UasFq3oWcf6qU1nVs8hALkkqRXP+qLoWHVvKvFK+h4OvYq8AXmxukJkjwFUAERHAs8VrRsUw9w8W+7wD+OftK1mSJEmSpNlT5pXyR4DVEXFWRJwAXAHc29wgIk4utgF8AtheBPUZRcSy4t8a8O+BW9teuSRJkiRJs6C0UJ6ZB4DrgfuBJ4GtmbkzIq6LiOuKZucCOyNiF+OztN84sX9E3Ak8BJwdEXsi4upi05UR8XfALsavvP9pWZ9BkiRJkqQylTl8ncy8D7hvyrpbm94/BKyeYd8rZ1h/Cwc/Ok2SJEmSpGNSmcPXJUmSJEnSIZR6pVySJEnS8afRSAaHRxkaqdO7xMdRSodiKJckSZLUNo1Gsm3ny2zeOkB9rEFnR40tG9fRv3a5wVyahsPXJUmSJLXN4PDoZCAHqI812Lx1gMHh0York+YmQ7kkSZKkthkaqU8G8gn1sQZ799crqkia2wzlkiRJktqmd0knnR0Hx4zOjhrLFndWVJE0txnKJUmSJLVNX3cXWzaumwzmE/eU93V3VVyZNDc50ZskSZKktqnVgv61yzln04Xs3V9n2WJnX5cOxVAuSZIkqa1qtWBVzyJW9SyquhRpznP4uiRJkiRJFTGUS5IkSZJUEYevS5IkSVI7RXD3tRdUXcXh1RYS4b3+7XTaipW88PxzR7SPoVySJEmS2imTy297sOoqVIGj+WOMoVySNK81Gsng8ChDI3V6lzgDsCRJmlsM5ZKkeavRSLbtfJnNWweojzUmn5Xbv3a5wVySJM0JTvQmSZq3BodHJwM5QH2sweatAwwOj1ZcmSTNb41G8sy+13joe6/wzL7XaDSy6pKkOcsr5ZKkeWtopM7SE0/gsnevYGIemy9/aw9799d9dq4klcRRStKRMZRLkuatt53UyUd/9kxueeDpyV8Mb7x4NcuXdFZdmiTNW8++Mv0opbNvuJCfWOYfRKWpHL4uSZq33mgwGchh/BfDWx54mjcaFRcmSfPY918dnTzvTqiPNXjuVW8dkqZjKJckzVt799en/cVw32v1iiqSpPmv64SFdHYcHDM6O2qceIKDdKXpGMolSfNW75LOaX8xXLbY4euSVJbeJW/hxotXT55/J24d6l3yloork+Ym/1wlSZq3+rq72LJx3Y9NNtTX3VV1aZI0b51xSherexdxzXtX0UioBazuXcQZp3julaZjKJckzWsnLIyDfjE8YaEz/0pSmWq14KKze1l16iL27q+zbHEnfd1dzrwuzcBQLkmatwaHR7n+zx496L7yzo4a92260EeiSVKJarVgVc8iz7VSC7ynXJI0bw2NTD/R29CIE71JkqS5wVAuSZq3TpxxBuAFFVUkSZJ0MEO5JGneev2NN9h00cEzAG+6aDVjPqhckiTNEaWG8ojoj4inImJ3RNw0zfalEXFPRDweEQ9HxDubtt0REXsj4okp+6yLiL+NiIGI2BERG8r8DJKkY1d311u4e8dzXP2eVVx/0du5+j2ruHvHc5zS5WN5JEnS3FDaRG8RsQD4DPABYA/wSETcm5nfbWr2aWAgMz8cEecU7S8utn0e+H+AL0459O8C/yEz/zoiPlQsv6+szyFJOnb1dXfxG/3n+kg0SZI0Z5U5+/oGYHdmPgMQEXcBlwLNoXwN8DsAmbkrIvoiojczhzJze0T0TXPcBJYU708CXizrA0iSjm21WtC/djnnbLrQx/JIkqQ5qcxQfjrwfNPyHuBnprR5DLgM+HoxDP1MYAUwdIjjfgq4PyJ+j/Hh9xdM1ygirgGuATjjjDOOonxJ0nzgY3kkSbOhOX9IR6LMe8qnuwyRU5ZvBpZGxABwA/AocOAwx/1V4NcycyXwa8DnpmuUmbdn5vrMXN/T03NEhUuSJEk6eo1G8sy+13joe6/wzL7XaDSmxoD5pzl/VF2Lji1lXinfA6xsWl7BlKHmmTkCXAUQEQE8W7wO5WPAjcX7Pwc+245iJUmSJL15jUaybefLPzafR//a5d4+JE2jzCvljwCrI+KsiDgBuAK4t7lBRJxcbAP4BLC9COqH8iLwz4r3FwFPt7FmSZIkSW/C4PDoZCAHqI812Lx1gMHh0York+am0q6UZ+aBiLgeuB9YANyRmTsj4rpi+63AucAXI+INxieAu3pi/4i4k/FZ1U+NiD3Ab2Xm54D/A7glIhYCdbxvQ5IkSZozhkbqk4F8Qn2swd79def3kKZR5vB1MvM+4L4p625tev8QsHqGfa+cYf3XgZ9uY5mSJEmS2qR3SSedHbWDgnlnR41lizsrrEqau8ocvi5JkiTpONPX3cWWjevo7BiPGhP3lPd1d1VcmTQ3lXqlXJIkSdLxpVYL+tcu55xNF7J3f51lizvp6+5ykjdpBoZySZIkSW1VqwWrehZ5D7nUgpaHr0fEWyPi7DKLkSRJkiTpeNJSKI+IfwEMANuK5XURce8hd5IkSZIkSYfU6pXy3wY2AP8AkJkDQF8ZBUmSJEmSdLxoNZQfyMwflFqJJEmSJEnHmVYnensiIj4CLIiI1cAm4MHyypIkSZIkaf5r9Ur5DcBa4EfAnwE/AD5VUk2SJEmSJB0XDnulPCIWAPdm5vuBf1d+SZIkSZKOZY1GMjg8ytBInd4lPqdcOpTDhvLMfCMifhgRJ3lfuSRJkqRDaTSS//bUEI/v+QGNhAUBP7niJC46u9dgLk2j1XvK68B3IuJrwOjEyszcVEpVkiRJko5Jz706ytNDr3H79meojzXo7Khx48WreXvPIvpOXVR1edKc02oo/6viJUnSMcUhlJI0u4ZGfsQtDzxNfawBQH2swS0PPM27z1hqKJem0VIoz8wvRMQJwDuKVU9l5lh5ZUmS9OY1Gsm2nS+zeevA5NWaLRvX0b92ucFckkoy+vqByUA+oT7W4IevH6ioImluaymUR8T7gC8Ag0AAKyPiY5m5vbTKJEl6kwaHRycDOYz/Urh56wDnbLqQVT1erZGkMpx5ShedHbWDgnlnR40zTumqsKpZFsHd115QdRWqwGkrVh7xPq0OX/994IOZ+RRARLwDuBP46SP+iZIkzZKhkfq0V2v27q8byiWpJGed2sWWjet+bJTSWaceR6E8k8tve7DqKuaku6+9gMysuow5pdVQ3jERyAEy8+8ioqOkmiRJaoveJZ3TXq1ZtrizwqokaX6r1YL+tcs5Z9OF7N1fZ9li5/OQDqXWYrsdEfG5iHhf8foT4FtlFiZJ0pvV1z1+taazY/x/dxNXa/q6j6OrNZJUgVotWNWziPNXncqqnkUGcukQWr1S/qvAJ4FNjN9Tvh34o7KKkiSpHWq14IPn9nL3Nefz0g/qvO2kTta+7SR/OZQkSXNGq6F8IXBLZm4BiIgFwFtKq0qSpDZoNJL//vReHt/zAxoJT740wr7XfsRFZ/cazCVJ0pzQaih/AHg/8Fqx/Fbgq4BTCkqS5qznXh3l6aHXuH37M5OTDd148Wre3rPIZ+VKkqQ5odV7yjszcyKQU7w/sZySJElqj6GRH3HLA08f9Ei0Wx54mqGRH1VcmSRJ0rhWQ/loRLx7YiEi1gP/WE5JkiS1x+jrB6Z9JNoPXz9QUUWSJEkHa3X4+o3An0fEi0ACpwGXl1aVJEltcOYpXdM+Eu2MU5x9XZIkzQ2tXik/C3gX47Owfw14ivFwLknSnHXWqdM/Eu2sUw3lkiRpbmj1Svn/mZl/HhEnAx8Afh/4Y+BnyipMkqQ3q1YL+tcu55xNF7J3f51lizvp6+5y5nVJkjRntBrK3yj+/efArZn5lxHx2+WUJElS+9RqwaqeRazqcbZ1SZotjUYyODzK0Eid3iX+QVQ6lFaHr78QEbcBG4H7IuItrewbEf0R8VRE7I6Im6bZvjQi7omIxyPi4Yh4Z9O2OyJib0Q8MWWfuyNioHgNRsRAi59BkiRJUskajWTbzpf50B/+T678k2/yoT/8n2zb+TKNhne/StNpNZRvBO4H+jPzH4BTgF8/1A4