Final Machine Learning Code.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "1931b3c6",
   "metadata": {
    "heading_collapsed": true
   },
   "source": [
    "# Options Pricing Group Project (Gabe Owens, Yuan-Cheng Tsai)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a69a9603",
   "metadata": {},
   "source": [
    "# Initial Data Quality Checks/Cleaning/Standardizing  "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1b47cfd7",
   "metadata": {},
   "source": [
    "## Load DataSets/Dummy Variable Conversion "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "acd6a0f7",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Load all datasets \n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import warnings\n",
    "warnings.filterwarnings('ignore')\n",
    "group_pred = pd.read_csv(\"group_0_prediction.csv\")\n",
    "option_test = pd.read_csv('option_test_wolabel.csv')\n",
    "option_train = pd.read_csv(\"option_train.csv\")\n",
    "#Convert 'BS' to a dummy variable (0 if Under, 1 if Over)\n",
    "option_train['BS']=option_train['BS'].apply(lambda x: 1 if x == 'Over' else 0)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a3f4b340",
   "metadata": {},
   "source": [
    "## Display All DataSets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "d536ae03",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style  type=\"text/css\" >\n",
       "</style><table id=\"T_91a6c_\" style='display:inline'><caption>Group Predictions</caption><thead>    <tr>        <th class=\"blank level0\" ></th>        <th class=\"col_heading level0 col0\" >Value</th>        <th class=\"col_heading level0 col1\" >BS</th>    </tr></thead><tbody>\n",
       "                <tr>\n",
       "                        <th id=\"T_91a6c_level0_row0\" class=\"row_heading level0 row0\" >0</th>\n",
       "                        <td id=\"T_91a6c_row0_col0\" class=\"data row0 col0\" >1.000000</td>\n",
       "                        <td id=\"T_91a6c_row0_col1\" class=\"data row0 col1\" >0</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_91a6c_level0_row1\" class=\"row_heading level0 row1\" >1</th>\n",
       "                        <td id=\"T_91a6c_row1_col0\" class=\"data row1 col0\" >2.000000</td>\n",
       "                        <td id=\"T_91a6c_row1_col1\" class=\"data row1 col1\" >0</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_91a6c_level0_row2\" class=\"row_heading level0 row2\" >2</th>\n",
       "                        <td id=\"T_91a6c_row2_col0\" class=\"data row2 col0\" >1.000000</td>\n",
       "                        <td id=\"T_91a6c_row2_col1\" class=\"data row2 col1\" >0</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_91a6c_level0_row3\" class=\"row_heading level0 row3\" >3</th>\n",
       "                        <td id=\"T_91a6c_row3_col0\" class=\"data row3 col0\" >1.000000</td>\n",
       "                        <td id=\"T_91a6c_row3_col1\" class=\"data row3 col1\" >0</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_91a6c_level0_row4\" class=\"row_heading level0 row4\" >4</th>\n",
       "                        <td id=\"T_91a6c_row4_col0\" class=\"data row4 col0\" >3.000000</td>\n",
       "                        <td id=\"T_91a6c_row4_col1\" class=\"data row4 col1\" >0</td>\n",
       "            </tr>\n",
       "    </tbody></table>   <style  type=\"text/css\" >\n",
       "</style><table id=\"T_a113d_\" style='display:inline'><caption>Test Data</caption><thead>    <tr>        <th class=\"blank level0\" ></th>        <th class=\"col_heading level0 col0\" >S</th>        <th class=\"col_heading level0 col1\" >K</th>        <th class=\"col_heading level0 col2\" >tau</th>        <th class=\"col_heading level0 col3\" >r</th>    </tr></thead><tbody>\n",
       "                <tr>\n",
       "                        <th id=\"T_a113d_level0_row0\" class=\"row_heading level0 row0\" >0</th>\n",
       "                        <td id=\"T_a113d_row0_col0\" class=\"data row0 col0\" >431.618600</td>\n",
       "                        <td id=\"T_a113d_row0_col1\" class=\"data row0 col1\" >460</td>\n",
       "                        <td id=\"T_a113d_row0_col2\" class=\"data row0 col2\" >0.293651</td>\n",
       "                        <td id=\"T_a113d_row0_col3\" class=\"data row0 col3\" >0.031470</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_a113d_level0_row1\" class=\"row_heading level0 row1\" >1</th>\n",
       "                        <td id=\"T_a113d_row1_col0\" class=\"data row1 col0\" >432.633296</td>\n",
       "                        <td id=\"T_a113d_row1_col1\" class=\"data row1 col1\" >420</td>\n",
       "                        <td id=\"T_a113d_row1_col2\" class=\"data row1 col2\" >0.182540</td>\n",
       "                        <td id=\"T_a113d_row1_col3\" class=\"data row1 col3\" >0.031470</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_a113d_level0_row2\" class=\"row_heading level0 row2\" >2</th>\n",
       "                        <td id=\"T_a113d_row2_col0\" class=\"data row2 col0\" >432.633296</td>\n",
       "                        <td id=\"T_a113d_row2_col1\" class=\"data row2 col1\" >430</td>\n",
       "                        <td id=\"T_a113d_row2_col2\" class=\"data row2 col2\" >0.182540</td>\n",
       "                        <td id=\"T_a113d_row2_col3\" class=\"data row2 col3\" >0.031470</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_a113d_level0_row3\" class=\"row_heading level0 row3\" >3</th>\n",
       "                        <td id=\"T_a113d_row3_col0\" class=\"data row3 col0\" >431.618600</td>\n",
       "                        <td id=\"T_a113d_row3_col1\" class=\"data row3 col1\" >415</td>\n",
       "                        <td id=\"T_a113d_row3_col2\" class=\"data row3 col2\" >0.293651</td>\n",
       "                        <td id=\"T_a113d_row3_col3\" class=\"data row3 col3\" >0.031470</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_a113d_level0_row4\" class=\"row_heading level0 row4\" >4</th>\n",
       "                        <td id=\"T_a113d_row4_col0\" class=\"data row4 col0\" >434.772855</td>\n",
       "                        <td id=\"T_a113d_row4_col1\" class=\"data row4 col1\" >420</td>\n",
       "                        <td id=\"T_a113d_row4_col2\" class=\"data row4 col2\" >0.043651</td>\n",
       "                        <td id=\"T_a113d_row4_col3\" class=\"data row4 col3\" >0.031470</td>\n",
       "            </tr>\n",
       "    </tbody></table>   <style  type=\"text/css\" >\n",
       "</style><table id=\"T_5e434_\" style='display:inline'><caption>Training Data</caption><thead>    <tr>        <th class=\"blank level0\" ></th>        <th class=\"col_heading level0 col0\" >Value</th>        <th class=\"col_heading level0 col1\" >S</th>        <th class=\"col_heading level0 col2\" >K</th>        <th class=\"col_heading level0 col3\" >tau</th>        <th class=\"col_heading level0 col4\" >r</th>        <th class=\"col_heading level0 col5\" >BS</th>    </tr></thead><tbody>\n",
       "                <tr>\n",
       "                        <th id=\"T_5e434_level0_row0\" class=\"row_heading level0 row0\" >0</th>\n",
       "                        <td id=\"T_5e434_row0_col0\" class=\"data row0 col0\" >21.670404</td>\n",
       "                        <td id=\"T_5e434_row0_col1\" class=\"data row0 col1\" >431.623898</td>\n",
       "                        <td id=\"T_5e434_row0_col2\" class=\"data row0 col2\" >420.000000</td>\n",
       "                        <td id=\"T_5e434_row0_col3\" class=\"data row0 col3\" >0.341270</td>\n",
       "                        <td id=\"T_5e434_row0_col4\" class=\"data row0 col4\" >0.030130</td>\n",
       "                        <td id=\"T_5e434_row0_col5\" class=\"data row0 col5\" >0</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_5e434_level0_row1\" class=\"row_heading level0 row1\" >1</th>\n",
       "                        <td id=\"T_5e434_row1_col0\" class=\"data row1 col0\" >0.125000</td>\n",
       "                        <td id=\"T_5e434_row1_col1\" class=\"data row1 col1\" >427.015526</td>\n",
       "                        <td id=\"T_5e434_row1_col2\" class=\"data row1 col2\" >465.000000</td>\n",
       "                        <td id=\"T_5e434_row1_col3\" class=\"data row1 col3\" >0.166667</td>\n",
       "                        <td id=\"T_5e434_row1_col4\" class=\"data row1 col4\" >0.031260</td>\n",
       "                        <td id=\"T_5e434_row1_col5\" class=\"data row1 col5\" >1</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_5e434_level0_row2\" class=\"row_heading level0 row2\" >2</th>\n",
       "                        <td id=\"T_5e434_row2_col0\" class=\"data row2 col0\" >20.691244</td>\n",
       "                        <td id=\"T_5e434_row2_col1\" class=\"data row2 col1\" >427.762336</td>\n",
       "                        <td id=\"T_5e434_row2_col2\" class=\"data row2 col2\" >415.000000</td>\n",
       "                        <td id=\"T_5e434_row2_col3\" class=\"data row2 col3\" >0.265873</td>\n",
       "                        <td id=\"T_5e434_row2_col4\" class=\"data row2 col4\" >0.031160</td>\n",
       "                        <td id=\"T_5e434_row2_col5\" class=\"data row2 col5\" >0</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_5e434_level0_row3\" class=\"row_heading level0 row3\" >3</th>\n",
       "                        <td id=\"T_5e434_row3_col0\" class=\"data row3 col0\" >1.035002</td>\n",
       "                        <td id=\"T_5e434_row3_col1\" class=\"data row3 col1\" >451.711658</td>\n",
       "                        <td id=\"T_5e434_row3_col2\" class=\"data row3 col2\" >460.000000</td>\n",
       "                        <td id=\"T_5e434_row3_col3\" class=\"data row3 col3\" >0.063492</td>\n",
       "                        <td id=\"T_5e434_row3_col4\" class=\"data row3 col4\" >0.029720</td>\n",
       "                        <td id=\"T_5e434_row3_col5\" class=\"data row3 col5\" >1</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_5e434_level0_row4\" class=\"row_heading level0 row4\" >4</th>\n",
       "                        <td id=\"T_5e434_row4_col0\" class=\"data row4 col0\" >39.553020</td>\n",
       "                        <td id=\"T_5e434_row4_col1\" class=\"data row4 col1\" >446.718974</td>\n",
       "                        <td id=\"T_5e434_row4_col2\" class=\"data row4 col2\" >410.000000</td>\n",
       "                        <td id=\"T_5e434_row4_col3\" class=\"data row4 col3\" >0.166667</td>\n",
       "                        <td id=\"T_5e434_row4_col4\" class=\"data row4 col4\" >0.029620</td>\n",
       "                        <td id=\"T_5e434_row4_col5\" class=\"data row4 col5\" >0</td>\n",
       "            </tr>\n",
       "    </tbody></table>   "
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Display All DataSets\n",
    "from IPython.core.display import display, HTML\n",
    "def display_side_by_side(dfs:list, captions:list):\n",
    "    output = \"\"\n",
    "    combined = dict(zip(captions, dfs))\n",
    "    for caption, df in combined.items():\n",
    "        output += df.style.set_table_attributes(\"style='display:inline'\").set_caption(caption)._repr_html_()\n",
    "        output += \"\\xa0\\xa0\\xa0\"\n",
    "    display(HTML(output))\n",
    "display_side_by_side([group_pred.head(), option_test.head(), option_train.head()], ['Group Predictions', 'Test Data', \n",
    "                                                                                    'Training Data'])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "04751874",
   "metadata": {},
   "source": [
    "## Quality Checks on Training Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "aeb049a4",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Info:\n",
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 1680 entries, 0 to 1679\n",
      "Data columns (total 6 columns):\n",
      " #   Column  Non-Null Count  Dtype  \n",
      "---  ------  --------------  -----  \n",
      " 0   Value   1679 non-null   float64\n",
      " 1   S       1679 non-null   float64\n",
      " 2   K       1678 non-null   float64\n",
      " 3   tau     1679 non-null   float64\n",
      " 4   r       1680 non-null   float64\n",
      " 5   BS      1680 non-null   int64  \n",
      "dtypes: float64(5), int64(1)\n",
      "memory usage: 78.9 KB\n",
      "None\n",
      "\n",
      "Descriptive Stats:\n",
      "             Value            S            K          tau            r  \\\n",
      "count  1679.000000  1679.000000  1678.000000  1679.000000  1680.000000   \n",
      "mean     15.072013   440.639445   438.241955     0.437519     0.030235   \n",
      "std      14.036492    13.130436    23.408989     7.057555     0.000557   \n",
      "min       0.125000     0.000000   375.000000     0.003968     0.029510   \n",
      "25%       2.255002   433.863864   420.000000     0.119048     0.029820   \n",
      "50%      11.220588   442.525366   440.000000     0.202381     0.030130   \n",
      "75%      25.745904   447.320414   455.000000     0.285714     0.030540   \n",
      "max      60.149367   455.880619   500.000000   250.000000     0.031880   \n",
      "\n",
      "                BS  \n",
      "count  1680.000000  \n",
      "mean      0.436905  \n",
      "std       0.496151  \n",
      "min       0.000000  \n",
      "25%       0.000000  \n",
      "50%       0.000000  \n",
      "75%       1.000000  \n",
      "max       1.000000  \n",
      "\n",
      "Null Value Count:\n",
      "Value    1\n",
      "S        1\n",
      "K        2\n",
      "tau      1\n",
      "r        0\n",
      "BS       0\n",
      "dtype: int64\n",
      "\n",
      "Duplicate Value Count:\n",
      "0\n",
      "\n",
      "Shape:\n",
      "(1680, 6)\n"
     ]
    }
   ],
   "source": [
    "#Quality Checks on Training Data\n",
    "print('Info:', sep='\\n')\n",
    "print(option_train.info())\n",
    "print()\n",
    "print('Descriptive Stats:', sep='\\n')\n",
    "print(option_train.describe())\n",
    "print()\n",
    "print('Null Value Count:', option_train.isnull().sum(), sep='\\n')\n",
    "print()\n",
    "print('Duplicate Value Count:')\n",
    "print(option_train.duplicated().sum())\n",
    "print()\n",
    "print('Shape:', option_train.shape, sep='\\n')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "162286b4",
   "metadata": {},
   "source": [
    "## Drop Null Values in Training Data "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "b5bdd55c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Null Value Count:\n",
      "Value    0\n",
      "S        0\n",
      "K        0\n",
      "tau      0\n",
      "r        0\n",
      "BS       0\n",
      "dtype: int64\n",
      "\n",
      "New Shape:\n",
      "(1678, 6)\n"
     ]
    }
   ],
   "source": [
    "#Drop Null Values in Training Data \n",
    "option_train = option_train.dropna(axis=0, how='any', thresh=None, subset=None, inplace=False)\n",
    "print('Null Value Count:', option_train.isnull().sum(), sep='\\n')\n",
    "print()\n",
    "print('New Shape:', option_train.shape, sep='\\n')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "53702327",
   "metadata": {},
   "source": [
    "## Outlier Visualization (Excluding 'BS')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "a302335f",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Original Skews:\n",
      "K skew: 0.09760174103507052\n",
      "S skew: -22.577206096818195\n",
      "Value skew: 0.7349212221963913\n",
      "r skew: 1.2358783187225972\n",
      "tau skew: 31.607019444736878\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x1080 with 5 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Outlier Visualization (Excluding 'BS') \n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "from matplotlib.pyplot import figure\n",
    "figure(figsize=(15, 15))\n",
    "count=1\n",
    "for i in option_train.columns.difference(['BS']):\n",
    "    plt.subplot(2,5,count)\n",
    "    sns.boxplot(data=option_train[i]).set_title(i)\n",
    "    count+=1\n",
    "plt.subplots_adjust(hspace=2, wspace=.7)\n",
    "print('Original Skews:', sep='\\n')\n",
    "for i in option_train.columns.difference(['BS']):\n",