RsQD4DHAJsAa4MiLWTGn2aWAgM88DPgrc0rTt80D/1ONm5uWZuS4z1wFfBr7S4meQJEmSVLLB4VE2bx046HGUm7cOMDg8WnFl0tzUUijPzB9m5lcy8+li+aXM/OphdtsA7M7MZzLzdeAu4NIpbdYADxTH3AX0RURvsbwdeHWmg0dEMP7Hgjtb+QySJEmSyjc0Up/2cZR799crqkia21q9Un40Tgeeb1reU6xr9hhwGUBEbADOBFa0ePwLgaGJPxRIkiRJql7vks7Jp15M6OyosWxxZ0UVSXNbmaF8upkcpt5IcjOwtLgv/AbgUeBAi8e/kkNcJY+IayJiR0Ts2LdvX4uHlCRJkvRm9HVP/zjKvu75/TjK5vxRdS06trQ6+/rR2AOsbFpeAbzY3CAzR4CrYHI4+rPF65AiYiHjV9h/eqY2mXk7cDvA+vXrnVVCkiRJmgXH6+Mom/NHRJg/1LIyQ/kjwOqIOAt4AbgC+Ehzg+K55z8s7jn/BLC9COqH835gV2buaW/JkiRJkt4sH0cpta604euZeQC4nvFZ258Etmbmzoi4LiKuK5qdC+yMiF2Mz9J+48T+EXEn8BBwdkTsiYirmw5/BU7wJkmSJEk6xpV5pZzMvA+4b8q6W5vePwSsnmHfKw9x3I+3qURJkiRJkipT5kRvkiRJkiTpEAzlkiRJkiRVxFAuSZIkSVJFSr2nXJKkqjUayeDwKEMjdXqXHB+P5ZEkSccOQ7kkad5qNJJtO19m89YB6mMNOjtqbNm4jv61yw3mkiRpTnD4uiRp3hocHp0M5AD1sQabtw4wODxacWWSJEnjDOWSpHlraKQ+Gcgn1Mca7N1fr6giSZKkgxnKJUnzVu+STjo7Dv5fXWdHjWWLOyuqSJIk6WCGcknSvNXX3cWWjesmg/nEPeV93V0VVyZJkjTOid4kSfNWrRb0r13OOZsuZO/+OssWO/u6JEmaWwzlkqR5rVYLVvUsYlXPoqpLkSRJ+jEOX5ckSZIkqSKGckmSJEmSKmIolyRJkiSpIoZySZIkSZIqYiiXJEmSJKkihnJJkiRJkipiKJckSZIkqSI+p1ySJEmS2imCu6+9oOoq5qTTVqysuoQ5x1AuSZIkSe2UyeW3PXjEu9197QVkZgkFaS5z+LokSZIkSRUxlEuSJEmSVBFDuSRJkiRJFTGUS5IkSZJUEUO5JEmSJEkVMZRLkiRJklQRQ7kkSZIkSRUxlEuSJEmSVJFSQ3lE9EfEUxGxOyJummb70oi4JyIej4iHI+KdTdvuiIi9EfHENPvdUBx3Z0T8bpmfQZIkSZKkspQWyiNiAfAZ4BJgDXBlRKyZ0uzTwEBmngd8FLiladvngf5pjvvzwKXAeZm5Fvi99lcvSZIkSVL5yrxSvgHYnZnPZObrwF2Mh+lma4AHADJzF9AXEb3F8nbg1WmO+6vAzZn5o6Ld3pLqlyRJkiSpVGWG8tOB55uW9xTrmj0GXAYQERuAM4EVhznuO4ALI+KbEfE/IuJ/ma5RRFwTETsiYse+ffuO6gNIkiRJUiua80fVtejYUmYoj2nW5ZTlm4GlETEA3AA8Chw4zHEXAkuB84FfB7ZGxI/9rMy8PTPXZ+b6np6eI61dkiRJklrWnD+qrkXHloUlHnsPsLJpeQXwYnODzBwBrgIogvWzxetwx/1KZibwcEQ0gFMBL4dLkiRJko4pZV4pfwRYHRFnRcQJwBXAvc0NIuLkYhvAJ4DtRVA/lP8CXFTs/w7gBOCVdhYuSZIkSdJsKC2UZ+YB4HrgfuBJYGtm7oyI6yLiuqLZucDOiNjF+CztN07sHxF3Ag8BZ0fEnoi4uth0B7CqeFTaXcDHiqvmkiRJkiQdU8ocvk5m3gfcN2XdrU3vHwJWz7DvlTOsfx345TaWKUmSJElSJcocvi5JkiRJkg7BUC5JkiRJUkUM5ZIkSZIkVcRQLkmSJElSRQzlkiRJkiRVxFAuSZIkSVJFDOWSJEmSJFXEUC5JkiRJUkUM5ZIkSZIkVcRQLkmSJElSRRZWXcDxotFIBodHGRqp07ukk77uLmq1qLosSZIkSVKFDOWzoNFItu18mc1bB6iPNejsqLFl4zr61y43mEuSJEnScczh67NgcHh0MpAD1McabN46wODwaMWVSZIkSZKqZCifBUMj9clAPqE+1mDv/npFFUmSJEmS5gJD+SzoXdJJZ8fBXd3ZUWPZ4s6KKpIkSZIkzQWG8lnQ193Flo3rJoP5xD3lfd1dFVcmSZIkSaqSE73Nglot6F+7nHM2Xcje/XWWLXb2dUmSJEmSoXzW1GrBqp5FrOpZVHUpkiRJkqQ5wuHrkiRJkiRVxCvls6TRSAaHRxkaqdO7xOHrkiRJkiRD+axoNJJtO1+efFb5xERv/WuXG8wlSZKk+SaCu6+94Mj3qy0kwnwwndNWrOSF55+ruoxSGMpnweDw6GQgh/FnlG/eOsA5my70HnNJkiRpvsnk8tserLqKeeWo/shxjPCe8lkwNFKfDOQT6mMN9u6vV1SRJEmSJGkuMJTPgt4lnZPPKJ/Q2VFj2eLOiiqSJEmSJM0FhvJZ0NfdxZaN6yaD+cQ95X3dXRVXJkmSJEmqkveUz4JaLehfu5xzNl3I3v11li129nVJkiRJkqF81tRqwaqeRU7sJkmSJEmaVOrw9Yjoj4inImJ3RNw0zfalEXFPRDweEQ9HxDubtt0REXsj4okp+/x2RLwQEQPF60NlfgZJkiRJkspSWiiPiAXAZ4BLgDXAlRGxZkqzTwMDmXke8FHglqZtnwf6Zzj8H2TmuuJ1X3srlyRJkiRpdpR5pXwDsDszn8nM14G7gEuntFkDPACQmbuAvojoLZa3A6+WWJ8kSZIkSZUqM5SfDjzftLynWNfsMeAygIjYAJwJrGjh2NcXQ97viIil0zWIiGsiYkdE7Ni3b9+RVy9JkiRJLWrOH1XXomNLmaF8uqnFc8ryzcDSiBgAbgAeBQ4c5rh/DPwEsA54Cfj96Rpl5u2ZuT4z1/f09BxB2eVoNJJn9r3GQ997hWf2vUajMbUrJEmSJB2rmvNH1bXo2FLm7Ot7gJVNyyuAF5sbZOYIcBVARATwbPGaUWYOTbyPiD8B/mub6i1No5Fs2/kym7cOUB9rTD6nvH/tch+LJkmSJEnHsTKvlD8CrI6IsyLiBOAK4N7mBhFxcrEN4BPA9iKozygi3ta0+GHgiZnazhWDw6OTgRygPtZg89YBBodHK65MkiRJklSl0kJ5Zh4ArgfuB54Etmbmzoi4LiKuK5qdC+yMiF2Mz9J+48T+EXEn8BBwdkTsiYiri02/GxHfiYjHgZ8Hfq2sz9AuQyP1yUA+oT7WYGikXlFFkiRJkqS5oMzh6xSPK7tvyrpbm94/BKyeYd8rZ1j/K+2scTaceMJCOjtqBwXzzo4aJ56woMKqJEmSJElVK3P4ugqvv/EGmy5aTWfHeHd3dtTYdNFqxt5oHGZPSZIkSdJ8VuqVco3r7noLd+94jqvfs4oIyIS7dzxH/zuXV12aJEmSJKlChvJZ0NfdxW/0n/tjs6/3dXdVXZokSZIkqUKG8llQqwX9a5dzzqYL2bu/zrLFnfR1d/k4NEmSJEk6zhnKZ0mtFqzqWcSqnkVVlyJJkiRJmiOc6E2SJEmSpIoYyiVJkiRJqoihXJIkSZKkihjKJUmSJEmqiKFckiRJkqSKGMolSZIkSaqIj0SbJY1GMjg8ytBInd4lPqdckiRJkmQonxWNRrJt58ts3jpAfaxBZ0eNLRvX0b92ucFckiRJko5jDl+fBYPDo5OBHKA+1mDz1gEGh0crrkySJEmSVCVD+SwYGqlPBvIJ9bEGe/fXK6pIkiRJkjQXGMpnwbLFnXR2HNzVnR01ehZ1VlSRJEmSJGkuMJTPggU1uPHi1ZPBvLOjxo0Xr2aBvS9JkiRJxzUnepsFL/2gzhcf+j5Xv2cVEZAJX3zo+7zrjJPpO3VR1eVJkiRJkipiKJ8FvUs6+fsfvs5n/mb35LrOjhrLFjt8XZIkSZKOZw6gngV93V1s2bjuoOHrWzauo6+7q+LKJEmSJElV8kr5LKjVgv61yzln04Xs3V9n2eJO+rq7fEa5JEmSJB3nvFI+yzKrrkCSJEmSNFd4pXwWNBrJtp0vs3nrAPWxxuTw9f61y71aLkmSJEnHMa+Uz4LB4dHJQA5QH2uweesAg8OjFVcmSZIkSaqSV8pnwdBIfTKQT6iPNdi7v86qHh+JJkmSJM0nHR0d3H3tBVWXMa+ctmJl1SWUxlA+C3qXdNLZUTsomPtINEmSJGl+Ou+889ixY0fVZegY4fD1WeAj0SRJkiRJ0/FK+SzwkWiSJEmSpOmUeqU8Ivoj4qmI2B0RN02zfWlE3BMRj0fEwxHxzqZtd0TE3oh4YoZj/+uIyIg4tczP0C61WrCqZxHnrzqVVT2LDOSSJEmSpPJCeUQsAD4DXAKsAa6MiDVTmn0aGMjM84CPArc0bfs80D/DsVcCHwCea3PZkiRJkiTNmjKvlG8AdmfmM5n5OnAXcOmUNmuABwAycxfQFxG9xfJ24NUZjv0HwL8BsozCJUmSJEmaDWWG8tOB55uW9xTrmj0GXAYQERuAM4EVhzpoRPwC8EJmPnaYdtdExI6I2LFv374jrV2SJEmSWmb+0NEqM5RPd9P01CvbNwNLI2IAuAF4FDgw4wEjTgT+HfCbh/vhmXl7Zq7PzPU9PT0tFy1JkiRJR8r8oaNV5uzre4DmJ7yvAF5sbpCZI8BVABERwLPFayY/AZwFPDbenBXAtyNiQ2a+3L7SJUmSJEkqX5mh/BFgdUScBbwAXAF8pLlBRJwM/LC45/wTwPYiqE8rM78DLGvafxBYn5mvtL16SZIkSZJKVtrw9cw8AFwP3A88CWzNzJ0RcV1EXFc0OxfYGRG7GJ+l/caJ/SPiTuAh4OyI2BMRV5dVqyRJkiRJVYjM+T+BeUTsA75fdR2FUwGv7JfH/i2X/Vsu+7dc9m+57N/y2Lflsn/LNZf695XMnPaRy+0WEdtm62fp2HdchPK5JCJ2ZOb6quuYr+zfctm/5bJ/y2X/lsv+LY99Wy77t1z2r3R4Zc6+LkmSJEmSDsFQLkmSJElSRQzls+/2qguY5+zfctm/5bJ/y2X/lsv+LY99Wy77t1z2r3QY3lMuSZIkSVJFvFIuSZIkSVJFDOWSJEmSJFXEUN5GEXFyRPxFROyKiCcj4menbI+I+MOI2B0Rj0fEu5u29UfEU8W2m2a/+rmvhf79paJfH4+IByPip5q2DUbEdyJiICJ2zH71c1sLffu+iPhB0X8DEfGbTdv87h5GC/376019+0REvBERpxTb/O4eQkSc3dR3AxExEhGfmtLGc+9RarF/PfcepRb71/PvUWqxfz3/vgkR8WsRsbPouzsjonPKds+/Uisy01ebXsAXgE8U708ATp6y/UPAXwMBnA98s1i/APgesKrY7zFgTdWfZ669WujfC4ClxftLJvq3WB4ETq36M8zVVwt9+z7gv06zn9/dNvTvlLb/AvhvTct+d1vv5wXAy8CZU9Z77i23fz33ltu/nn9L7N8pbTz/Hlmfng48C7y1WN4KfHxKG8+/vny18PJKeZtExBLgvcDnADLz9cz8hynNLgW+mOP+Fjg5It4GbAB2Z+Yzmfk6cFfRVoVW+jczH8zMvy8W/xZYMatFHqNa/O7OxO/uYRxF/14J3DkLpc1HFwPfy8zvT1nvubc9pu1fz71tM9P3dyZ+f49MK/3r+ffILQTeGhELgROBF6ds9/wrtcBQ3j6rgH3An0bEoxHx2YjomtLmdOD5puU9xbqZ1uuftNK/za5m/C+zExL4akR8KyKuKbPQY1CrffuzEfFYRPx1RKwt1vndPbyWv7sRcSLQD3y5abXf3dZdwfS/UHvubY+Z+reZ596jd6j+9fz75h3y++v598hl5gvA7wHPAS8BP8jMr05p5vlXaoGhvH0WAu8G/jgz3wWMAlPvj4lp9stDrNc/aaV/AYiIn2f8F8PfaFr9c5n5bsaHVn4yIt5bcr3Hklb69tuMD/n7KeD/Bv5Lsd7v7uG1/N1lfOjkNzLz1aZ1fndbEBEnAL8A/Pl0m6dZ57n3CBymfyfaeO49SofpX8+/b1Ir3188/x6xiFjK+NXts4DTgK6I+OWpzabZ1fOvNIWhvH32AHsy85vF8l8w/ov41DYrm5ZXMD7MZ6b1+iet9C8RcR7wWeDSzByeWJ+ZLxb/7gXuYXzYlMYdtm8zcyQzXyve3wd0RMSp+N1tRUvf3cKPXcnxu9uyS4BvZ+bQNNs89755h+pfz71v3oz96/m3LQ75/S14/j1y7weezcx9mTkGfIXxOSaaef6VWmAob5PMfBl4PiLOLlZdDHx3SrN7gY8WM1Gez/gwn5eAR4DVEXFW8dfcK4q2KrTSvxFxBuP/Q/iVzPy7pvVdEbF44j3wQeCJWSn8GNBi3y6PiCjeb2D83DGM393DavHcQEScBPwz4C+b1vndbd2h7gX13Pvmzdi/nnvb4lD96/n3zTvkveKef4/ac8D5EXFi8R29GHhyShvPv1ILFlZdwDxzA/Cl4uTyDHBVRFwHkJm3AvcxPgvlbuCHwFXFtgMRcT1wP+OzUd6RmTsrqH+uO1z//ibQDfxR8fvLgcxcD/QC9xTrFgJ/lpnbKqh/Ljtc3/4i8KsRcQD4R+CKzEzA725rDte/AB8GvpqZo037+d1tQXEv6AeAa5vWee5tkxb613Pvm9BC/3r+fRNa6F/w/HtUMvObEfEXjN9icQB4FLjd86905GL8vC5JkiRJkmabw9clSZIkSaqIoVySJEmSpIoYyiVJkiRJqoihXJIkSZKkihjKJUmSJEmqiKFckiRJkqSKGMolScediBiMiFOPct+PR8Rp7TiWJEmSoVySpCPzceC0wzWSJElqhaFcklSZiOiLiF0R8dmIeCIivhQR74+Ib0TE0xGxoXg9GBGPFv+eXey7OSLuKN7/ZLH/iTP8nO6I+GpxjNuAaNr2yxHxcEQMRMRtEbGgWP9aRPx+RHw7Ih6IiJ6I+EVgPfClov1bi8PcULT7TkScU2afSZKk+cVQLkmq2tuBW4DzgHOAjwDvAf418GlgF/DezHwX8JvAfy72+7+At0fEh4E/Ba7NzB/O8DN+C/h6cYx7gTMAIuJc4HLg5zJzHfAG8EvFPl3AtzPz3cD/AH4rM/8C2AH8Umauy8x/LNq+UrT746JuSZKkliysugBJ0nHv2cz8DkBE7AQeyMyMiO8AfcBJwBciYjWQQAdAZjYi4uPA48BtmfmNQ/yM9wKXFfv9VUT8fbH+YuCngUciAuCtwN5iWwO4u3j//wJfOcTxJ7Z9a+LnSJIktcJQLkmq2o+a3jealhuM/3/qPwF/k5kfjog+4L83tV8NvEZr93jnNOsC+EJm/tuj3H/CRM1v4P9bJUnSEXD4uiRprjsJeKF4//GJlRFxEuPD3t8LdBf3e89kO8Ww9Ii4BFharH8A+MWIWFZsOyUiziy21YCJY34E+Hrxfj+w+E18HkmSpEmGcknSXPe7wO9ExDeABU3r/wD4o8z8O+Bq4OaJcD2N/wC8NyK+DXwQeA4gM78L/HvgqxHxOPA14G3FPqPA2oj4FnAR8B+L9Z8Hbp0y0ZskSdJRicxDjcaTJOn4FBGvZeaiquuQJEnzm1fKJUmSJEmqiFfKJUnzRkRcBdw4ZfU3MvOTVdQjSZJ0OIZySZIkSZIq4vB1SZIkSZIqYiiXJEmSJKkihnJJkiRJkipiKJckSZIkqSL/P1D17fslyMe9AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 1080x360 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x360 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x360 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x360 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x360 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x360 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "rfc_grid = tm.evaluate_params(estimator=rfc,\n",