    "    print(i,'skew:',(option_train[i].skew()))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "273c2b9f",
   "metadata": {},
   "source": [
    "## Find Extreme Values and Remove Them "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "d9d69c16",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "K :\n",
      "Empty DataFrame\n",
      "Columns: [Value, S, K, tau, r, BS]\n",
      "Index: []\n",
      "S :\n",
      "     Value    S      K       tau        r  BS\n",
      "879  4.125  0.0  455.0  0.170635  0.03003   1\n",
      "Value :\n",
      "Empty DataFrame\n",
      "Columns: [Value, S, K, tau, r, BS]\n",
      "Index: []\n",
      "r :\n",
      "Empty DataFrame\n",
      "Columns: [Value, S, K, tau, r, BS]\n",
      "Index: []\n",
      "tau :\n",
      "       Value           S      K    tau        r  BS\n",
      "12  2.315001  448.688109  470.0  250.0  0.03013   1\n",
      "33  2.565000  445.042240  455.0  146.0  0.03003   1\n"
     ]
    }
   ],
   "source": [
    "#Find Extreme Values and Remove Them \n",
    "col = option_train.columns.difference(['BS'])\n",
    "for i in col:\n",
    "    Q1 = option_train[i].quantile(0.25)\n",
    "    Q3 = option_train[i].quantile(0.75)\n",
    "    IQR = Q3 - Q1\n",
    "    print(i,':')\n",
    "    print(option_train.loc[(option_train[i] < (Q1 - 3 * IQR)) | (option_train[i] > (Q3 + 3 * IQR))])\n",
    "    condition = ~((option_train[i] < (Q1 - 3 * IQR)) | (option_train[i] > (Q3 + 3 * IQR)))\n",
    "    option_train = option_train[condition]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4d61b565",
   "metadata": {},
   "source": [
    "## Final Skew Values and Dataset Shape "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "a6d2b864",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Final Skews:\n",
      "K skew: 0.10083342984947408\n",
      "S skew: -0.22652462714817947\n",
      "Value skew: 0.7322637711951554\n",
      "r skew: 1.2335201609700461\n",
      "tau skew: -0.08777009408890274\n",
      "\n",
      "Final Shape:\n",
      "(1675, 6)\n"
     ]
    }
   ],
   "source": [
    "#Final Skew Values and Dataset Shape \n",
    "print('Final Skews:', sep='\\n')\n",
    "for i in option_train.columns.difference(['BS']):\n",
    "    print(i,'skew:',(option_train[i].skew()))\n",
    "print()\n",
    "print('Final Shape:', option_train.shape, sep='\\n')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "17f1184e",
   "metadata": {},
   "source": [
    "## Train/Test Split on Dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "9d6fdce4",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "X_train shape: (1172, 4)\n",
      "y_train shape: (1172, 1)\n",
      "X_test shape: (503, 4)\n",
      "y_test shape: (503, 1)\n"
     ]
    }
   ],
   "source": [
    "#Train/Test Split on Dataset\n",
    "from sklearn.model_selection import train_test_split\n",
    "X=option_train[['K','S','r','tau']]\n",
    "y=option_train[['Value']]\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3,random_state=1)\n",
    "option_train_train = pd.DataFrame(X_train, columns = ['K', 'S', 'r', 'tau'])\n",
    "y_train = pd.DataFrame(y_train, columns = ['Value'])\n",
    "option_train_train['Value']= y_train\n",
    "option_train_test = pd.DataFrame(X_test, columns = ['K', 'S', 'r', 'tau'])\n",
    "y_test = pd.DataFrame(y_test, columns = ['Value'])\n",
    "option_train_test['Value']= y_test\n",
    "print('X_train shape:', X_train.shape)\n",
    "print('y_train shape:', y_train.shape)\n",
    "print('X_test shape:', X_test.shape)\n",
    "print('y_test shape:', y_test.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d415881d",
   "metadata": {},
   "source": [
    "## Normalizing data using MinMaxScaler()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a2964c4d",
   "metadata": {},
   "source": [
    "- We added a very small number (.0000000000001) to both the train/test split data to allow for more transformations (ex: 1/x or log(x))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "7b93dac0",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "         K      S      r    tau   Value\n",
      "64    0.60  0.674  0.304  0.908  11.250\n",
      "481   0.64  0.235  0.827  0.459   0.970\n",
      "282   0.40  0.719  0.046  0.490  26.666\n",
      "1611  0.68  0.453  0.304  0.735   1.755\n",
      "1018  0.36  0.258  0.177  0.020  14.120\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>K</th>\n",
       "      <th>S</th>\n",
       "      <th>r</th>\n",
       "      <th>tau</th>\n",
       "      <th>Value</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>1172.000</td>\n",
       "      <td>1172.000</td>\n",
       "      <td>1172.000</td>\n",
       "      <td>1172.000</td>\n",
       "      <td>1172.000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>0.510</td>\n",
       "      <td>0.512</td>\n",
       "      <td>0.304</td>\n",
       "      <td>0.510</td>\n",
       "      <td>14.750</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>0.185</td>\n",
       "      <td>0.248</td>\n",
       "      <td>0.234</td>\n",
       "      <td>0.254</td>\n",
       "      <td>13.915</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.125</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>0.360</td>\n",
       "      <td>0.277</td>\n",
       "      <td>0.131</td>\n",
       "      <td>0.306</td>\n",
       "      <td>2.125</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>0.520</td>\n",
       "      <td>0.590</td>\n",
       "      <td>0.262</td>\n",
       "      <td>0.510</td>\n",
       "      <td>10.666</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>0.640</td>\n",
       "      <td>0.719</td>\n",
       "      <td>0.435</td>\n",
       "      <td>0.724</td>\n",
       "      <td>24.918</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000</td>\n",
       "      <td>60.149</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              K         S         r       tau     Value\n",
       "count  1172.000  1172.000  1172.000  1172.000  1172.000\n",
       "mean      0.510     0.512     0.304     0.510    14.750\n",
       "std       0.185     0.248     0.234     0.254    13.915\n",
       "min       0.000     0.000     0.000     0.000     0.125\n",
       "25%       0.360     0.277     0.131     0.306     2.125\n",
       "50%       0.520     0.590     0.262     0.510    10.666\n",
       "75%       0.640     0.719     0.435     0.724    24.918\n",
       "max       1.000     1.000     1.000     1.000    60.149"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Normalizing data using MinMaxScaler \n",
    "from sklearn.preprocessing import MinMaxScaler\n",
    "num_cols = option_train_train.columns.difference(['Value'])\n",
    "for i in num_cols:\n",
    "    scale = MinMaxScaler().fit(option_train_train[[i]])\n",
    "    option_train_train[i] = scale.transform(option_train_train[[i]])\n",
    "num_cols = option_train_test.columns.difference(['Value'])\n",
    "for i in num_cols:\n",
    "    scale = MinMaxScaler().fit(option_train_test[[i]])\n",
    "    option_train_test[i] = scale.transform(option_train_test[[i]])\n",
    "option_train_train = option_train_train + .0000000000001\n",
    "option_train_test = option_train_test +   .0000000000001\n",
    "y_test = option_train_test['Value']\n",
    "option_train_test = option_train_test.drop(columns=['Value'])\n",
    "print(round(option_train_train.head(),3))\n",
    "round(option_train_train.describe(),3)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c4fe5639",
   "metadata": {},
   "source": [
    "# Data Analysis "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f37b2548",
   "metadata": {
    "heading_collapsed": true
   },
   "source": [
    "## Regression (Value)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9a2b7bea",
   "metadata": {
    "hidden": true
   },
   "source": [
    "### GridSearchCV"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c00cbc79",
   "metadata": {
    "hidden": true
   },
   "source": [
    "For all regressors, we will use GridSearchCV to find the best combination of tuned hyperparameters.\n",
    "\n",
    "We will then return the average cross-validated score of the specific Regressor."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 485,
   "id": "9412b780",
   "metadata": {
    "hidden": true
   },
   "outputs": [],
   "source": [
    "from sklearn.model_selection import GridSearchCV"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2e08bd79",
   "metadata": {
    "hidden": true
   },
   "source": [
    "### Lasso Regression "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 480,
   "id": "dcb76a6e",
   "metadata": {
    "hidden": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best Parameters:  {'alpha': 0.005}\n",
      "RSquared:  0.9094789702205934\n"
     ]
    }
   ],
   "source": [
    "from sklearn import linear_model\n",
    "model = linear_model.Lasso()\n",
    "forest_params = {'alpha': [5, 0.5, 0.05, 0.005, 0.0005, 1, 0.1, 0.01, 0.001, 0.0001]}\n",
    "clf = GridSearchCV(model, forest_params, cv = 10, scoring='r2')\n",
    "clf.fit(X_train, y_train)\n",
    "print('Best Parameters: ', clf.best_params_)\n",
    "print('RSquared: ', clf.best_score_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 481,
   "id": "de796ac8",
   "metadata": {
    "hidden": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mean Cross Val (R2) Score 0.9105461522997921\n"
     ]
    }
   ],
   "source": [
    "from sklearn.model_selection import cross_val_score\n",
    "from sklearn.model_selection import RepeatedKFold\n",
    "from sklearn import linear_model\n",
    "X = option_train[feature_cols]\n",
    "y = option_train[\"Value\"]\n",
    "model = linear_model.Lasso(alpha=0.005).fit(X_train, y_train)\n",
    "y_pred = model.predict(X_test) \n",
    "cv = RepeatedKFold(n_splits=5, n_repeats= 100, random_state=1)\n",
    "print(\"Mean Cross Val (R2) Score\", np.mean(cross_val_score(model, X, y, scoring='r2', cv=cv, n_jobs=-1)))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "553c6437",
   "metadata": {
    "hidden": true
   },
   "source": [
    "### Ridge Regression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 491,
   "id": "1a45a8b3",
   "metadata": {
    "hidden": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best Parameters:  {'alpha': 1, 'solver': 'auto'}\n",
      "RSquared:  0.9091558617141151\n"
     ]
    }
   ],
   "source": [
    "from sklearn.linear_model import Ridge\n",
    "model = Ridge()\n",
    "forest_params = {'solver': ['auto', 'svd', 'cholesky', 'lsqr', 'sparse_cg', 'sag', 'saga', 'lbfgs'], 'alpha':[1]}\n",
    "clf = GridSearchCV(model, forest_params, cv = 10, scoring='r2')\n",
    "clf.fit(X_train, y_train)\n",
    "print('Best Parameters: ', clf.best_params_)\n",
    "print('RSquared: ', clf.best_score_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 492,
   "id": "05a6fb64",
   "metadata": {
    "hidden": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mean Cross Val (R2) Score 0.910311667096764\n"
     ]
    }
   ],
   "source": [
    "from sklearn.linear_model import Ridge\n",
    "model = Ridge(alpha=1.0, solver = 'auto').fit(X_train, y_train)\n",
    "y_pred = model.predict(X_test) \n",
    "cv = RepeatedKFold(n_splits=5, n_repeats= 100, random_state=1)\n",
    "print(\"Mean Cross Val (R2) Score\", np.mean(cross_val_score(model, X, y, scoring='r2', cv=cv, n_jobs=-1)))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "32479c31",
   "metadata": {
    "hidden": true
   },
   "source": [
    "### Decision Tree Regressor "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 494,
   "id": "15b1f1b9",
   "metadata": {
    "hidden": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best Parameters:  {'max_depth': 11, 'splitter': 'best'}\n",
      "RSquared:  0.9888647563862423\n"
     ]
    }
   ],
   "source": [
    "from sklearn.tree import DecisionTreeRegressor\n",
    "model = DecisionTreeRegressor()\n",
    "forest_params = {\"splitter\":[\"best\",\"random\"],\n",
    "            \"max_depth\" : [1,3,5,7,9,11,12]}\n",
    "clf = GridSearchCV(model, forest_params, cv = 10, scoring='r2')\n",
    "clf.fit(X_train, y_train)\n",
    "print('Best Parameters: ', clf.best_params_)\n",
    "print('RSquared: ', clf.best_score_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 495,
   "id": "221a5f35",
   "metadata": {
    "hidden": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mean Cross Val (R2) Score 0.990953127367164\n"
     ]
    }
   ],
   "source": [
    "from sklearn.tree import DecisionTreeRegressor\n",
    "model = DecisionTreeRegressor(max_depth=11, splitter='best', random_state=1).fit(X_train, y_train)\n",
    "y_pred = model.predict(X_test) \n",
    "cv = RepeatedKFold(n_splits=5, n_repeats= 100, random_state=1)\n",
    "print(\"Mean Cross Val (R2) Score\", np.mean(cross_val_score(model, X, y, scoring='r2', cv=cv, n_jobs=-1)))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "78862167",
   "metadata": {
    "hidden": true
   },
   "source": [
    "### ElasticNet Regression "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 499,
   "id": "1d58ab94",
   "metadata": {
    "hidden": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best Parameters:  {'alpha': 0.0001, 'l1_ratio': 0.0, 'max_iter': 5}\n",
      "RSquared:  0.9094837210433822\n"
     ]
    }
   ],
   "source": [
    "from sklearn.linear_model import ElasticNet\n",
    "model = ElasticNet()\n",
    "forest_params = {\"max_iter\": [1, 5, 10],\n",
    "                      \"alpha\": [0.0001, 0.001, 0.01, 0.1, 1, 10, 100],\n",
    "                      \"l1_ratio\": np.arange(0.0, 1.0, 0.1)}\n",
    "clf = GridSearchCV(model, forest_params, cv = 10, scoring='r2')\n",
    "clf.fit(X_train, y_train)\n",
    "print('Best Parameters: ', clf.best_params_)\n",
    "print('RSquared: ', clf.best_score_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 500,
   "id": "f0de6af6",
   "metadata": {
    "hidden": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mean Cross Val (R2) Score 0.9105567386332925\n"
     ]
    }
   ],
   "source": [
    "from sklearn.linear_model import ElasticNet\n",
    "model = ElasticNet(alpha = 0.0001, l1_ratio = 0.0, max_iter =5, random_state=1).fit(X_train, y_train)\n",
    "y_pred = model.predict(X_test) \n",
    "cv = RepeatedKFold(n_splits=5, n_repeats= 100, random_state=1)\n",
    "print(\"Mean Cross Val (R2) Score\", np.mean(cross_val_score(model, X, y, scoring='r2', cv=cv, n_jobs=-1)))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1ace51a0",
   "metadata": {
    "hidden": true
   },
   "source": [
    "### Linear Regression "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 467,
   "id": "00932522",
   "metadata": {
    "hidden": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mean Cross Val (R2) Score 0.9107469509112591\n"
     ]
    }
   ],
   "source": [
    "from sklearn.linear_model import LinearRegression\n",
    "model = LinearRegression(n_jobs=-1).fit(X_train, y_train)\n",
    "y_pred = model.predict(X_test) \n",
    "cv = RepeatedKFold(n_splits=5, n_repeats= 100, random_state=1)\n",
    "print(\"Mean Cross Val (R2) Score\", np.mean(cross_val_score(model, X, y, scoring='r2', cv=cv, n_jobs=-1)))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e88a4862",
   "metadata": {
    "hidden": true
   },
   "source": [
    "### Attempting to Form the Best Linear Regression Equation:"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "90b15170",
   "metadata": {
    "hidden": true
   },
   "source": [
    "While we divided the samples into training and testing sets, we will be doing the regression model fit tests on our training data, and will find the equation with the highest cross-validated Adjusted R2. We will then check the performance of the final model on the unseen test data. This helps endure there was no overfitting. We are also doing it this way to avoid data leakage, where knowledge of the test set leaks into the dataset used to train the model. \n",
    "\n",
    "In addition, from now on, we are using Adjusted RSquared instead of R2 for training data because: \n",
    "- R^2 does not include a penalty for model complexity and will pick the model with the most variables while Adjusted RSquared does include a penalty for adding more predictors. \n",