    "                              cv_method='randomized',\n",
    "                              param_grid=param_grid_rfc,\n",
    "                              x=train_X_pca,\n",
    "                              y=train_y_pca,\n",
    "                              cv=3,\n",
    "                              scoring='roc_auc',\n",
    "                              n_iter=25,\n",
    "                              verbose=2,\n",
    "                              n_jobs=6,\n",
    "                              random_state=123)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f66b856a",
   "metadata": {},
   "source": [
    "# Random Forest Classifier - Final"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "id": "5f4a8867",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Using the best parameters\n",
    "rfc_final = RandomForestClassifier(n_estimators = 800,\n",
    "                                   criterion = 'gini',\n",
    "                                   max_depth = 8,\n",
    "                                   max_features = 4,\n",
    "                                   max_leaf_nodes = 300,\n",
    "                                   min_samples_leaf = 0.0035,\n",
    "                                   min_samples_split = 0.006,\n",
    "                                   verbose = False,\n",
    "                                   random_state = 42,\n",
    "                                   n_jobs = -1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "id": "17a11dfa",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "RandomForestClassifier(max_depth=8, max_features=4, max_leaf_nodes=300,\n",
       "                       min_samples_leaf=0.0035, min_samples_split=0.006,\n",
       "                       n_estimators=800, n_jobs=-1, random_state=42,\n",
       "                       verbose=False)"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Fit the training set\n",
    "rfc_final.fit(train_X_pca, train_y_pca)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "id": "493d3332",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Predict on training set\n",
    "train_y_pred_rfc = rfc_final.predict(train_X_pca)\n",
    "train_y_prob_rfc = rfc_final.predict_proba(train_X_pca)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "id": "8abba7a4",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Predict on test set\n",
    "test_y_pred_rfc = rfc_final.predict(test_X_pca)\n",
    "test_y_prob_rfc = rfc_final.predict_proba(test_X_pca)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "id": "37a5a78d",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[17380  1333]\n",
      " [ 2170  5130]]\n"
     ]
    }
   ],
   "source": [
    "print(confusion_matrix(test_y_pca, rfc_final.predict(test_X_pca) > 0.5))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "id": "91726d07",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.7454770035602702\n"
     ]
    }
   ],
   "source": [
    "print(f1_score(test_y_pca, rfc_final.predict(test_X_pca) > 0.5))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "id": "8f151523",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ROC AUC - Train\n",
      "0.9224565508853997\n",
      "ROC AUC - Test\n",
      "0.9220258241102626\n"
     ]
    }
   ],
   "source": [
    "# ROC AUC Scores\n",
    "print('ROC AUC - Train')\n",
    "print(roc_auc_score(train_y_removed_outliers, train_y_prob_rfc[:,1]))\n",
    "print('ROC AUC - Test')\n",
    "print(roc_auc_score(test_y, test_y_prob_rfc[:,1]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "id": "be7b73e4",
   "metadata": {
    "scrolled": false
   },
   "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": [
    "# ROC AUC curves\n",
    "fpr_rfc_tr, tpr_rfc_tr, _ = roc_curve(train_y_removed_outliers, train_y_prob_rfc[:,1])\n",
    "fpr_rfc_te, tpr_rfc_te, _ = roc_curve(test_y, test_y_prob_rfc[:,1])\n",
    "\n",
    "vz.plot_roc_curve(fpr_rfc_tr, tpr_rfc_tr, \"Training\")\n",
    "vz.plot_roc_curve(fpr_rfc_te, tpr_rfc_te, \"Test\")\n",
    "\n",
    "plt.title('Random Forest Classifier', fontsize=16)\n",
    "plt.xlabel('False Positive Rate', fontsize=12)\n",
    "plt.ylabel('True Positive Rate', fontsize=12)\n",
    "plt.legend(loc=\"lower right\", fontsize=12)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6a5e2bde",
   "metadata": {},
   "source": [
    "# XGBoost"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "id": "a6c37eb4",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3-fold AUC:  0.9296105493299902\n"
     ]
    }
   ],
   "source": [
    "from sklearn.model_selection import cross_val_score\n",
    "cross_val_scores = cross_val_score(pipeline, train_X_pca, train_y_pca, scoring='roc_auc', cv=3, error_score='raise')\n",
    "\n",
    "# Print avg. AUC\n",
    "print(\"3-fold AUC: \", np.mean(cross_val_scores))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8dfec3d2",
   "metadata": {},
   "source": [
    "# Neural Network "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "id": "25550c9c",
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras import Sequential\n",
    "from keras.layers import Dense"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "id": "ebceb177",
   "metadata": {},
   "outputs": [],
   "source": [
    "nn = Sequential()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "id": "d792d62c",
   "metadata": {},
   "outputs": [],
   "source": [
    "# First layer\n",
    "nn.add(Dense(4, activation='relu', kernel_initializer='random_normal'))\n",
    "# Second layer\n",
    "nn.add(Dense(4, activation='relu', kernel_initializer='random_normal'))\n",
    "# Output layer\n",
    "nn.add(Dense(1, activation='sigmoid', kernel_initializer='random_normal'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "id": "1f52edd5",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Compile the neural network\n",
    "nn.compile(optimizer='adam', loss='binary_crossentropy', metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "id": "8ea5969b",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/100\n",
      "99/99 [==============================] - 1s 2ms/step - loss: 0.6781 - accuracy: 0.7419\n",
      "Epoch 2/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.6242 - accuracy: 0.7713\n",
      "Epoch 3/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.5347 - accuracy: 0.8142\n",
      "Epoch 4/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.4630 - accuracy: 0.8195\n",
      "Epoch 5/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.4362 - accuracy: 0.8208\n",
      "Epoch 6/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.4236 - accuracy: 0.8235\n",
      "Epoch 7/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.4086 - accuracy: 0.8287\n",
      "Epoch 8/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3912 - accuracy: 0.8349\n",
      "Epoch 9/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3723 - accuracy: 0.8419\n",
      "Epoch 10/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3556 - accuracy: 0.8484\n",
      "Epoch 11/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3432 - accuracy: 0.8531\n",
      "Epoch 12/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3344 - accuracy: 0.8588\n",
      "Epoch 13/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3283 - accuracy: 0.8635\n",
      "Epoch 14/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3240 - accuracy: 0.8650\n",
      "Epoch 15/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3209 - accuracy: 0.8669\n",
      "Epoch 16/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3187 - accuracy: 0.8673\n",