    "- R2-value is a function of the mean square error (MSE). \n",
    "- R2-value increases only if MSE decreases. That is, the adjusted R2-value and MSE criteria always yield the same \"best\" models\n",
    "\n",
    "For the linear regression portion of this project, we will compare Adjusted RSquared values for different models using our training data, then test our final model on the test data. "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3b03f35a",
   "metadata": {
    "hidden": true
   },
   "source": [
    "### Initial Linear Regression Tests"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 391,
   "id": "78b5e8cf",
   "metadata": {
    "hidden": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Predictors: ('K',)   RSquared: 0.7860988875720014\n",
      "Predictors: ('S',)   RSquared: 0.005840156504013705\n",
      "Predictors: ('r',)   RSquared: 0.0014867135061327197\n",
      "Predictors: ('tau',)   RSquared: 0.0024604877873176667\n",
      "Predictors: ('K', 'S')   RSquared: 0.8650910086134567\n",
      "Predictors: ('K', 'r')   RSquared: 0.8000859342499749\n",
      "Predictors: ('K', 'tau')   RSquared: 0.8109362561744098\n",
      "Predictors: ('S', 'r')   RSquared: 0.012594582272823174\n",
      "Predictors: ('S', 'tau')   RSquared: 0.0098411623771536\n",
      "Predictors: ('r', 'tau')   RSquared: 0.0038882515000855378\n",
      "Predictors: ('K', 'S', 'r')   RSquared: 0.865103598800637\n",
      "Predictors: ('K', 'S', 'tau')   RSquared: 0.9127552344188956\n",
      "Predictors: ('K', 'r', 'tau')   RSquared: 0.8263809415834936\n",
      "Predictors: ('S', 'r', 'tau')   RSquared: 0.017360962869397034\n",
      "Predictors: ('K', 'S', 'r', 'tau')   RSquared: 0.9130010611706952\n",
      "\n",
      "The best regression equation has predictors: ('K', 'S', 'r', 'tau') with an RSquared of: 0.9130010611706952\n",
      "However, the most important predictors are K and S and tau with an RSquared of .91276.\n"
     ]
    }
   ],
   "source": [
    "#Initial Linear Regression Tests\n",
    "from sklearn.linear_model import LinearRegression\n",
    "from itertools import chain, combinations\n",
    "X = list(option_train_train.columns.difference(['Value','BS']))\n",
    "y = list(option_train_train[\"Value\"])\n",
    "def powerset(iterable): \n",
    "    s = list(iterable) \n",
    "    return chain.from_iterable(combinations(s, r) for r in range(len(s)+1)) \n",
    "R2 = {}\n",
    "model=LinearRegression()\n",
    "for subset in powerset(X):\n",
    "    if len(subset) > 0:\n",
    "        model.fit(option_train_train[list(subset)],y)\n",
    "        R2_value = model.score(option_train_train[list(subset)],y)\n",
    "        R2.update({subset: R2_value})\n",
    "        print('Predictors:', subset,'  RSquared:', R2_value)\n",
    "print()\n",
    "print('The best regression equation has predictors:', max(R2, key=R2.get) ,'with an RSquared of:', max(R2.values()))\n",
    "print('However, the most important predictors are K and S and tau with an RSquared of .91276.')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "25a65287",
   "metadata": {
    "hidden": true
   },
   "source": [
    "### Analysis "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e22ad21a",
   "metadata": {
    "hidden": true
   },
   "source": [
    "- We see that K is a necessity for a strong model fit, so we include it in every future model. Including S and tau provides stronger R2 values than when including r. Therefore, K+S+tau is the most necessary (as the jump from RSquared values of .91276 to .91300 is not large enough to justify adding the additional variable r. Next, we will test transformations on the predictors. "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4fe961fa",
   "metadata": {
    "hidden": true
   },
   "source": [
    "### Interaction Terms "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 291,
   "id": "cb31cb6f",
   "metadata": {
    "hidden": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Preditors: K + r + S + tau + r*K    Adj RSquared: 0.9134008657996627\n",
      "Preditors: K + r + S + tau + r*S    Adj RSquared: 0.9127797786198377\n",
      "Preditors: K + r + S + tau + r*tau    Adj RSquared: 0.9127825851191467\n",
      "Preditors: K + r + S + tau + S*K    Adj RSquared: 0.913500468539426\n",
      "Preditors: K + r + S + tau + S*tau    Adj RSquared: 0.9126364013079261\n",
      "Preditors: K + S + r + tau + K*tau   Adj RSquared: 0.9178016895756905\n",
      "Preditors: K + r + S + tau + K*S*r   Adj RSquared: 0.8697521324225653\n",
      "Preditors: K + r + S + tau + K*S*tau   Adj RSquared: 0.9193029021929917\n",
      "Preditors: K + r + S + tau + K*r*tau   Adj RSquared: 0.9196938924159708\n",
      "Preditors: K + r + S + tau + S*r*tau   Adj RSquared: 0.9127478695272224\n",
      "Preditors: K + r + S + tau + K*r*S*tau   Adj RSquared: 0.9222609306798204\n",
      "\n",
      "The Adj RSquared values for all interaction terms are very similar and equally beneficial\n",
      "For simplicity, and also to limit the total number of predictors, we will include the most important predictor as K*S, with Adj RSquared value: 0.9135\n"
     ]
    }
   ],
   "source": [
    "#Interaction Terms \n",
    "import statsmodels.formula.api as smf\n",
    "for i in list(option_train.columns.difference(['Value','BS','r'])):\n",
    "    result1 = smf.ols(f'Value ~ K + r + S + tau + r*{i}', data=option_train_train).fit()\n",
    "    print(f'Preditors: K + r + S + tau + r*{i} ', '  Adj RSquared:', result1.rsquared_adj)\n",
    "for i in list(option_train.columns.difference(['Value','BS','r','S'])):\n",
    "    result1 = smf.ols(f'Value ~ K + r + S + tau + S*{i}', data=option_train_train).fit()\n",
    "    print(f'Preditors: K + r + S + tau + S*{i} ', '  Adj RSquared:', result1.rsquared_adj)\n",
    "result1 = smf.ols(f'Value ~ K + S + r + tau + K * tau', data=option_train_train).fit()\n",
    "print(f'Preditors: K + S + r + tau + K*tau', '  Adj RSquared:', result1.rsquared_adj)\n",
    "result1 = smf.ols(f'Value ~ K*S*r', data=option_train_train).fit()\n",
    "print(f'Preditors: K + r + S + tau + K*S*r', '  Adj RSquared:', result1.rsquared_adj)\n",
    "result1 = smf.ols(f'Value ~ K + r + S + tau + K*S*tau', data=option_train_train).fit()\n",
    "print(f'Preditors: K + r + S + tau + K*S*tau', '  Adj RSquared:', result1.rsquared_adj)\n",
    "result1 = smf.ols(f'Value ~ K + r + S + tau + K*r*tau', data=option_train_train).fit()\n",
    "print(f'Preditors: K + r + S + tau + K*r*tau', '  Adj RSquared:', result1.rsquared_adj)\n",
    "result1 = smf.ols(f'Value ~ K + r + S + tau + S*r*tau', data=option_train_train).fit()\n",
    "print(f'Preditors: K + r + S + tau + S*r*tau', '  Adj RSquared:', result1.rsquared_adj)\n",
    "result1 = smf.ols(f'Value ~ K + r + S + tau + K*S*r*tau', data=option_train_train).fit()\n",
    "print(f'Preditors: K + r + S + tau + K*r*S*tau', '  Adj RSquared:', result1.rsquared_adj)\n",
    "print()\n",
    "print('The Adj RSquared values for all interaction terms are very similar and equally beneficial')\n",
    "print('For simplicity, and also to limit the total number of predictors, \\\n",
    "we will include the most important predictor as K*S, with Adj RSquared value: 0.9135')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2cbd30c2",
   "metadata": {
    "hidden": true
   },
   "source": [
    "### Exponential Transformations (Non-linear)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 292,
   "id": "531b115c",
   "metadata": {
    "hidden": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Preditors: K + r + S + tau + K^2    Adj RSquared: 0.9678956954046828\n",
      "Preditors: K + r + S + tau + S^2    Adj RSquared: 0.9130281249790448\n",
      "Preditors: K + r + S + tau + r^2    Adj RSquared: 0.9128321205704222\n",
      "Preditors: K + r + S + tau + tau^2    Adj RSquared: 0.9133403141562983\n",
      "Preditors: K + r + S + tau + K^2 + S^2   Adj RSquared: 0.9679111666334539\n",
      "Preditors: K + r + S + tau + K^2 + r^2   Adj RSquared: 0.9679360016967685\n",
      "Preditors: K + r + S + tau + K^2 + tau^2   Adj RSquared: 0.968116945456705\n",
      "Preditors: K + r + S + tau + S^2 + r^2   Adj RSquared: 0.9131357068520343\n",
      "Preditors: K + r + S + tau + S^2 + tau^2   Adj RSquared: 0.913683889730024\n",
      "Preditors: K + r + S + tau + r^2 + tau^2   Adj RSquared: 0.9134415817670417\n",
      "Preditors: K + r + S + tau + K^2 + S^2 + r^2   Adj RSquared: 0.9679554428085427\n",
      "Preditors: K + r + S + tau + K^2 + S^2 + tau^2   Adj RSquared: 0.9681366197319549\n",
      "Preditors: K + r + S + tau + K^2 + r^2 + tau^2   Adj RSquared: 0.9681670027742612\n",
      "Preditors: K + r + S + tau + S^2 + r^2 + tau^2   Adj RSquared: 0.9137643212946193\n",
      "Preditors: K + r + S + tau + K^2 + r^2 + S^2 + tau^2   Adj RSquared: 0.968191209186904\n",
      "\n",
      "We can see the most important predictor is K^2 from its impact on Adj RSquared values - every Adj R2 is >.96\n"
     ]
    }
   ],
   "source": [
    "#Non-linear Transformations (Exponential) \n",
    "for i in list(option_train.columns.difference(['Value','BS'])):\n",
    "    result1 = smf.ols(f'Value ~ K + r + S + tau + np.power({i}, 2)', data=option_train_train).fit()\n",
    "    print(f'Preditors: K + r + S + tau + {i}^2 ', '  Adj RSquared:', result1.rsquared_adj)\n",
    "for i in list(option_train.columns.difference(['Value','BS','K'])):\n",
    "    result1 = smf.ols(f'Value ~ K + r + S + tau + np.power(K, 2) + np.power({i}, 2)', data=option_train_train).fit()\n",
    "    print(f'Preditors: K + r + S + tau + K^2 + {i}^2 ', ' Adj RSquared:', result1.rsquared_adj)\n",
    "for i in list(option_train.columns.difference(['Value','BS','K','S'])):\n",
    "    result1 = smf.ols(f'Value ~ K + r + S + tau + np.power(S, 2) + np.power({i}, 2)', data=option_train_train).fit()\n",
    "    print(f'Preditors: K + r + S + tau + S^2 + {i}^2 ', ' Adj RSquared:', result1.rsquared_adj)\n",
    "result1 = smf.ols(f'Value ~ K + r + S + tau + np.power(r, 2) + np.power(tau, 2)', data=option_train_train).fit()\n",
    "print(f'Preditors: K + r + S + tau + r^2 + tau^2 ', ' Adj RSquared:', result1.rsquared_adj)\n",
    "result1 = smf.ols(f'Value ~ K + r + S + tau + np.power(K, 2) + np.power(S, 2) + np.power(r, 2)', \n",
    "                  data=option_train_train).fit()\n",
    "print(f'Preditors: K + r + S + tau + K^2 + S^2 + r^2 ', ' Adj RSquared:', result1.rsquared_adj)\n",
    "result1 = smf.ols(f'Value ~ K + r + S + tau + np.power(K, 2) + np.power(S, 2) + np.power(tau, 2)', \n",
    "                  data=option_train_train).fit()\n",
    "print(f'Preditors: K + r + S + tau + K^2 + S^2 + tau^2 ', ' Adj RSquared:', result1.rsquared_adj)\n",
    "result1 = smf.ols(f'Value ~ K + r + S + tau + np.power(K, 2) + np.power(r, 2) + np.power(tau, 2)', \n",
    "                  data=option_train_train).fit()\n",
    "print(f'Preditors: K + r + S + tau + K^2 + r^2 + tau^2 ', ' Adj RSquared:', result1.rsquared_adj)\n",
    "result1 = smf.ols(f'Value ~ K + r + S + tau + np.power(S, 2) + np.power(r, 2) + np.power(tau, 2)', \n",
    "                  data=option_train_train).fit()\n",
    "print(f'Preditors: K + r + S + tau + S^2 + r^2 + tau^2 ', ' Adj RSquared:', result1.rsquared_adj)\n",
    "result1 = smf.ols(f'Value ~ K + r + S + tau + np.power(K, 2) + np.power(r, 2) + np.power(S,2) + np.power(tau,2)', \n",
    "                  data=option_train_train).fit()\n",
    "print(f'Preditors: K + r + S + tau + K^2 + r^2 + S^2 + tau^2', '  Adj RSquared:', result1.rsquared_adj)\n",
    "print()\n",
    "print('We can see the most important predictor is K^2 from its impact on Adj RSquared values - every Adj R2 is >.96')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "071e5b36",
   "metadata": {
    "hidden": true
   },
   "source": [
    "### Log Transformations"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 293,
   "id": "b284bda3",
   "metadata": {
    "hidden": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Preditors: K + r + S + tau + log(K)   Adj RSquared: 0.915765397602837\n",
      "Preditors: K + r + S + tau + log(S)   Adj RSquared: 0.912977547812878\n",
      "Preditors: K + r + S + tau + log(r)   Adj RSquared: 0.9127602047405832\n",
      "Preditors: K + r + S + tau + log(tau)   Adj RSquared: 0.9130538673860299\n",
      "Preditors: K + r + S + tau + log(S) + log(r)   Adj RSquared: 0.9130200016740212\n",
      "Preditors: K + r + S + tau + log(S) + log(K)   Adj RSquared: 0.9157974674259968\n",
      "Preditors: K + r + S + tau + log(S) + log(tau)   Adj RSquared: 0.9133316663884898\n",
      "Preditors: K + r + S + tau + log(tau) + log(r)   Adj RSquared: 0.913205535861779\n",
      "Preditors: K + r + S + tau + log(tau) + log(K)   Adj RSquared: 0.9161248768528022\n",
      "Preditors: K + r + S + tau + log(K) + log(r)   Adj RSquared: 0.9157998058248551\n",
      "Preditors: K + r + S + tau + log(S) + log(r) + log(tau)   Adj RSquared: 0.9134637287056592\n",
      "Preditors: K + r + S + tau + log(S) + log(r) + log(tau) + log(K)   Adj RSquared: 0.9162700056806314\n",
      "\n",
      "The Adj RSquared values for all log terms are very similar and equally beneficial\n",
      "For simplicity are variety, we will include log(tau) with Adj R2 = .9131\n"
     ]
    }
   ],
   "source": [
    "#Log Transformations\n",
    "import warnings\n",
    "warnings.filterwarnings('ignore')\n",
    "result1 = smf.ols(f'Value ~ K + S + r + tau + np.log(K)', data=option_train_train).fit()\n",
    "print(f'Preditors: K + r + S + tau + log(K)', '  Adj RSquared:', result1.rsquared_adj)\n",
    "result1 = smf.ols(f'Value ~ K + S + r + tau + np.log(S)', data=option_train_train).fit()\n",
    "print(f'Preditors: K + r + S + tau + log(S)', '  Adj RSquared:', result1.rsquared_adj)\n",
    "result1 = smf.ols(f'Value ~ K + S + r + tau + np.log(r)', data=option_train_train).fit()\n",
    "print(f'Preditors: K + r + S + tau + log(r)', '  Adj RSquared:', result1.rsquared_adj)\n",
    "result1 = smf.ols(f'Value ~ K + S + r + tau + np.log(tau)', data=option_train_train).fit()\n",
    "print(f'Preditors: K + r + S + tau + log(tau)', '  Adj RSquared:', result1.rsquared_adj)\n",
    "result1 = smf.ols(f'Value ~ K + S + r + tau + np.log(r) + np.log(S)', data=option_train_train).fit()\n",
    "print(f'Preditors: K + r + S + tau + log(S) + log(r)', '  Adj RSquared:', result1.rsquared_adj)\n",
    "result1 = smf.ols(f'Value ~ K + S + r + tau + np.log(K) + np.log(S)', data=option_train_train).fit()\n",
    "print(f'Preditors: K + r + S + tau + log(S) + log(K)', '  Adj RSquared:', result1.rsquared_adj)\n",
    "result1 = smf.ols(f'Value ~ K + S + r + tau + np.log(tau) + np.log(S)', data=option_train_train).fit()\n",
    "print(f'Preditors: K + r + S + tau + log(S) + log(tau)', '  Adj RSquared:', result1.rsquared_adj)\n",
    "result1 = smf.ols(f'Value ~ K + S + r + tau + np.log(r) + np.log(tau)', data=option_train_train).fit()\n",
    "print(f'Preditors: K + r + S + tau + log(tau) + log(r)', '  Adj RSquared:', result1.rsquared_adj)\n",
    "result1 = smf.ols(f'Value ~ K + S + r + tau + np.log(K) + np.log(tau)', data=option_train_train).fit()\n",
    "print(f'Preditors: K + r + S + tau + log(tau) + log(K)', '  Adj RSquared:', result1.rsquared_adj)\n",
    "result1 = smf.ols(f'Value ~ K + S + r + tau + np.log(r) + np.log(K)', data=option_train_train).fit()\n",
    "print(f'Preditors: K + r + S + tau + log(K) + log(r)', '  Adj RSquared:', result1.rsquared_adj)\n",
    "result1 = smf.ols(f'Value ~ K + S + r + tau + np.log(S) + np.log(r) + np.log(tau)', data=option_train_train).fit()\n",
    "print(f'Preditors: K + r + S + tau + log(S) + log(r) + log(tau)', '  Adj RSquared:', result1.rsquared_adj)\n",
    "result1 = smf.ols(f'Value ~ K + S + r + tau + np.log(S) + np.log(r) + np.log(tau) + np.log(K)', data=option_train_train).fit()\n",
    "print(f'Preditors: K + r + S + tau + log(S) + log(r) + log(tau) + log(K)', '  Adj RSquared:', result1.rsquared_adj)\n",
    "print()\n",
    "print('The Adj RSquared values for all log terms are very similar and equally beneficial')\n",
    "print('For simplicity are variety, we will include log(tau) with Adj R2 = .9131')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "95c6d14e",
   "metadata": {
    "hidden": true