      "Epoch 17/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3170 - accuracy: 0.8682\n",
      "Epoch 18/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3158 - accuracy: 0.8684\n",
      "Epoch 19/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3148 - accuracy: 0.8688\n",
      "Epoch 20/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3141 - accuracy: 0.8691\n",
      "Epoch 21/100\n",
      "99/99 [==============================] - 0s 1ms/step - loss: 0.3135 - accuracy: 0.8692\n",
      "Epoch 22/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3129 - accuracy: 0.8693\n",
      "Epoch 23/100\n",
      "99/99 [==============================] - 0s 1ms/step - loss: 0.3125 - accuracy: 0.8695\n",
      "Epoch 24/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3120 - accuracy: 0.8695\n",
      "Epoch 25/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3117 - accuracy: 0.8696\n",
      "Epoch 26/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3114 - accuracy: 0.8697\n",
      "Epoch 27/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3110 - accuracy: 0.8699\n",
      "Epoch 28/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3108 - accuracy: 0.8696\n",
      "Epoch 29/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3106 - accuracy: 0.8698\n",
      "Epoch 30/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3103 - accuracy: 0.8697\n",
      "Epoch 31/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3103 - accuracy: 0.8696\n",
      "Epoch 32/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3099 - accuracy: 0.8697\n",
      "Epoch 33/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3098 - accuracy: 0.8694\n",
      "Epoch 34/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3097 - accuracy: 0.8697\n",
      "Epoch 35/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3095 - accuracy: 0.8697\n",
      "Epoch 36/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3095 - accuracy: 0.8696\n",
      "Epoch 37/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3094 - accuracy: 0.8697\n",
      "Epoch 38/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3093 - accuracy: 0.8697\n",
      "Epoch 39/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3093 - accuracy: 0.8698\n",
      "Epoch 40/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3093 - accuracy: 0.8697\n",
      "Epoch 41/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3091 - accuracy: 0.8697\n",
      "Epoch 42/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3092 - accuracy: 0.8696\n",
      "Epoch 43/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3091 - accuracy: 0.8697\n",
      "Epoch 44/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3090 - accuracy: 0.8698\n",
      "Epoch 45/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3090 - accuracy: 0.8694\n",
      "Epoch 46/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3090 - accuracy: 0.8697\n",
      "Epoch 47/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3090 - accuracy: 0.8696\n",
      "Epoch 48/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3089 - accuracy: 0.8696\n",
      "Epoch 49/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3088 - accuracy: 0.8698\n",
      "Epoch 50/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3088 - accuracy: 0.8699\n",
      "Epoch 51/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3088 - accuracy: 0.8698\n",
      "Epoch 52/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3088 - accuracy: 0.8694\n",
      "Epoch 53/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3088 - accuracy: 0.8699\n",
      "Epoch 54/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3090 - accuracy: 0.8696\n",
      "Epoch 55/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3087 - accuracy: 0.8697\n",
      "Epoch 56/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3087 - accuracy: 0.8701\n",
      "Epoch 57/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3086 - accuracy: 0.8697\n",
      "Epoch 58/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3087 - accuracy: 0.8698\n",
      "Epoch 59/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3086 - accuracy: 0.8696\n",
      "Epoch 60/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3086 - accuracy: 0.8699\n",
      "Epoch 61/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3085 - accuracy: 0.8696\n",
      "Epoch 62/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3086 - accuracy: 0.8697\n",
      "Epoch 63/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3085 - accuracy: 0.8695\n",
      "Epoch 64/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3085 - accuracy: 0.8699\n",
      "Epoch 65/100\n",
      "99/99 [==============================] - 0s 1ms/step - loss: 0.3085 - accuracy: 0.8695\n",
      "Epoch 66/100\n",
      "99/99 [==============================] - 0s 1ms/step - loss: 0.3084 - accuracy: 0.8698\n",
      "Epoch 67/100\n",
      "99/99 [==============================] - 0s 1ms/step - loss: 0.3084 - accuracy: 0.8696\n",
      "Epoch 68/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3085 - accuracy: 0.8696\n",
      "Epoch 69/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3084 - accuracy: 0.8700\n",
      "Epoch 70/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3084 - accuracy: 0.8696\n",
      "Epoch 71/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3083 - accuracy: 0.8697\n",
      "Epoch 72/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3083 - accuracy: 0.8695\n",
      "Epoch 73/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3082 - accuracy: 0.8696\n",
      "Epoch 74/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3082 - accuracy: 0.8695\n",
      "Epoch 75/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3082 - accuracy: 0.8695\n",
      "Epoch 76/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3082 - accuracy: 0.8696\n",
      "Epoch 77/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3081 - accuracy: 0.8694\n",
      "Epoch 78/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3082 - accuracy: 0.8694\n",
      "Epoch 79/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3081 - accuracy: 0.8695\n",
      "Epoch 80/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3080 - accuracy: 0.8696\n",
      "Epoch 81/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3080 - accuracy: 0.8694\n",
      "Epoch 82/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3080 - accuracy: 0.8695\n",
      "Epoch 83/100\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3080 - accuracy: 0.8695\n",
      "Epoch 84/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3079 - accuracy: 0.8694\n",
      "Epoch 85/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3080 - accuracy: 0.8693\n",
      "Epoch 86/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3079 - accuracy: 0.8695\n",
      "Epoch 87/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3079 - accuracy: 0.8694\n",
      "Epoch 88/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3078 - accuracy: 0.8694\n",
      "Epoch 89/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3077 - accuracy: 0.8693\n",
      "Epoch 90/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3078 - accuracy: 0.8694\n",
      "Epoch 91/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3078 - accuracy: 0.8691\n",
      "Epoch 92/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3076 - accuracy: 0.8694\n",
      "Epoch 93/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3077 - accuracy: 0.8690\n",
      "Epoch 94/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3076 - accuracy: 0.8689\n",
      "Epoch 95/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3076 - accuracy: 0.8694\n",
      "Epoch 96/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3075 - accuracy: 0.8694\n",
      "Epoch 97/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3074 - accuracy: 0.8694\n",