   },
   "source": [
    "### Exponential Transformations (e^x)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 294,
   "id": "e8526dc7",
   "metadata": {
    "hidden": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Preditors: K + r + S + tau + e^(K)   Adj RSquared: 0.9692024836776321\n",
      "Preditors: K + r + S + tau + e^(S)   Adj RSquared: 0.91296251479459\n",
      "Preditors: K + r + S + tau + e^(r)   Adj RSquared: 0.9128180156333\n",
      "Preditors: K + r + S + tau + e^(tau)   Adj RSquared: 0.9132579683371659\n",
      "Preditors: K + r + S + tau + e^(S) + e^(r)   Adj RSquared: 0.913081282614678\n",
      "Preditors: K + r + S + tau + e^(S) + e^(K)   Adj RSquared: 0.9692530673657499\n",
      "Preditors: K + r + S + tau + e^(S) + e^(tau)   Adj RSquared: 0.9135290447142419\n",
      "Preditors: K + r + S + tau + e^(tau) + e^(r)   Adj RSquared: 0.9133392914632987\n",
      "Preditors: K + r + S + tau + e^(tau) + e^(K)   Adj RSquared: 0.969418969312162\n",
      "Preditors: K + r + S + tau + e^(K) + e^(r)   Adj RSquared: 0.9692349615034728\n",
      "Preditors: K + r + S + tau + e^(K) + e^(S) + e^(r) + e^(tau)   Adj RSquared: 0.9695174044649394\n",
      "\n",
      "We can see the most important predictor is e^K from its impact on Adj RSquared values. Alone it has Adj R2=0.969\n"
     ]
    }
   ],
   "source": [
    "#Exponential Transformations (e^x)\n",
    "import warnings\n",
    "warnings.filterwarnings('ignore')\n",
    "result1 = smf.ols(f'Value ~ K + S + r + tau + np.exp(K)', data=option_train_train).fit()\n",
    "print(f'Preditors: K + r + S + tau + e^(K)', '  Adj RSquared:', result1.rsquared_adj)\n",
    "result1 = smf.ols(f'Value ~ K + S + r + tau + np.exp(S)', data=option_train_train).fit()\n",
    "print(f'Preditors: K + r + S + tau + e^(S)', '  Adj RSquared:', result1.rsquared_adj)\n",
    "result1 = smf.ols(f'Value ~ K + S + r + tau + np.exp(r)', data=option_train_train).fit()\n",
    "print(f'Preditors: K + r + S + tau + e^(r)', '  Adj RSquared:', result1.rsquared_adj)\n",
    "result1 = smf.ols(f'Value ~ K + S + r + tau + np.exp(tau)', data=option_train_train).fit()\n",
    "print(f'Preditors: K + r + S + tau + e^(tau)', '  Adj RSquared:', result1.rsquared_adj)\n",
    "result1 = smf.ols(f'Value ~ K + S + r + tau + np.exp(r) + np.exp(S)', data=option_train_train).fit()\n",
    "print(f'Preditors: K + r + S + tau + e^(S) + e^(r)', '  Adj RSquared:', result1.rsquared_adj)\n",
    "result1 = smf.ols(f'Value ~ K + S + r + tau + np.exp(K) + np.exp(S)', data=option_train_train).fit()\n",
    "print(f'Preditors: K + r + S + tau + e^(S) + e^(K)', '  Adj RSquared:', result1.rsquared_adj)\n",
    "result1 = smf.ols(f'Value ~ K + S + r + tau + np.exp(tau) + np.exp(S)', data=option_train_train).fit()\n",
    "print(f'Preditors: K + r + S + tau + e^(S) + e^(tau)', '  Adj RSquared:', result1.rsquared_adj)\n",
    "result1 = smf.ols(f'Value ~ K + S + r + tau + np.exp(r) + np.exp(tau)', data=option_train_train).fit()\n",
    "print(f'Preditors: K + r + S + tau + e^(tau) + e^(r)', '  Adj RSquared:', result1.rsquared_adj)\n",
    "result1 = smf.ols(f'Value ~ K + S + r + tau + np.exp(K) + np.exp(tau)', data=option_train_train).fit()\n",
    "print(f'Preditors: K + r + S + tau + e^(tau) + e^(K)', '  Adj RSquared:', result1.rsquared_adj)\n",
    "result1 = smf.ols(f'Value ~ K + S + r + tau + np.exp(r) + np.exp(K)', data=option_train_train).fit()\n",
    "print(f'Preditors: K + r + S + tau + e^(K) + e^(r)', '  Adj RSquared:', result1.rsquared_adj)\n",
    "result1 = smf.ols(f'Value ~ K + S + r + tau + np.exp(K) + np.exp(r) + np.exp(S) + np.exp(tau)', data=option_train_train).fit()\n",
    "print(f'Preditors: K + r + S + tau + e^(K) + e^(S) + e^(r) + e^(tau)', '  Adj RSquared:', result1.rsquared_adj)\n",
    "print()\n",
    "print('We can see the most important predictor is e^K from its impact on Adj RSquared values. Alone it has Adj R2=0.969')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "85e8106a",
   "metadata": {
    "hidden": true
   },
   "source": [
    "### Fractional transformations "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 295,
   "id": "ab75e08f",
   "metadata": {
    "hidden": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Preditors: K + r + S + tau + 1/K   Adj RSquared: 0.9144629360904243\n",
      "Preditors: K + r + S + tau + 1/S   Adj RSquared: 0.9129131137118671\n",
      "Preditors: K + r + S + tau + 1/r   Adj RSquared: 0.9127455224071006\n",
      "Preditors: K + r + S + tau + 1/tau   Adj RSquared: 0.9127170237644765\n",
      "Preditors: K + r + S + tau + 1/K + 1/S + 1/r + 1/tau   Adj RSquared: 0.9145933108009753\n",
      "\n",
      "The Adj RSquared values for all fractional terms are very similar and equally beneficial\n",
      "For simplicity and variety, we will include the most important predictor as 1/r. Alone is has Adj R2=0.9127\n"
     ]
    }
   ],
   "source": [
    "#Fractional transformations \n",
    "import warnings\n",
    "warnings.filterwarnings('ignore')\n",
    "result1 = smf.ols(f'Value ~ K + S + r + tau + np.power(K,-1)', data=option_train_train).fit()\n",
    "print(f'Preditors: K + r + S + tau + 1/K', '  Adj RSquared:', result1.rsquared_adj)\n",
    "result1 = smf.ols(f'Value ~ K + S + r + tau + np.power(S,-1)', data=option_train_train).fit()\n",
    "print(f'Preditors: K + r + S + tau + 1/S', '  Adj RSquared:', result1.rsquared_adj)\n",
    "result1 = smf.ols(f'Value ~ K + S + r + tau + np.power(r,-1)', data=option_train_train).fit()\n",
    "print(f'Preditors: K + r + S + tau + 1/r', '  Adj RSquared:', result1.rsquared_adj)\n",
    "result1 = smf.ols(f'Value ~ K + S + r + tau + np.power(tau,-1)', data=option_train_train).fit()\n",
    "print(f'Preditors: K + r + S + tau + 1/tau', '  Adj RSquared:', result1.rsquared_adj)\n",
    "result1 = smf.ols(f'Value ~ K + S + r + tau + np.power(K,-1) + np.power(r,-1) + np.power(S,-1) + np.power(tau,-1)', \n",
    "                  data=option_train_train).fit()\n",
    "print(f'Preditors: K + r + S + tau + 1/K + 1/S + 1/r + 1/tau', '  Adj RSquared:', result1.rsquared_adj)\n",
    "print()\n",
    "print('The Adj RSquared values for all fractional terms are very similar and equally beneficial')\n",
    "print('For simplicity and variety, we will include the most important predictor as 1/r. Alone is has Adj R2=0.9127')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9164aabe",
   "metadata": {
    "hidden": true
   },
   "source": [
    "### More Exponential Transformations"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 296,
   "id": "a28a1fd6",
   "metadata": {
    "hidden": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Preditors: K + r + S + tau + K^r  Adj RSquared: 0.9166188774579507\n",
      "Preditors: K + r + S + tau + S^r  Adj RSquared: 0.9127897811475457\n",
      "Preditors: K + r + S + tau + tau^r  Adj RSquared: 0.9131951872623336\n",
      "Preditors: K + r + S + tau + K^-r  Adj RSquared: 0.9133679885737379\n",
      "Preditors: K + r + S + tau + S^-r  Adj RSquared: 0.912913156239497\n",
      "Preditors: K + r + S + tau + tau^-r  Adj RSquared: 0.9131889213629474\n",
      "Preditors: K + r + S + tau + S^K  Adj RSquared: 0.9130328715098092\n",
      "Preditors: K + r + S + tau + r^K  Adj RSquared: 0.9133189180633805\n",
      "Preditors: K + r + S + tau + tau^K  Adj RSquared: 0.9230102756258632\n",
      "Preditors: K + r + S + tau + S^-K  Adj RSquared: 0.9135002625554809\n",
      "Preditors: K + r + S + tau + r^-K  Adj RSquared: 0.913530533764135\n",
      "Preditors: K + r + S + tau + tau^-K  Adj RSquared: 0.9127170507089629\n",
      "Preditors: K + r + S + tau + K^S  Adj RSquared: 0.920241683994025\n",
      "Preditors: K + r + S + tau + r^S  Adj RSquared: 0.9127766448080645\n",
      "Preditors: K + r + S + tau + tau^S  Adj RSquared: 0.9128393647195423\n",
      "Preditors: K + r + S + tau + K^-S  Adj RSquared: 0.9138719721407443\n",
      "Preditors: K + r + S + tau + r^-S  Adj RSquared: 0.9126314709352255\n",
      "Preditors: K + r + S + tau + tau^-S  Adj RSquared: 0.9127655019021931\n",
      "Preditors: K + r + S + tau + K^tau  Adj RSquared: 0.9126328473814216\n",
      "Preditors: K + r + S + tau + S^tau  Adj RSquared: 0.912964094912216\n",
      "Preditors: K + r + S + tau + r^tau  Adj RSquared: 0.9126748474760328\n",
      "Preditors: K + r + S + tau + K^-tau  Adj RSquared: 0.9140220398508674\n",
      "Preditors: K + r + S + tau + S^-tau  Adj RSquared: 0.9129131562396716\n",
      "Preditors: K + r + S + tau + r^-tau  Adj RSquared: 0.9128354862977203\n",
      "\n",
      "The Adj RSquared values for all exponential terms are very similar and equally beneficial\n",
      "For simplicity and variety, we will include the most important predictor as tau^K. Alone is has Adj R2=0.9230\n"
     ]
    }
   ],
   "source": [
    "#More Exponential Transformations (Because we would like to include r in the model)\n",
    "for i in list(option_train.columns.difference(['Value','BS','r'])):\n",
    "    result1 = smf.ols(f'Value ~ K + r + S + tau + np.power({i}, r)', data=option_train_train).fit()\n",
    "    print(f'Preditors: K + r + S + tau + {i}^r', ' Adj RSquared:', result1.rsquared_adj)\n",
    "for i in list(option_train.columns.difference(['Value','BS','r'])):\n",
    "    result1 = smf.ols(f'Value ~ K + r + S + tau + np.float_power({i}, -r)', data=option_train_train).fit()\n",
    "    print(f'Preditors: K + r + S + tau + {i}^-r', ' Adj RSquared:', result1.rsquared_adj)\n",
    "for i in list(option_train.columns.difference(['Value','BS','K'])):\n",
    "    result1 = smf.ols(f'Value ~ K + r + S + tau + np.float_power({i}, K)', data=option_train_train).fit()\n",
    "    print(f'Preditors: K + r + S + tau + {i}^K', ' Adj RSquared:', result1.rsquared_adj)\n",
    "for i in list(option_train.columns.difference(['Value','BS','K'])):\n",
    "    result1 = smf.ols(f'Value ~ K + r + S + tau + np.float_power({i}, -K)', data=option_train_train).fit()\n",
    "    print(f'Preditors: K + r + S + tau + {i}^-K', ' Adj RSquared:', result1.rsquared_adj)\n",
    "for i in list(option_train.columns.difference(['Value','BS','S'])):\n",
    "    result1 = smf.ols(f'Value ~ K + r + S + tau + np.float_power({i}, S)', data=option_train_train).fit()\n",
    "    print(f'Preditors: K + r + S + tau + {i}^S', ' Adj RSquared:', result1.rsquared_adj)\n",
    "for i in list(option_train.columns.difference(['Value','BS','S'])):\n",
    "    result1 = smf.ols(f'Value ~ K + r + S + tau + np.float_power({i}, -S)', data=option_train_train).fit()\n",
    "    print(f'Preditors: K + r + S + tau + {i}^-S', ' Adj RSquared:', result1.rsquared_adj)\n",
    "for i in list(option_train.columns.difference(['Value','BS','tau'])):\n",
    "    result1 = smf.ols(f'Value ~ K + r + S + tau + np.float_power({i}, tau)', data=option_train_train).fit()\n",
    "    print(f'Preditors: K + r + S + tau + {i}^tau', ' Adj RSquared:', result1.rsquared_adj)\n",
    "for i in list(option_train.columns.difference(['Value','BS','tau'])):\n",
    "    result1 = smf.ols(f'Value ~ K + r + S + tau + np.float_power({i}, -tau)', data=option_train_train).fit()\n",
    "    print(f'Preditors: K + r + S + tau + {i}^-tau', ' Adj RSquared:', result1.rsquared_adj)\n",
    "print()\n",
    "print('The Adj RSquared values for all exponential terms are very similar and equally beneficial')\n",
    "print('For simplicity and variety, we will include the most important predictor as tau^K. Alone is has Adj R2=0.9230')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6d54f7b2",
   "metadata": {
    "hidden": true
   },
   "source": [
    "### Final Equation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 398,
   "id": "524791e3",
   "metadata": {
    "hidden": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Preditors: K + S + tau + K*S + K^2 + log(tau) + e^(K)+ 1/r + tau^K \n",
      "Adj RSquared:  0.9934322929033138\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<table class=\"simpletable\">\n",
       "<caption>OLS Regression Results</caption>\n",
       "<tr>\n",
       "  <th>Dep. Variable:</th>          <td>Value</td>      <th>  R-squared:         </th> <td>   0.993</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Model:</th>                   <td>OLS</td>       <th>  Adj. R-squared:    </th> <td>   0.993</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Method:</th>             <td>Least Squares</td>  <th>  F-statistic:       </th> <td>2.214e+04</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Date:</th>             <td>Thu, 21 Apr 2022</td> <th>  Prob (F-statistic):</th>  <td>  0.00</td>  \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Time:</th>                 <td>14:41:31</td>     <th>  Log-Likelihood:    </th> <td> -1799.3</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>No. Observations:</th>      <td>  1172</td>      <th>  AIC:               </th> <td>   3617.</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Df Residuals:</th>          <td>  1163</td>      <th>  BIC:               </th> <td>   3662.</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Df Model:</th>              <td>     8</td>      <th>                     </th>     <td> </td>    \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Covariance Type:</th>      <td>nonrobust</td>    <th>                     </th>     <td> </td>    \n",
       "</tr>\n",
       "</table>\n",
       "<table class=\"simpletable\">\n",
       "<tr>\n",
       "          <td></td>            <th>coef</th>     <th>std err</th>      <th>t</th>      <th>P>|t|</th>  <th>[0.025</th>    <th>0.975]</th>  \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Intercept</th>        <td>   47.1691</td> <td>    0.459</td> <td>  102.817</td> <td> 0.000</td> <td>   46.269</td> <td>   48.069</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>K</th>                <td> -162.3278</td> <td>    1.098</td> <td> -147.859</td> <td> 0.000</td> <td> -164.482</td> <td> -160.174</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>S</th>                <td>   45.6838</td> <td>    0.427</td> <td>  107.079</td> <td> 0.000</td> <td>   44.847</td> <td>   46.521</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>tau</th>              <td>    6.4078</td> <td>    0.531</td> <td>   12.065</td> <td> 0.000</td> <td>    5.366</td> <td>    7.450</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>K:S</th>              <td>  -53.7405</td> <td>    0.821</td> <td>  -65.445</td> <td> 0.000</td> <td>  -55.352</td> <td>  -52.129</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>np.power(K, 2)</th>   <td>  108.4854</td> <td>    0.911</td> <td>  119.090</td> <td> 0.000</td> <td>  106.698</td> <td>  110.273</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>np.log(tau)</th>      <td>    0.1613</td> <td>    0.042</td> <td>    3.833</td> <td> 0.000</td> <td>    0.079</td> <td>    0.244</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>np.exp(K)</th>        <td>    2.7895</td> <td>    0.128</td> <td>   21.809</td> <td> 0.000</td> <td>    2.539</td> <td>    3.040</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>np.power(r, -1)</th>  <td> 1.999e-14</td> <td> 2.16e-14</td> <td>    0.927</td> <td> 0.354</td> <td>-2.23e-14</td> <td> 6.23e-14</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>np.power(tau, K)</th> <td>    2.5550</td> <td>    0.727</td> <td>    3.513</td> <td> 0.000</td> <td>    1.128</td> <td>    3.982</td>\n",
       "</tr>\n",
       "</table>\n",
       "<table class=\"simpletable\">\n",
       "<tr>\n",
       "  <th>Omnibus:</th>       <td>15.157</td> <th>  Durbin-Watson:     </th> <td>   1.958</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Prob(Omnibus):</th> <td> 0.001</td> <th>  Jarque-Bera (JB):  </th> <td>  22.573</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Skew:</th>          <td>-0.105</td> <th>  Prob(JB):          </th> <td>1.25e-05</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Kurtosis:</th>      <td> 3.647</td> <th>  Cond. No.          </th> <td>1.02e+15</td>\n",
       "</tr>\n",
       "</table><br/><br/>Notes:<br/>[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.<br/>[2] The condition number is large, 1.02e+15. This might indicate that there are<br/>strong multicollinearity or other numerical problems."