      "Epoch 98/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3073 - accuracy: 0.8692\n",
      "Epoch 99/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3073 - accuracy: 0.8694\n",
      "Epoch 100/100\n",
      "99/99 [==============================] - 0s 2ms/step - loss: 0.3073 - accuracy: 0.8693\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.History at 0x1f7954b6ca0>"
      ]
     },
     "execution_count": 58,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Fitting the data\n",
    "nn.fit(train_X_pca, train_y_pca, batch_size=1000, epochs=100)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "id": "25c6158d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3089/3089 [==============================] - 4s 1ms/step - loss: 0.3072 - accuracy: 0.8692\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[0.30718597769737244, 0.8691930770874023]"
      ]
     },
     "execution_count": 59,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Evaluate the model on the training set\n",
    "nn.evaluate(train_X_pca, train_y_pca)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "id": "de112b3d",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "813/813 [==============================] - 1s 1ms/step\n",
      "[[17432  1281]\n",
      " [ 2171  5129]]\n"
     ]
    }
   ],
   "source": [
    "print(confusion_matrix(test_y_pca, nn.predict(test_X_pca) > 0.5))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "id": "af2d53ab",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "813/813 [==============================] - 1s 1ms/step\n",
      "0.7482129832239242\n"
     ]
    }
   ],
   "source": [
    "print(f1_score(test_y_pca, nn.predict(test_X_pca) > 0.5))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "id": "1c4af881",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3089/3089 [==============================] - 3s 1ms/step\n",
      "3089/3089 [==============================] - 3s 1ms/step\n"
     ]
    }
   ],
   "source": [
    "# Predict on training set\n",
    "train_y_pred_nn = nn.predict(train_X_pca)\n",
    "#tr_y_prob_nnc = nnc.predict_proba(tr_X_pca)\n",
    "train_y_prob_nn =(nn.predict(train_X_pca) > 0.5).astype(\"int32\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "id": "c14a2856",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "813/813 [==============================] - 1s 1ms/step\n",
      "813/813 [==============================] - 1s 1ms/step\n"
     ]
    }
   ],
   "source": [
    "# Predict on test set\n",
    "test_y_pred_nn = nn.predict(test_X_pca)\n",
    "#te_y_prob_nnc = nnc.predict_proba(te_X_pca)\n",
    "test_y_prob_nn =(nn.predict(test_X_pca) > 0.5).astype(\"int32\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "id": "daa65baa",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ROC AUC - Train\n",
      "0.8022435463870788\n",
      "ROC AUC - Test\n",
      "0.8170738275127758\n"
     ]
    }
   ],
   "source": [
    "# ROC AUC Scores\n",
    "print('ROC AUC - Train')\n",
    "print(roc_auc_score(train_y_removed_outliers, train_y_prob_nn))\n",
    "print('ROC AUC - Test')\n",
    "print(roc_auc_score(test_y, test_y_prob_nn))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "id": "c3fd8810",
   "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": [
    "# ROC AUC curves\n",
    "fpr_nn_tr, tpr_nn_tr, _ = roc_curve(train_y_removed_outliers, train_y_prob_nn)\n",
    "fpr_nn_te, tpr_nn_te, _ = roc_curve(test_y, test_y_prob_nn)\n",
    "\n",
    "vz.plot_roc_curve(fpr_nn_tr, tpr_nn_tr, \"Training\")\n",
    "vz.plot_roc_curve(fpr_nn_te, tpr_nn_te, \"Test\")\n",
    "\n",
    "plt.title('Neural Network Classifier', fontsize=16)\n",
    "plt.xlabel('False Positive Rate', fontsize=12)\n",
    "plt.ylabel('True Positive Rate', fontsize=12)\n",
    "plt.legend(loc=\"lower right\", fontsize=12)\n",
    "plt.show()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.9.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}