      ],
      "text/plain": [
       "<class 'statsmodels.iolib.summary.Summary'>\n",
       "\"\"\"\n",
       "                            OLS Regression Results                            \n",
       "==============================================================================\n",
       "Dep. Variable:                  Value   R-squared:                       0.993\n",
       "Model:                            OLS   Adj. R-squared:                  0.993\n",
       "Method:                 Least Squares   F-statistic:                 2.214e+04\n",
       "Date:                Thu, 21 Apr 2022   Prob (F-statistic):               0.00\n",
       "Time:                        14:41:31   Log-Likelihood:                -1799.3\n",
       "No. Observations:                1172   AIC:                             3617.\n",
       "Df Residuals:                    1163   BIC:                             3662.\n",
       "Df Model:                           8                                         \n",
       "Covariance Type:            nonrobust                                         \n",
       "====================================================================================\n",
       "                       coef    std err          t      P>|t|      [0.025      0.975]\n",
       "------------------------------------------------------------------------------------\n",
       "Intercept           47.1691      0.459    102.817      0.000      46.269      48.069\n",
       "K                 -162.3278      1.098   -147.859      0.000    -164.482    -160.174\n",
       "S                   45.6838      0.427    107.079      0.000      44.847      46.521\n",
       "tau                  6.4078      0.531     12.065      0.000       5.366       7.450\n",
       "K:S                -53.7405      0.821    -65.445      0.000     -55.352     -52.129\n",
       "np.power(K, 2)     108.4854      0.911    119.090      0.000     106.698     110.273\n",
       "np.log(tau)          0.1613      0.042      3.833      0.000       0.079       0.244\n",
       "np.exp(K)            2.7895      0.128     21.809      0.000       2.539       3.040\n",
       "np.power(r, -1)   1.999e-14   2.16e-14      0.927      0.354   -2.23e-14    6.23e-14\n",
       "np.power(tau, K)     2.5550      0.727      3.513      0.000       1.128       3.982\n",
       "==============================================================================\n",
       "Omnibus:                       15.157   Durbin-Watson:                   1.958\n",
       "Prob(Omnibus):                  0.001   Jarque-Bera (JB):               22.573\n",
       "Skew:                          -0.105   Prob(JB):                     1.25e-05\n",
       "Kurtosis:                       3.647   Cond. No.                     1.02e+15\n",
       "==============================================================================\n",
       "\n",
       "Notes:\n",
       "[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n",
       "[2] The condition number is large, 1.02e+15. This might indicate that there are\n",
       "strong multicollinearity or other numerical problems.\n",
       "\"\"\""
      ]
     },
     "execution_count": 398,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Final Equation\n",
    "from sklearn.metrics import mean_squared_error\n",
    "model = smf.ols('Value ~ K + S + tau + \\\n",
    "K*S + \\\n",
    "np.power(K, 2) + \\\n",
    "np.log(tau) + \\\n",
    "np.exp(K) + \\\n",
    "np.power(r,-1) + \\\n",
    "np.power(tau,K)', data=option_train_train).fit() \n",
    "print(f'Preditors: K + S + tau + \\\n",
    "K*S + \\\n",
    "K^2 + \\\n",
    "log(tau) + \\\n",
    "e^(K)+ \\\n",
    "1/r + \\\n",
    "tau^K \\nAdj RSquared: ', model.rsquared_adj)\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6a61ac56",
   "metadata": {
    "hidden": true
   },
   "source": [
    "### Test R2 Score on final model with Test Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 399,
   "id": "20653512",
   "metadata": {
    "hidden": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RSquared: 0.9938345194320443\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import r2_score\n",
    "from sklearn.model_selection import cross_val_score\n",
    "from sklearn.model_selection import RepeatedKFold\n",
    "model = smf.ols('Value ~ K + S + tau + \\\n",
    "K*S + \\\n",
    "np.power(K, 2) + \\\n",
    "np.log(tau) + \\\n",
    "np.exp(K) + \\\n",
    "np.power(r,-1) + \\\n",
    "np.power(tau,K)', data=option_train_train).fit()\n",
    "y_pred = model.predict(option_train_test)\n",
    "print('RSquared:',r2_score(y_test, y_pred))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4b330bb0",
   "metadata": {
    "hidden": true
   },
   "source": [
    "- We are confident there was no overfitting or data leakage during our trials.\n",
    "- We believe this equation will perform well (high out-of-sample Rsquared) when predicting Value using the unseen options test data."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "54a54958",
   "metadata": {},
   "source": [
    "## Classification (BS)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "91fefe84",
   "metadata": {},
   "source": [
    "### Normalizing Original Training Data (No Change in BS Binary Variable)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "654bddbc",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Normalizing Original Training Data (No Change in BS Binary Variable)\n",
    "from sklearn import metrics\n",
    "from sklearn.model_selection import train_test_split\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import pandas as pd\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.preprocessing import MinMaxScaler\n",
    "feature_cols = ['S','K','tau','r']\n",
    "X = option_train[feature_cols]\n",
    "y = option_train[\"BS\"]\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3,random_state=1)\n",
    "mms= MinMaxScaler()\n",
    "X_train = mms.fit_transform(X_train)\n",
    "X_test = mms.transform(X_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "66311371",
   "metadata": {},
   "source": [
    "### GridSearchCV"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b18328c4",
   "metadata": {},
   "source": [
    "- For all classifiers, we will use GridSearchCV to find the best combination of tuned hyperparameters.\n",
    "\n",
    "- We will return the average cross-validated score of the specific Classifier."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "1b79bbe0",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import GridSearchCV"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "32124ce0",
   "metadata": {},
   "source": [
    "### Logistic Regression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "id": "2972f17f",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best Parameters:  {'C': 100, 'penalty': 'l2'}\n",
      "Accuracy Score:  0.9155512096190062\n"
     ]
    }
   ],
   "source": [
    "#Logistic Regression\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "model = LogisticRegression()\n",
    "forest_params = {'C': [0.01, 0.1, 1, 2, 10, 100], 'penalty': ['l1', 'l2']}\n",
    "clf = GridSearchCV(model, forest_params, cv = 10, scoring='accuracy')\n",
    "clf.fit(X_train, y_train)\n",
    "print('Best Parameters: ', clf.best_params_)\n",
    "print('Accuracy Score: ', clf.best_score_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 432,
   "id": "58902f39",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion Matrix:\n",
      " [[266  18]\n",
      " [ 24 195]]\n",
      "Accuracy: 0.9165009940357853\n",
      "Classification Error: 0.08349900596421467\n",
      "Mean Cross Val (Accuracy) Score 0.9098268656716417\n",
      "Mean Classification Error: 0.09017313432835827\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x216 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Area Under Curve: 0.9767187600488777\n"
     ]
    }
   ],
   "source": [
    "#Logistic Regression\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.metrics import auc\n",
    "from sklearn.model_selection import cross_val_score\n",
    "from sklearn.model_selection import RepeatedKFold\n",
    "model = LogisticRegression(C= 100, penalty= 'l2', random_state=1).fit(X_train, y_train)\n",
    "y_pred = model.predict(X_test)\n",
    "cnf_matrix = metrics.confusion_matrix(y_test, y_pred)\n",
    "print(\"Confusion Matrix:\\n\", cnf_matrix)\n",
    "print(\"Accuracy:\",metrics.accuracy_score(y_test, y_pred))\n",
    "print(\"Classification Error:\",1-metrics.accuracy_score(y_test, y_pred))\n",
    "cv = RepeatedKFold(n_splits=5, n_repeats= 100, random_state=1)\n",
    "print(\"Mean Cross Val (Accuracy) Score\", np.mean(cross_val_score(model, X, y, scoring='accuracy', cv=cv, n_jobs=-1)))\n",
    "print(\"Mean Classification Error:\", 1 - np.mean(cross_val_score(model, X, y, scoring='accuracy', cv=cv, n_jobs=-1)))\n",
    "y_pred_proba = model.predict_proba(X_test)[:,1]\n",
    "fpr, tpr, _ = metrics.roc_curve(y_test, y_pred_proba)\n",
    "roc_auc = auc(fpr,tpr)\n",
    "plt.figure(figsize=(4,3))\n",
    "plt.plot(fpr, tpr, color='orange',\n",
    "lw=2, label='ROC curve (area = {0:.4f})'.format(roc_auc))\n",
    "plt.plot([0, 1], [0, 1], color='blue', lw=2, linestyle='--')\n",
    "plt.xlim([0.0, 1.0])\n",
    "plt.ylim([0.0, 1.05])\n",
    "plt.xlabel('False Positive Rate')\n",
    "plt.ylabel('True Positive Rate')\n",
    "plt.title('ROC curve')\n",
    "plt.legend(loc=\"lower right\")\n",
    "plt.show()\n",
    "print('Area Under Curve:', roc_auc)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f9f826c4",
   "metadata": {},
   "source": [
    "### Decision Tree Classifier "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "id": "8030d758",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best Parameters:  {'criterion': 'entropy', 'max_depth': 8, 'min_samples_leaf': 3, 'min_samples_split': 2}\n",
      "Accuracy Score:  0.9257279443720122\n"
     ]
    }
   ],
   "source": [
    "#Decision Tree Classifier\n",
    "from sklearn.tree import DecisionTreeClassifier\n",
    "model = DecisionTreeClassifier()\n",
    "forest_params = {\"criterion\" : ['gini', 'entropy'], \n",
    "                 'max_depth':range(1,10),\"min_samples_split\": range(1,10),\n",
    "                 \"min_samples_leaf\":range(1,5)}\n",
    "clf = GridSearchCV(model, forest_params, cv = 10, scoring='accuracy')\n",
    "clf.fit(X_train, y_train)\n",
    "print('Best Parameters: ', clf.best_params_)\n",
    "print('Accuracy Score: ', clf.best_score_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "id": "7eb791f6",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.9165009940357853\n",
      "Classification Error: 0.08349900596421467\n",
      "Mean Cross Val (Accuracy) Score 0.916179104477612\n",
      "Mean Classification Error: 0.08382089552238803\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x216 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Area Under Curve: 0.9432278603125605\n"
     ]
    }
   ],
   "source": [
    "#Decision Tree Classifier \n",
    "from sklearn.tree import DecisionTreeClassifier\n",
    "model = DecisionTreeClassifier(criterion='entropy', max_depth= 8, min_samples_leaf= 3, min_samples_split= 2, \n",
    "                               random_state=1).fit(X_train, y_train)\n",
    "y_pred = model.predict(X_test)\n",
    "print(\"Accuracy:\",metrics.accuracy_score(y_test, y_pred))\n",
    "print(\"Classification Error:\",1-metrics.accuracy_score(y_test, y_pred))\n",
    "cv = RepeatedKFold(n_splits=5, n_repeats= 100, random_state=1)\n",
    "print(\"Mean Cross Val (Accuracy) Score\", np.mean(cross_val_score(model, X, y, cv=cv)))\n",
    "print(\"Mean Classification Error:\", 1 - np.mean(cross_val_score(model, X, y, scoring='accuracy', cv=cv, n_jobs=-1)))\n",
    "y_pred_proba = model.predict_proba(X_test)[:,1]\n",
    "fpr, tpr, _ = metrics.roc_curve(y_test, y_pred_proba)\n",
    "roc_auc = auc(fpr,tpr)\n",
    "plt.figure(figsize=(4,3))\n",
    "plt.plot(fpr, tpr, color='orange',\n",
    "lw=2, label='ROC curve (area = {0:.4f})'.format(roc_auc))\n",
    "plt.plot([0, 1], [0, 1], color='blue', lw=2, linestyle='--')\n",
    "plt.xlim([0.0, 1.0])\n",
    "plt.ylim([0.0, 1.05])\n",
    "plt.xlabel('False Positive Rate')\n",
    "plt.ylabel('True Positive Rate')\n",
    "plt.title('ROC curve')\n",
    "plt.legend(loc=\"lower right\")\n",
    "plt.show()\n",
    "print('Area Under Curve:', roc_auc) "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b6859921",
   "metadata": {},
   "source": [
    "### KNN Classifier"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c97d5a7f",
   "metadata": {},
   "source": [
    "- We use a different method to determine optimal k value "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "id": "003ed01e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy value for k= 2 is: 0.9006  and RMSE value is: 0.31528\n",
      "Accuracy value for k= 3 is: 0.90258  and RMSE value is: 0.31211\n",
      "Accuracy value for k= 4 is: 0.89463  and RMSE value is: 0.3246\n",
      "Accuracy value for k= 5 is: 0.89463  and RMSE value is: 0.3246\n",
      "Accuracy value for k= 6 is: 0.90258  and RMSE value is: 0.31211\n",
      "Accuracy value for k= 7 is: 0.91451  and RMSE value is: 0.29238\n",
      "Accuracy value for k= 8 is: 0.9165  and RMSE value is: 0.28896\n",
      "Accuracy value for k= 9 is: 0.92247  and RMSE value is: 0.27845\n",
      "Accuracy value for k= 10 is: 0.91849  and RMSE value is: 0.2855\n",
      "Accuracy value for k= 11 is: 0.92048  and RMSE value is: 0.282\n",
      "Accuracy value for k= 12 is: 0.91252  and RMSE value is: 0.29576\n",
      "Accuracy value for k= 13 is: 0.91252  and RMSE value is: 0.29576\n",
      "Accuracy value for k= 14 is: 0.9165  and RMSE value is: 0.28896\n",
      "Accuracy value for k= 15 is: 0.92247  and RMSE value is: 0.27845\n",
      "Accuracy value for k= 16 is: 0.92247  and RMSE value is: 0.27845\n",
      "Accuracy value for k= 17 is: 0.92843  and RMSE value is: 0.26753\n",
      "Accuracy value for k= 18 is: 0.9165  and RMSE value is: 0.28896\n",
      "Accuracy value for k= 19 is: 0.92445  and RMSE value is: 0.27486\n",
      "Accuracy value for k= 20 is: 0.91849  and RMSE value is: 0.2855\n",
      "Accuracy value for k= 21 is: 0.92247  and RMSE value is: 0.27845\n",
      "Accuracy value for k= 22 is: 0.9165  and RMSE value is: 0.28896\n",
      "Accuracy value for k= 23 is: 0.92843  and RMSE value is: 0.26753\n",
      "Accuracy value for k= 24 is: 0.91849  and RMSE value is: 0.2855\n",
      "Accuracy value for k= 25 is: 0.92644  and RMSE value is: 0.27122\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAR8AAADgCAYAAAA65GI1AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuNCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8QVMy6AAAACXBIWXMAAAsTAAALEwEAmpwYAAA45ElEQVR4nO2deXyU1dX4vychhC2QEMIqEPYlEKJE8VWUDIg7KlqCti7YWktrrdXX31tbrYbaKkV9rX2tWq1YUIsErVqtbVVMWFUEBMmEVQgQiEnYl5CQ5fz+eJ6JQzKTmSQzmWTmfj+f+cwzz93OMzAn95577jmiqhgMBkNLExVqAQwGQ2RilI/BYAgJRvkYDIaQYJSPwWAICUb5GAyGkGCUj8FgCAlG+YQRIlIgIpeEWo5IQ0SyROS1UMvR1jDKJ0KwFVOxiHR2u3eHiOS6fVYR2SQiUW73fisif21ZacMX8wfiW4zyiSzaAff4qNMXuLEFZGk2ItIu1DIYmo5RPmGKiIwUkV0i4q5IngDuF5H4BprOA+b488MWkQQReV9ESkXksH19llt5dxF5RUT22+XvuJVdKyIbROSYiHwtIpfb98+YGbgvaUQk2Z6d/UBE9gCf2PeXiMg3InJURJaLSIpb+44i8pSI7LbLV9r3/ikid9d5nq9E5DoPz+ka9077WYpE5L8b+F6uERGniBwRkVwRGWXffxUYALwnIidE5H98fcfhjFE+YYiInAN8CNytqm+4Fa0FcoH7G2j+d+AYMMuPoaKAV4CBWD+qU8CzbuWvAp2AFKAn8LQt33nAQuD/AfHAxUCBH+O5mASMAi6zP/8LGGaPsR543a3uk8B44AKgO/A/QA2wALjZVUlExgH9gA8aGNdhj3Mp8ICn5ZOIDAcWAT8Hkuz+3hOR9qp6C7AHmKaqXVR1XiOeOfxQVfMKkxfWD3gOUAg4PJRdAowBjmL9MO4Act3qKDAUuBLrRxIL/Bb4q5/jpwGH7es+WD/yBA/1/gw83cAzXOL2OQt4zb5OtmUc3IAM8XadbljK8RQwzkO9WOAQMMz+/CTwnJc+XeOOdLs3D3jZg4y/BrLd6kUB+4AMT88XyS8z8wk/ZgOrVTXHU6Gq5gHvAw9460BVP8BSPnc2NJCIdBKRP9tLmmPAciBeRKKB/sAhVT3soWl/4Gu/nsYze91kiBaRufbS7RjfzqB62K8OnsZS1QogG7jZNrDfhDVT82tcYDeWfawufe0y1zg1drt+PvqOOIzyCT9mAwNE5OkG6jwC/JCGfxAPAQ9iLZu88d/ACGCCqnbFWj4BCNYPrrsX+9JeYIiXPk/WGbO3hzruoRi+C1yLNavrhjVLcclwAChvYKwFwPeAKUCZqn7qpZ6L/m7XA4D9Hursx1qGWkKIiN1unwfZIxqjfMKP48DlwMUiMtdTBVXdASwGfuatE1XNBTYBtzUwVhzWsuaIiHTHUmqu9kVYtpjnbMN0jIi4lNPLwO0iMkVEokSkn4iMtMs2ADfa9dOB7/h43jigAjiIpbQec5OhBpgP/K+I9LVnSf8lIrF2+adYS8On8D3rAfi1PdtLAW7H+g7rkg1cZT9bDJaCrgBW2+XFwGA/xgp7jPIJQ1T1CDAVuEJEHvVS7TdAZy9lLh7CMtJ64w9AR6wZxmfAv+uU3wJUAluAEiwjLKq6BuvH+zSW/WkZ384Wfo01UzmMZb/6mw8ZF2Itc/YB+bYc7tyPpUS/wLLx/J4z/98vBMYC/jgJLgN2AEuBJ1X1w7oVVHUrliH7/7C+l2lYBubTdpXHgYfsnbCGDP9hj9hGMIMhIhGRW4E7VXViA3WSgV1AjKpWtZRs4Y6Z+RgiFhHpBPwEeDHUskQiRvkYIhIRuQwoxbLB+FraGYJAUJWPiFwuIltFZIeI1NvatQ2Rb9uepWtEZIx9v4P9eaPtKTrHrU13EflIRLbb7wnBfAZDeKKq/1HVzqp6ra+llKoWqKqYJVdgCZrysX09/gRcAYwGbhKR0XWq/QrYoKqpwK3AM/b9CmCyqo7Dcly7XETOt8seAJaq6jAsw59XfxWDwdB6CebM5zxgh6rutC39b2D5Y7gzGkuBoKpbgGQR6aUWJ+w6MfbLZRm/Fss/A/v9uuA9gsFgCBbBPBXcjzM9QguBCXXqbASuB1ba530GAmcBxfbMaR2Wu/+fVPVzu00v24cEVS0SkZ6eBheRO7E9dDt37jx+5MiRnqoZDIYgsm7dugOqmuSpLJjKRzzcq7uvPxd4RkQ2YPlifAlUAahqNZBme8i+LSJj7KMBfqGqL2LvYqSnp+vatWsb/QAGg6F5iMhub2XBVD6FnOmOfhZ13NFV9RiWs5nLDX2X/XKvc0SsgFeXA3lYs6I+9qynD5bzmsFgaGME0+bzBTBMRAaJSHusAFX/cK8gIvF2GVgnrJer6jERSXKdCRKRjljndrbY9f7Bty7/twHvBvEZDAZDkAjazEdVq0Tkp8B/gGhgvqo6RWS2Xf4CVkyWhSJSjeUa/wO7eR9ggW33icIKUfC+XTYXyBaRH2CdvJ4RrGcwGAzBIyKOV3iy+VRWVlJYWEh5eXmIpDIEiw4dOnDWWWcRExMTalEiHhFZp6rpnsoiNgZuYWEhcXFxJCcnY5mbDOGAqnLw4EEKCwsZNGhQk/s5WHaQP37+Rx66+CFiosNDiRUdL+KVDa/wwMQHiJLQH24IvQQhory8nMTERKN4wgwRITExsdkz2ue+eI7fLP8Nq/auCpBkoWfhxoU8+MmDbPhmQ6hFASJY+QBG8YQpgfh3zc7PBsBZ4mx2X60FZ6n1LGv3tw63k4hWPqHk4MGDpKWlkZaWRu/evenXr1/t59OnT/tsn5uby+rVq2s/v/DCCyxcuDBg8pWWlhITE8Of//zngPXZVsgvzSevxHIpc/1gw4HWpnwi1uYTahITE9mwYQMAWVlZdOnShfvv9z+2VG5uLl26dOGCCy4AYPbs2QGVb8mSJZx//vksWrSIH/3oRwHt252qqiratWtd/w2zndlESRRDEobUKqG2TnVNNfml+UDrUT5m5tOKWLduHZMmTWL8+PFcdtllFBUVAfDHP/6R0aNHk5qayo033khBQQEvvPACTz/9NGlpaaxYsYKsrCyefPJJADIyMvjFL37Beeedx/Dhw1mxYgUAZWVlZGZmkpqaysyZM5kwYQLePL8XLVrEU089RWFhIfv27au9v3DhQlJTUxk3bhy33HILAMXFxUyfPp1x48Yxbtw4Vq9eTUFBAWPGjKlt9+STT5KVlVUr369+9SsmTZrEM888w3vvvceECRM4++yzueSSSyguLgbgxIkT3H777YwdO5bU1FTeeustXn75Ze69997afl966SXuu+++AP0LWAbrxc7FTBo4iYzkDJylTsJhR3jXkV2UV5XTN64vm0o2UV4V+l3e1vUnJ0T8/N8/D7gRLq13Gn+4/A9+11dV7r77bt59912SkpJYvHgxDz74IPPnz2fu3Lns2rWL2NhYjhw5Qnx8PLNnzz5jtrR06dIz+quqqmLNmjV88MEHzJkzh48//pjnnnuOhIQEvvrqK/Ly8khLS/Moy969e/nmm28477zzyMzMZPHixdx33304nU5+97vfsWrVKnr06MGhQ4cA+NnPfsakSZN4++23qa6u5sSJExw+7ClpxbccOXKEZcuWAXD48GE+++wzRIS//OUvzJs3j6eeeopHH32Ubt26sWnTptp67du3JzU1lXnz5hETE8Mrr7wS0KVhXkkeWw5s4Z4J91BRVcFL61+i+GQxvbt4imPfdnDZrm4eezPzVs9jU/Emzu13bkhlMjOfVkJFRQV5eXlMnTqVtLQ0fvvb31JYWAhAamoq3/ve93jttdf8XqJcf/31AIwfP56CggIAVq5cyY03WglMx4wZQ2pqqse2b7zxBpmZmQDceOONLFq0CIBPPvmE73znO/To0QOA7t27197/8Y9/DEB0dDTdunXzKd/MmTNrrwsLC7nssssYO3YsTzzxBE6n9UP5+OOPueuuu2rrJSQk0LlzZyZPnsz777/Pli1bqKysZOzYsX59J/6w2LmYKIni+lHXM6anNXMLB6Ozy95zyzhrttoall5m5gONmqEEC1UlJSWFTz+tn73ln//8J8uXL+cf//gHjz76aO2PsyFiY2MBSxlUVVXVjuEPixYtori4mNdftxJ/7t+/n+3bt6Oqfu8ktWvXjpqamtrPdbe+O3f+Nnb93XffzX333cc111xDbm5u7fLM23h33HEHjz32GCNHjuT222/3Sx5/UFWyndlMHjSZnp17ktLTyrrsLHUyZfCUgI0TCpylTgZ0G0BKUgo9OvVolvKp0ZqA+AmZmU8rITY2ltLS0lrlU1lZidPppKamhr179+JwOJg3bx5HjhzhxIkTxMXFcfz48UaNMXHiRLKzrS3k/Pz82uWMO1u3buXkyZPs27ePgoICCgoK+OUvf8kbb7zBlClTyM7O5uDBgwC1y64pU6bw/PPPA1BdXc2xY8fo1asXJSUlHDx4kIqKCt5///16Y7k4evQo/fpZKcQWLFhQe//SSy/l2We/zb7sWspNmDCBvXv38re//Y2bbrqpUd9BQ2ws3sj2Q9vJHG3N+np17kX3jt3DY+ZT4mRMzzGICOl901lb1HTlM/PNmdyQfUOzZTLKp5UQFRXFm2++yS9+8QvGjRtHWloaq1evprq6mptvvpmxY8dy9tlnc++99xIfH8+0adN4++23aw3O/vCTn/yE0tJSUlNT+f3vf09qamq9JdKiRYuYPn36GfduuOEGFi1aREpKCg8++CCTJk1i3LhxtYbeZ555hpycHMaOHcv48eNxOp3ExMTw8MMPM2HCBK6++moaiqeUlZXFjBkzuOiii2qXdAAPPfQQhw8fZsyYMYwbN46cnG+TsGZmZnLhhReSkBC4KLqL8xYTLdFcP8pasooIKUkp5JW27R2vqpoqNh/YTEqSNZNL75OOs8RJWWVZo/uq0RqW7lxKfGx88wULdb7mlniNHz9e65Kfn1/vXrhTVVWlp06dUlXVHTt26MCBA7WioiLEUjWNq666Sj/++GOv5Y39962pqdHBzwzWy1697Iz7s9+brd0e76Y1NTVNkrM1sKV0i5KF/vXLv6qq6jub31Gy0NV7Vje6ry+LvlSy0Fc3vupXfWCtmlzthrKyMiZOnMi4ceOYPn06zz//PO3bt/fdsBVx5MgRhg8fTseOHZkyJXB2mHVF69h5eCeZKZln3E/pmcLRiqPsP+4pM3LbwGVsdtmwxvcdD1jP3Fhydlmzz4zkjGbLFVSDs4hcjhUUPhr4i6rOrVOegJXOdghWTu3vq2qeiPTHyiTZGyud7Yuq+ozdJgsrz3ip3c2vVPWDYD5HuBAXF+fVr6etEB8fz7Zt2wLeb7Yzm5ioGKaPPHPJWbvjVeqkX9eGUtu3Xlw2q1E9RgHQL64fvTr3apLROacgh6Hdh3JW17OaLVdrzV5RBfy3qo4CzgfuqtP2aVVNs19G8Riahdq7XFOHTCWh45k2JJedpC0bnZ2lTgbFD6Jze2uHsdbo3EjlU11TzfLdy3EkOwIiV2vNXlGkquvt+8eBzVgB6QOKhoHnqqE+jf13XbNvDbuP7q7d5XInqXMSSZ2S2vQZL2eps3YG5yK9bzqbD2zmxOkTXlrVZ8M3GzhacbRNKB9P2SvqKhBX9grqZK+oxc6TfTbwudvtn9qJBud7SxooIneKyFoRWVtaWlqvvEOHDhw8eNAooDBD7Xg+HTp08LtNtjOb9tHtuXZk3b+NFik9U9rsGa/K6kq2HthaO4Nzkd43nRqtaZRnf05B4Ow90IqzVwCISBfgLeDnagWbB3geeNTu61HgKeD79Qaqk72ibvlZZ51FYWEhnhSToW3jimToDzVaQ3Z+NpcNuYz4DvEe66QkpbBw48JGOVm6qKiqoKqmqnbJ09JsP7SdyprKWmOzi/F9bKPz/nVMHDDRr75yCnIYkTiCPnF9AiJbq81eISIxWIrndVX9u1ubYte1iLwEePdea4CYmJhmRbozhAefFX5G4bFC5k6Z67VOSlIKx08fZ++xvQzoNqBR/d/x3h1sLt3M2jtDY+h32arqznz6xPWhb1xfv50Nq2qqWLF7Bd8b+72AydZas1cI8DKwWVX/t04bd7U7HSudjsHQJBbnLSY2OpZpI6Z5rdPUM141WsM/t/2TdUXrQmawdpY6iZIoRvao7+TZGKPz+qL1HD99HMegwNh7IIjKR1WrAFf2is1YGSicIjLblcECK3uFU0S2YO2K3WPfvxC4BZgsIhvs15V22TwR2SQiXwEO4Nv4CgZDI6jRGpbkL+GKYVfQNbar13ruZ7waw1fFX3G43DoSku3MbrqgzcBZ6mRwwmA6xnSsV5beJ52tB7ZyrOKYh5Zn4vLvmTRwUsBkC6qfj70N/kGdey+4XX8KDPPQbiWebUao6i0BFtMQoazcs5KiE0XMTJnZYL3uHbvTu0vvRisf1w92dNJoFjsXk5WR1eKhe50lznpLLhfpfdNRlC+LvmRScsNKJacgh9FJo+nVpVfAZDMezoaIJduZTYd2Hbh6+NU+66YkNX7Hy+WQ97PzfsbWg1vZVFL/IG8wqaiqYPuh7fW22V24PJ19Lb0qqytZuWdlwLbYXRjlY4hIqmuqeTP/Ta4adhVd2nfxWT8lKYX80nxqtMZnXVf/y3cvJ2NgBtePup5oiWZx3uLmit0oth3cRlVNldeZT8/OPenftb/PYxZr96/lZOVJo3wMhkCwfPdyik8W+1xyuUjpmUJZZRm7j+z2q36tQ94gB0mdk5g8aDLZ+dkt6ldW90yXJ/wxOrv8e3wtzRqLUT6GiCTbmU2nmE5cOexK35U584yXP7h+sK7ZQmZKJjsO7WjRnFnOEifREs2IxBFe66T3TWf7oe0cKT/itU5OQQ5je46lR6ceXus0BRPJMML56OuPGBg/kOGJw0MtSsB4K/8tth7c2mCdJflLmDZ8mt/Of6OTrKOFzhKnXzaiug5500dO58f//DGLnYs5u8/Zfo3pzpv5b3LRgIsaZfB1ljoZ2n0ose1ivdZJ72tlMl5ftJ7JgybXKz9dfZpVe1bxw3N+2GiZfWGUTwRzrOIY0xZNI6VnCmt/uDYskigWHS9ixpIZaD1n+jOJlmhuT/M/BGt8h3j6xfXza+bjcsj77tjv1t5L7JTIJYMvIduZzeNTHm/Ud71893JmLJnBT9J/wp+u+pPf7Tyd6aqLy9N57f61HpXPmn1rOFV1KqD+PS6M8olg3tv6HhXVFawvWs8/tv7D69mmtsRbm99CUb6a/RUjenhfbkRJFO2iGvff398zXrUOeXUMtJmjM/n+P77PuqJ1tTMOf3gk9xEA3tz8Jn+84o9ER0X7bFNeVc6OQzu4MeXGBusldkpkUPwgr3afnF05CMLFAy/2W15/MTafCCY7P5uzup7FsO7DeCT3Eb93cloz2c5sxvYcy9heY2kf3d7rq7GKB6wdr80HNlNdU91gPW8Bt64beR0xUTGN2vXK2ZVDbkEukwdNpuRkCct3L/er3dYDW6nRmgaNzS7G9x3vdccrpyCHcb3H0b1jd79l9hejfCKUo+VH+feOf5M5OpNfX/xrNhZv5J0t74RarGax79g+Vu5ZWS8aYaBISUqhvKqcXUd2NVjPm0NeQscELh1yqd+7XqrKI7mP0DeuL0tmLKFzTGcWO/1TXK4ZmrdtdnfS+6Sz8/BODp06dMb98qpyPi38NOBb7C6M8olQ3t36LqerT5OZkslNY29ieOJwsnKz2vTs5838N1GUGaNnBKV/f854+XLIy0zJZM/RPXy+73OP5e58susTVuxZwS8n/pLuHbszbcQ03tr8FlU1VT7bOkudtItqx7DEegcI6uFaAq7bf+bs5/PCzymvKjfKxxBYsp3ZDOg2gPP6nUe7qHY8MukRNpVs4q38t0ItWpPJzs9mXK9xDdp6mkPtjlcDRmeXQ563mDfXjriW9tHtfZ71UlUezn2Ys7qexR3n3AFYNqMDZQfILcj1Kauz1MnwxOG0j/Ydo/ucPufUyu5OTkEOURLFRQMv8tlHUzDKJwI5fOowH379IZmjM2t3XWamzGRUj1FkLcvyadNojew9upfVe1f77TTYFOJi4xjQbUCDysdXwK1uHbpx+dDLWZK/pMFZ5kc7P2L13tX8auKv6NDOCox2+dDL6dK+i1+HVBs601WXhI4JDO0+tF54jdyCXM7ufbbXOEfNxSifCOSdLe9QWVPJzDHf/lCjo6J5ZNIj5JfmsyR/SQilaxoumWekBGfJ5cLXGS9/HPJmpsyk8Fghn+6tn50WvrX1DOg2gO+f/W2cvI4xHbl2xLW8tfktKqsrvfZfVlnGzsM7/VY+UN/T+VTlqaDaeyDIykdELheRrSKyQ0Qe8FCeICJv2yFR14jIGPt+fxHJEZHNIuIUkXvc2nQXkY9EZLv9HriscRFCdn42g+IH1fp4uJiRMoOUpBTmLJvTqNnP5tLN/ODdH9QzWLYk2c5szulzDkO7Dw3qOClJKWw5sMWj3aWiqoJVe1b5/MFOGz6N2OhYrzOYf+/4N58VfsaDFz1Yz0EwMyWTQ6cO8cmuT7z2v+XAFhT16ePjzvg+49lzdA+lJ63Inp8Wfsrp6tNB8e9x0RazVzwALFXVYVjB5+spNYN3DpYd5OOdH5OZklnP0S1Konhk0iNsObClUbsqGQsymL9hfsjsRQVHCvh83+ceA8AHmpSeKZyuPs3Xh76uV/bF/i/8csiLi43jymFXsiR/ST0l75r1DOw2kFlps+q1vXTIpXSN7drg0qt2p8uPbXYXtUZne8s9Z1cO0RLtd4jVptAWs1dcC7gSei8ArgviM4Qd72x5h6qaKq/b0TeMvoGxPccyZ9kcn7sqm4o3MXnBZKIlmsSOibX2jpZmidNacgVri92dhs54NcYhb2bKTIpOFLFq76oz7n+w/QO+2P8Fv7741x6NxR3adeDaEdfy9pa3OV192mPfzhIn7aPbN2oWWNfonLs7l/F9xzcYZK25tMXsFb1UtQjAfu/paXBf2SsilcXOxQztPpSze3s+XxQlUWRlZLHt4DYWbVrktZ+N32xk8sLJxETHkDsrl6lDppJbkBuSbCDZ+dmc2/dcBiUEPya3K/Gep+32xjjkXTX8Kjq263jGDMY16xmcMJhbx93qte3MlJkcLj/Mxzs/9ljuLHUyInFEoxwpu8Z2ZUTiCNbuX0tZZRmfF34eVHsPBFf5+Ju9IsHOXnE3/mWv8AtVfVFV01U1PSkpqVGChyulJ0v5ZNcnZ+xyeeK6kdcxrtc4frP8Nx5nPxu+2cDkhZPp0K4Dy2YtY3jicBzJDopOFLHtYOCziTbEzsM7Wbt/bYvMegA6t+/MoPhB9WY+5VXlrN672u8fbJf2Xbh6+NW8mf9m7dLrvW3vsa5oHQ9d9BAx0TFe204dMpVusd28Lr2cpc5GLblcuIzOq/asorKmsk0rH7+yV6jq7aqahmXzScJH9gqg2BVE3n4vCdoThBlvb3mbaq32+UONkijmZMxhx6EdvPbVa2eUrS9az+QFk+kc05nc23Jrp/au/6gtvfRyLbmC5VjoCU9nvD4v/JyK6opG/WAzUzIpPlnM8t3La2c9QxKGcMu4hiMFt49uz/RR03lnyztUVFWcUXbi9AkKjhQ0aqfLxfg+49l3fB9v5L1Bu6h2XDjgwkb30RjaXPYKu4/b7OvbgHeD9gRhRrYzm+GJw0ntleqz7jUjruGcPufw6PJHa7d11+1fx5SFU4iLjSN3Vi5Dug+prT+0+1D6xvX1ywEukCx2LmZCvwkMjB/YYmOmJKWw7eC2M7a7m+KQd+WwK2uPTLyz5R02fLOBhyc97NdyKXN0JkcrjvLh1x+ecX9z6WaARu10uXAZnV/b9Brn9j3XrwiPzaEtZq+YC0wVke3AVPuzwQfFJ4rJKcjxueRyISLMyZjDzsM7WbhxIV/s+4IpC6cQ3yGeZbOWMThhcL36jmRHi9p9th/czpfffBlUx0JPjOk5hsqaSrYf2l57L6cgp9EOeZ1iOtUemchalsXwxOFnhOFoiCmDp5DQIYHs/DOXXrXRC5sw8zm7z9kIYm2xB3nJBW0ze8VBYEpgJQ1//r7579RozRmOhb64athVnNv3XB7OfZgTp09YO1q35XidZTiSHby+6XU2H9hcexQhmLgcC78z+jtBH8sd1w/bWeJkdNJoTlWe4rPCz/jZeT9rdF+ZozN5I+8NDpQd4LXpr/ltJG4f3Z7rR11PtjOb8qryWi/ovJI8OrTrUO+Pgz90ad+FUUmjyC/ND6p/jwvj4RwhZOdnM6rHqEb9RRQRsjKy2H98P0mdklg2a1mDyxvXkYKWWnplO7O5oP8F9O/W33flADKyx0iiJKp2ltEch7zLh15OXPs4RvYYyY1jGo69U5fMlEyOnz7Of3b8p/aes9TJyB4j/Yr544lz+55LTFQMF/S/oEntG4NRPhHANye+YVnBMo+Ohb64YugVZH8nmxW3r/D5Ix+cMJj+Xfu3iNF564GtbCze2CKOhXXpGNORwQmDa5VPcxzyOsZ05J0b32HJjCWNVhiOZAeJHRPPcAhtzJkuT2RlZPH+d9+nU0ynJvfhLyaSYQTgCjXRlO1oEfH7vJSI4Bjk4IPtH1CjNURJ8P62ZTuzEaTFl1wu3M945RTkNMshz1P4Un+IiY7h+lHX87dNf+NU5SkqayrZe2xvs5RPcnwyyfHJTW7fGMzMJwLIdmYzpueYFrHDOJIdHCg7QH5pflDHyc7PZuKAifTrWtdvtWVISUph+8HtHD51mDX71rSIgdYTM1NmcrLyJP/a8a/a77wpPj6hwCifMKc2ul8LLU9cdh9XKNFgkF+aT15JXos5FnpiTM8xVGs1r2x4pUUc8rwxKXkSSZ2SWOxcXOt13ZRt9lBglE+Y4wqoHuxQEy5c0/Zg2n1cS64bRt0QtDF84ZpdPPfFcy3ikOeNdlHtuGHUDby/7X3W7FtDp5hOLbZsai5G+YQ5i52LSe2VysgeI1tsTEeyg2W7lwUlJKuqku3MZlLypNqcWKFgROIIoiWarw9/3SIOeQ2RmZJJWWUZC79ayKgeo4JqawskbUNKQ5Noieh+nshIzuDQqUNsKt4U8L6dpU42H9gckl0ud2LbxdY7WhIqLh54Mb0696K8qrzN2HvAKJ+w5s38N4GWPfcEwT3ntThvMVESxfWjrg94343F9UNvCYe8hoiOiq7d9WvOTldLY5RPmJJXksfjKx8nvW+6XxkMAkn/bv0ZkjAk4MqnoqqCBRsX4Eh2NCptcLCY0G8C3WK7tYhDni9uTr0ZsGRqKxjlE4Z8VfwVjgUOYqJjeP3610MiQ0ZyBst3Lw9oMPr5X85n77G9/M+F/xOwPpvDveffy7a7t7WIQ54vzj/rfArvLWRS8qRQi+I3PpWPiFwt0kYsWAYryNeCycRGx5J7Wy7DE4eHRA5HsoMj5UfYWLwxIP1VVFXw2MrHuKD/BUwdPDUgfTaXmOgYenb2GMsuJITK56mp+KNUbgS2i8g8ERkVbIEMTefLoi+ZvHAynWI6sWzWshZfbrkTaH+fv6z/C4XHCvlNxm8afUTE0DrxqXxU9WasMKZfA6+IyKd2iNI4X22bmr3CLpsvIiUiklenTZaI7PMQaiOiccXa6dK+S71YO6GgX9d+DOs+jNzduc3uq7yqnMdWPsZFAy5q8lEEQ+vDr+WUHcL0Lawg8H2A6cB6EbnbW5tmZq8A+CtwuZfun1bVNPv1gZc6EcPa/Wu55NVL6Brb1WOsnVDhSHawfPdyv9L7NsSL615k//H9zMmYY2Y9YYQ/Np9pIvI28AkQA5ynqlcA44D7G2ja5OwV9uflQOgSQbUR1uxbwyULLyGhQwLLZi1rVd6tjkEOjlUc48uiL5vcx6nKUzy+8nEykjNCvqVtCCz+zHxmYM00UlX1CVUtAVDVMuD7DbQLSPYKL/zUXqrN95Y0MBKyV3xW+BlTX51KYqdEcmfltmgoUX+YNNDaeWlOfJ8X1r7ANye+YU7GnABJZWgt+KN8HgHWuD6ISEc7nQ2qurSBds3OXuGF54EhQBpQBDzlqVK4Z69YX7SeS1+9lKROSeTelsuAbgNCLVI9+sT1YWSPkU329zl5+iRzV81l8qDJfuXCMrQt/FE+SwD3QzrV9j1fNCt7hTdUtVhVq1W1BngJa3kXcTz3xXNESRTLZi1r8Uh+jcGR7GDFnhUN5hb3xvNrn6fkZImZ9YQp/iifdrbNBgD7un4qxfo0OXtFQ5260ubYTAfyvNUNZ/JK8ji7z9mt3rcjIzmDE6dPsL5ofaPanTx9knmr5jF18NSgpuw1hA5/lE+piFzj+iAi1wIHfDVqZvYKRGQR8CkwQkQKReQHdtE8EdkkIl8BDuBeP54hrFBV8kvz28Q5nlp/n0Yuvf70xZ8oLSs1s54wxp8wqrOB10XkWSw7zl6sJZJPmpq9wi67ycv9hjOqRQB7j+3l+OnjbUL59Ozck5SkFHIKcnhgYj1XL48crzjOvFXzuHzo5fxX//8KsoSGUOFT+ajq18D5dupiUdXjwRfL0BBtLWKdI9nB/A3zOV19mvbRvlfsz655loOnDppZT5jjl5OhiFwF/AS4V0QeFpGHgyuWoSFcgcvbSuyWjOQMyirLWLt/rc+6xyqO8cTqJ7hy2JWc1y8i9xIiBn+cDF8AZmJthQuW30/rciiJMJylTnp36U33jt1DLYpfuE5a+3PO64+f/5HD5YfNrCcC8Gfmc4Gq3gocVtU5wH9x5ha6oYVxljYvN1NL06NTD1J7pfo0Oh8tP8pTnz7FtOHTavOGG8IXfwzO5fZ7mYj0BQ4Cg4InkqEharSG/NJ87jj7jlCL0igyBmbw3NrnSHnOu9I8XnGcI+VHyMrIajnBDCHDH+XznojEA08A67G8lF8KplAG7+w+spuyyrI2Y+9xMTt9NiVlJT4Pmd7V5y7O6XNOC0llCCUNKh87iNhSVT0CvCUi7wMdVPVoSwhnqI8rRW9b2elyMSppFItuWBRqMQytiAZtPvYRhqfcPlcYxRNaXNvsLZF91GAIJv4YnD8UkRvEBFJpFeSV5tEvrh/xHeJDLYrB0Cz8sfncB3QGqkSkHGu7XVW1a1AlM3jEWeJsc/Yeg8ET/ng4+wyXamgZqmuq2Xxgc+15KYOhLeNT+YiIx0AqdqRBQwuy68guKytlG/LxMRi84c+y6/+5XXfAip+zDvAZyVtELseKyxwN/EVV59YpTwDmYwUHKwe+r6p5dtl84GqgRFXdA8t3BxYDyUABkKmqh/14jjZPWzvTZTA0hD/ZK6a5vaYCY4BiX+2CGED+Aazt/2FY8Z/9OyodBri22c1OlyEcaEoywEIsBeSLYAWQvxZYYF8vAK5r7AO0VfJK8hjQbQBxscYMZ2j7+GPz+T++jb0chRU72Z80lJ4CyNdNJO0KIL+yTgD5hmZWvVS1CEBVi0Sk9aSMDDJt7UyXwdAQ/th83OMgVAGLVHWVH+38DSD/jB1AfhP+BZD3CxG5E7gTYMCA1hdcvbFU1VSx5cAWLh18aahFMRgCgj/K502gXFWrwbLliEgnO3VOQ/gVQB643e5XsILHNxhAHigWkT72rKcPUOKpkqq+CLwIkJ6eXlfptTm+PvQ1p6tPGx8fQ9jgj81nKdDR7XNH4GM/2gUlgLzdx2329W3Au37I0uZpq2e6DAZv+KN8OqjqCdcH+7qTr0ZBDCA/F5gqItuBqfbnsMe1zT6qx6gQS2IwBAZ/ll0nReQcVV0PICLjgVP+dB6kAPIHgSn+jB9O5JXmMSh+EJ3bdw61KAZDQPBH+fwcWCIiLntNH6ywqoYWxJzpMoQb/pzt+kJERgIjsHawtqhq49NPGppMZXUl2w5u4+rhV4daFIMhYPgTQP4uoLOq5qnqJqCLiPwk+KIZXGw/tJ3Kmkrj42MIK/wxOP/QjmQIgH2O6odBk8hQD3OmyxCO+KN8otwDidlntvzJ1W4IEM5SJ1ESxcgeI0MtisEQMPwxOP8HyLbzdylW+uR/BVUqwxk4S50MThhMx5iOvisbDG0Ef5TPL7COKfwYy+D8JdaOl6GFyCvJM/YeQ9jhT0iNGuAzYCeQjuVjsznIchlsKqoq2H5wu1E+hrDD68xHRIZjHYm4CStR4GIAVXW0jGgGgG0Ht1Gt1cbHxxB2NLTs2gKsAKap6g4AEbm3RaQy1GLOdBnClYaWXTcA3wA5IvKSiEzBc5gMQxBxljiJlmhGJI4ItSgGQ0DxqnxU9W1VnQmMBHKBe4FeIvK8iJigMi2Es9TJ0O5DiW0XG2pRDIaA4o/B+aSqvq6qV2PF5NlABMVNDjV5JXnG3mMISxoVw1lVD6nqn1XVZ+YKsLJXiMhWEdkhIvUUlogkiMjbIvKViKwRkTG+2opIlojsE5EN9uvKxjxDW6K8qpyvD39tdroMYUlTAsj7RXOyV/jR9mlVTbNfHxCmbDmwhRqtMcrHEJYETfnQvOwV/rQNe8yZLkM4E0zl4yl7Rb86dVzZK6iTvcJX25/aS7X5duLBsMRZ6qRdVDuGJXqMt2YwtGmCqXz8zV6RYGevuJtvs1c01PZ5rAynaUAR8JTHwUXuFJG1IrK2tLS00cK3BpylToYnDqd9tDnHawg//Dnb1VSak72ik7e2qlqb00tEXgLe9zR4OGSvcJY4OafPOaEWw2AICsGc+TQne4XXtna6HBfTgbwgPkPIKKssY+fhncbYbAhbgjbzUdUqEXFlr4gG5ruyV9jlL2Blr1goItVAPvCDhtraXc8TkTSsZVgB8KNgPUMo2Vy6GUWNj48hbAnmsqu52SvqtbXv3xJgMVsl5kyXIdwJ5rLL0AycJU7aR7dnaPehoRbFYAgKRvm0UpylTkYkjqBdVFAnpwZDyDDKp5XiLDV5ugzhjVE+rZATp09QcKTA7HQZwhqjfBrBkfIjnDh9wnfFZpJfmg9glI8hrDHKx0/W7FtD8h+SufO9O4M+1sZvNgJmp8sQ3hjl4wefFX7G1FencrTiKP/a8S+qa6qDOl7u7lx6d+ltdroMYY1RPj5YtWcVl756KUmdknh8yuMcKT/CV8VfBW08VSW3IJeM5AzccjUaDGGHUT4NsHz3ci577TJ6d+nNslnLuCXV8m/MKcgJ2pjbD21n//H9ZAzMCNoYBkNrwCgfL+QW5HLF61fQv1t/ls1aRr+u/ejXtR/Dug8LqvLJ2WX17RhkMhQZwhujfDywdOdSrnz9SpLjk8m5LYc+cd+eZXUkO1i+ezlVNVVBGTunIIe+cX0Z1t3E8DGEN0b51OHDrz/k6kVXM6T7EHJuy6F3l95nlDsGOThWcYwN32wI+Ngue48j2WHsPYawxygfN/61/V9cs+gaRiSOIOe2HHp27lmvzqSBk4Bvl0eBZMuBLRSfLCYjOSPgfRsMrY2gKp8gZa/oLiIfich2+z0gYVTf3/Y+1y2+jtFJo1l661J6dOrhsV6fuD6M7DEyKHYfV5+OZGPvMYQ/bTF7xQPAUlUdhhV8PiA5xJbkLyG1VypLb11KYqfEBus6kh2s2LOCyurKQAxdS05BDv279mdwwuCA9mswtEbaYvaKa4EF9vUC4LpACPvyNS+z9NalJHT0PZFyJDs4cfoE64vWB2JowM3eM8jYewyRQVvMXtFLVYsA7Pf6hpkm0C6qHV1ju/pVd1KybfcJ4NLLWerkQNkB499jiBjaYvYK/wYPYvaKnp17kpKUElDlY/x7DJFGMJWPX9krVPV2VU3DsvkkYWWvaKhtsSuIvP1e4mlwVX1RVdNVNT0pKSkAj3MmjmQHK/es5HT16YD0l1OQQ3J8MsnxyQHpz2Bo7bS57BX2+2329W3Au0F8Bq84Bjkoqyzji31fNLuvGq1h2e5lZpfLEFEETfmoahXgykCxGch2Za9wZbDAyl7hFJEtWDtb9zTU1m4zF5gqItuBqfbnFufigRcD1jGM5rKpeBOHTh0y/j2GiKItZq84CEwJrKSNp0enHqT2SiWnIIcHL36wWX0Z/x5DJGI8nJuBI9nBqr2rqKiqaFY/OQU5DEkYQv9u/X1XNhjCBKN8moEj2UF5VTmf7/u8yX1U11SzrMDYewyRh1E+zeDigRcjSLPsPhuLN3K04qix9xgiDqN8mkFCxwTSeqc1y9/H+PcYIhWjfJqJI9nBp3s/pbyqvEntcwpyGJ44nL5xfQMsmcHQujHKp5k4BjmoqK7g072fNrptVU0Vy3cvN/YeQ0RilE8zuWjARURJVJOWXuuL1nP89HFj7zFEJEb5NJNuHbpxTp9zmmR0drUxyscQiRjlEwAcyQ4+K/yMssqyRrXLKchhVI9R9UK1GgyRgFE+AcCR7KCyppLVe1f73aayupIVu1cYe48hYjHKJwBMHDCRaIluVFzntfvXcrLypFlyGSIWo3wCQFxsHOl908ndnet3G2PvMUQ6RvkECEeygzX71nDi9Am/6ucU5DCm5xiSOgc+1pDB0BYIdfaKbiLynohsFBGniNzuVnaPiOTZ93/udj9LRPaJyAb7dWUwn8FfHIMcVNVUsWrPKp91T1efZtXeVcbeY4hoQp294i4gX1XHARnAUyLS3k6h80OsQPLjgKtFxD30xtOqmma/6oXdCAUX9r+QdlHt/PL3WbNvDWWVZUb5GCKaUGevUCBOrHQNXYBDWDGcRwGfqWqZHVhsGTA9iLI2m87tO3Nev/P88vfJLchFkNqAZAZDJBLq7BXPYima/cAm4B5VrQHygItFJFFEOgFXcmZM55/aiQbnByppYCBwJDtYu38txyuON1gvpyCH1F6pPvODGQzhTDAjGfqTgeIyYAMwGRgCfCQiK1R1s4j8HvgIOIGVYqfKbvM88Kjd16PAU8D36w0ucidwJ8CAAQOa+yx+4Uh28LsVv2PuyrmM7DHSYx1FWb13NbPHz/ZYbjBECsFUPj6zVwC3A3NVVYEdIrILGAmsUdWXgZcBROQxuz9UtdjVWEReAt73NLiqvgi8CJCent6otDtN5YL+F9AtthuPrXzMZ90rh7UKO7nBEDKCqXxqM1AA+7AyUHy3Tp09WPGYV9iZSkcAOwFEpKeqlojIAKzEgv9l3+/jShqIZQfKC+IzNIqOMR3Zdc8uDp061GC9Du060K9r3RWowRBZBE35qGqViLgyUEQD813ZK+zyF7CWTX8VkU1Yy7RfqOoBu4u3RCQRqATuUtXD9v15IpKGtewqAH4UrGdoCgkdE/xKuWwwRDpirXjCm/T0dF27dm2oxTAYIg4RWaeq6Z7KjIezwWAICUb5GAyGkGCUj8FgCAlG+RgMhpAQEQZnESkFdodajjr0AA74rNV6MPIGl3CVd6CqegzdEBHKpzUiImu97QK0Roy8wSUS5TXLLoPBEBKM8jEYDCHBKJ/Q8WKoBWgkRt7gEnHyGpuPwWAICWbmYzAYQoJRPi2MiBSIyCY7/nSrO3BmB2grEZE8t3vdReQjEdluv7eqk7NeZG6Vsb5FpL+I5IjIZjs++T32/Vb5HTcgb7O/X7PsamFEpABIdzu936oQkYuxArgtVNUx9r15wCFVnWsnAkhQ1V+EUk53vMicBZxQ1SdDKVtdRKQP0EdV14tIHLAOuA6YRSv8jhuQN5Nmfr9m5mM4A1VdjhVL251rgQX29QKs/3ytBi8yt0pUtUhV19vXx4HNWOGFW+V33IC8zcYon5ZHgQ9FZJ0d6rUt0MsVwM1+7xliefylVcb6diEiycDZwOe0ge+4jrzQzO/XKJ+W50JVPQcrpdBd9pLBEHiex4oLngYUYcX6bjWISBfgLeDnqnos1PL4woO8zf5+jfJpYVR1v/1eAryNlWKotVNsr/1dNoCSEMvjE1UtVtVqOxvKS7Si71lEYrB+yK+r6t/t2632O/YkbyC+X6N8WhAR6Wwb7RCRzsCltKIY1A3wD+A2+/o24N0QyuIXrh+yTauJ9W3nqHsZ2Kyq/+tW1Cq/Y2/yBuL7NbtdLYiIDMaa7YAVP/tvqvq7EIpUDxFZhJU9tgdQDDwCvANkAwOwgv7PUNVWY+D1InMG1pKgNta3W+KBkCEiE4EVWHnqauzbv8Kyo7S677gBeW+imd+vUT4GgyEkmGWXwWAICUb5GAyGkGCUj8FgCAlG+RgMhpBglI/BYAgJRvmEGSKiIvKU2+f77UOWgej7ryLynUD05WOcGfYp6pw695Pt57vb7d6zIjLLR3+zReRWH3VmicizXspONEJ8g58Y5RN+VADXi0iPUAvijohEN6L6D4CfqKrDQ1kJcI+ItPe3M1V9QVUXNmL8gCEi7UIxblvAKJ/woworxOW9dQvqzlxcf9FFJENElolItohsE5G5IvI9EVljxx4a4tbNJSKywq53td0+WkSeEJEv7IOGP3LrN0dE/oblpFZXnpvs/vNE5Pf2vYeBicALIvKEh+crBZbyrTewe39DROTf9qHdFSIy0r6fJSL329fn2jJ+asvs7pnb126/3Q4j4t73UyKyXkSWikiSfS9NRD6z+3vbdbhSRHJF5DERWYalKGfYz7hRRJZ7eKbIRFXNK4xeWHFtumJ5nXYD7gey7LK/At9xr2u/ZwBHgD5ALLAPmGOX3QP8wa39v7H+aA0DCoEOwJ3AQ3adWGAtMMju9yQwyIOcfbE8eZOwvL0/Aa6zy3KxYh7VbZOM5cY/CNgCRAPPArPs8qXAMPt6AvCJfZ0F3G9f5wEX2NdzgTz7ehaw0/7OOmDleetvlynwPfv6YeBZ+/orYJJ9/Ru37ykXeM5N7k1AP/s6PtT/R1rLy8x8whC1Th0vBH7WiGZfqBW7pQL4GvjQvr8J60fvIltVa1R1O9aPdSTWGbVbRWQD1jGBRCzlBLBGVXd5GO9cIFdVS1W1Cngd8OuEv93fGuC7rnv2qesLgCW2HH/GUqa41YkH4lR1tX3rb3W6XqqqR1W1HMgHBtr3a4DF9vVrwEQR6YalSJbZ9xfUkX+x2/Uq4K8i8kMshWnA+otjCE/+AKwHXnG7V4W91LYPDLrbTSrcrmvcPtdw5v+TuudxFBDgblX9j3uBiGRgzXw8IT7k98VjwJuAaxkTBRxR1bQG2vga0/07qMb778OfM0m1z62qs0VkAnAVsEFE0lT1oB99hDVm5hOmqHUoMRvLeOuiABhvX18LxDSh6xkiEmXbgQYDW4H/AD+2Qy8gIsPtU/sN8TkwSUR62Mbom4BlPtrUoqpbsGYnV9ufjwG7RGSGLYOIyLg6bQ4Dx0XkfPvWjX4OFwW4bGXfBVaq6lHgsIhcZN+/xZv8IjJEVT9X1YexUgz393PcsMbMfMKbp4Cfun1+CXhXRNZg2Ue8zUoaYivWj6wXMFtVy0XkL1hLs/X2jKoUH2FAVbVIRH4J5GDNSD5Q1caGkfgd8KXb5+8Bz4vIQ1iK9Q1gY502PwBeEpGTWLaZo36McxJIEZF1dv2Z9v3bsAzjnbCWoLd7af+EiAzDes6lHmSKSMypdkNEISJdVNW1y/cAVnD0e0IsVkRiZj6GSOMqe8bVDmtHa1ZoxYlczMzHYDCEBGNwNhgMIcEoH4PBEBKM8jEYDCHBKB+DwRASjPIxGAwhwSgfg8EQEv4/clWJLkwFCywAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 288x216 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x216 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "The best value is K= 17 with an accuracy of: 0.92843 and an RMSE of 0.26753\n",
      "Accuracy: 0.9284294234592445\n",
      "Classification Error: 0.07157057654075549\n",
      "Mean Cross Val (Accuracy) Score 0.9051820895522387\n",
      "Mean Classification Error: 0.09481791044776133\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x216 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Area Under Curve: 0.982024567496302\n"
     ]
    }
   ],
   "source": [
    "#KNN Classifier\n",
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "accuracy = {}\n",
    "error = {}\n",
    "for K in range(1,25):\n",
    "    K = K+1\n",
    "    knn = KNeighborsClassifier(n_neighbors=K)\n",
    "    knn.fit(X_train, y_train)\n",
    "    y_pred = knn.predict(X_test)\n",
    "    accuracy_value = round(metrics.accuracy_score(y_test, y_pred),5)\n",
    "    accuracy.update({K: accuracy_value})\n",
    "    error_value = round((mean_squared_error(y_test, y_pred, squared=False)),5)\n",
    "    error.update({K: error_value})\n",
    "    print('Accuracy value for k=',K,'is:', accuracy_value, ' and RMSE value is:', error_value)\n",
    "import matplotlib.pyplot as plt\n",
    "plt.figure(figsize=(4,3))\n",
    "plt.title('kNN accuracy plot')\n",
    "plt.plot(accuracy.keys(), accuracy.values(), label = 'Testing Accuracy', color=\"green\")\n",
    "plt.legend()\n",
    "plt.xlabel('Number of Neighbors')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.show()\n",
    "plt.figure(figsize=(4,3))\n",
    "plt.title('kNN RMSE plot')\n",
    "plt.plot(error.keys(), error.values(), label = 'RMSE', color=\"red\")\n",
    "plt.legend()\n",
    "plt.xlabel('Number of Neighbors')\n",
    "plt.ylabel('Testing RMSE')\n",
    "plt.show()\n",
    "print()\n",
    "print('The best value is K=', max(accuracy, key=accuracy.get), 'with an accuracy of:', max(accuracy.values()), \n",
    "      'and an RMSE of', min(error.values()))\n",
    "\n",
    "feature_cols = ['S','K','tau','r']\n",
    "X = option_train[feature_cols]\n",
    "y = option_train[\"BS\"]\n",
    "model = KNeighborsClassifier(n_neighbors=17)\n",
    "model.fit(X_train, y_train)\n",
    "y_pred = model.predict(X_test)\n",
    "print(\"Accuracy:\",metrics.accuracy_score(y_test, y_pred))\n",
    "print(\"Classification Error:\",1-metrics.accuracy_score(y_test, y_pred))\n",
    "cv = RepeatedKFold(n_splits=5, n_repeats= 100, random_state=1)\n",
    "print(\"Mean Cross Val (Accuracy) Score\", np.mean(cross_val_score(model, X, y, cv=cv)))\n",
    "print(\"Mean Classification Error:\", 1 - np.mean(cross_val_score(model, X, y, scoring='accuracy', cv=cv, n_jobs=-1)))\n",
    "y_pred_proba = model.predict_proba(X_test)[:,1]\n",
    "fpr, tpr, _ = metrics.roc_curve(y_test, y_pred_proba)\n",
    "roc_auc = auc(fpr,tpr)\n",
    "plt.figure(figsize=(4,3))\n",
    "plt.plot(fpr, tpr, color='orange',\n",
    "lw=2, label='ROC curve (area = {0:.4f})'.format(roc_auc))\n",
    "plt.plot([0, 1], [0, 1], color='blue', lw=2, linestyle='--')\n",
    "plt.xlim([0.0, 1.0])\n",
    "plt.ylim([0.0, 1.05])\n",
    "plt.xlabel('False Positive Rate')\n",
    "plt.ylabel('True Positive Rate')\n",
    "plt.title('ROC curve')\n",
    "plt.legend(loc=\"lower right\")\n",
    "plt.show()\n",
    "print('Area Under Curve:', roc_auc)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fb35f388",
   "metadata": {},
   "source": [
    "### Linear Discriminant Analysis "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 433,
   "id": "fc35a5ca",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best Parameters:  {'solver': 'svd'}\n",
      "Accuracy Score:  0.9095683036361002\n"
     ]
    }
   ],
   "source": [
    "#Linear Discriminant Analysis\n",
    "from sklearn.discriminant_analysis import LinearDiscriminantAnalysis\n",
    "model = LinearDiscriminantAnalysis()\n",
    "forest_params = {'solver' : ['svd', 'lsqr', 'eigen']}\n",
    "clf = GridSearchCV(model, forest_params, cv = 10, scoring='accuracy')\n",
    "clf.fit(X_train, y_train)\n",
    "print('Best Parameters: ', clf.best_params_)\n",
    "print('Accuracy Score: ', clf.best_score_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 434,
   "id": "b3f11b3b",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.9065606361829026\n",
      "Classification Error: 0.09343936381709739\n",
      "Mean Cross Val (Accuracy) Score 0.9122388059701492\n",
      "Mean Classification Error: 0.08776119402985083\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x216 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Area Under Curve: 0.9757219113769374\n"
     ]
    }
   ],
   "source": [
    "#Linear Discriminant Analysis \n",
    "from sklearn.discriminant_analysis import LinearDiscriminantAnalysis\n",
    "from sklearn.model_selection import train_test_split\n",
    "model = LinearDiscriminantAnalysis(solver = 'svd').fit(X_train, y_train)\n",
    "y_pred = model.predict(X_test)\n",
    "print(\"Accuracy:\",metrics.accuracy_score(y_test, y_pred))\n",
    "print(\"Classification Error:\",1-metrics.accuracy_score(y_test, y_pred))\n",
    "cv = RepeatedKFold(n_splits=5, n_repeats= 100, random_state=1)\n",
    "print(\"Mean Cross Val (Accuracy) Score\", np.mean(cross_val_score(model, X, y, cv=cv)))\n",
    "print(\"Mean Classification Error:\", 1 - np.mean(cross_val_score(model, X, y, scoring='accuracy', cv=cv, n_jobs=-1)))\n",
    "y_pred_proba = model.predict_proba(X_test)[:,1]\n",
    "fpr, tpr, _ = metrics.roc_curve(y_test, y_pred_proba)\n",
    "roc_auc = auc(fpr,tpr)\n",
    "plt.figure(figsize=(4,3))\n",
    "plt.plot(fpr, tpr, color='orange',\n",
    "lw=2, label='ROC curve (area = {0:.4f})'.format(roc_auc))\n",
    "plt.plot([0, 1], [0, 1], color='blue', lw=2, linestyle='--')\n",
    "plt.xlim([0.0, 1.0])\n",
    "plt.ylim([0.0, 1.05])\n",
    "plt.xlabel('False Positive Rate')\n",
    "plt.ylabel('True Positive Rate')\n",
    "plt.title('ROC curve')\n",
    "plt.legend(loc=\"lower right\")\n",
    "plt.show()\n",
    "print('Area Under Curve:', roc_auc)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fcb5fa3e",
   "metadata": {},
   "source": [
    "### Random Forest Classifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "id": "057849f8",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best Parameters:  {'class_weight': 'balanced', 'max_depth': 13, 'max_features': 4, 'n_estimators': 100}\n",
      "Accuracy Score:  0.9351296537737216\n"
     ]
    }
   ],
   "source": [
    "#Random Forest Classifier\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "model = RandomForestClassifier()\n",
    "forest_params = {\n",
    "    'n_estimators': [100,500],\n",
    "    'max_depth' : [1,3,5,7,9,11,13,15],\n",
    "    'max_features': [4],\n",
    "    'class_weight': ['balanced']\n",
    "}\n",
    "clf = GridSearchCV(model, forest_params, cv = 10, scoring='accuracy')\n",
    "clf.fit(X_train, y_train)\n",
    "print('Best Parameters: ', clf.best_params_)\n",
    "print('Accuracy Score: ', clf.best_score_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 111,
   "id": "d1aafc9f",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.904572564612326\n",
      "Classification Error: 0.09542743538767395\n",
      "Mean Cross Val (Accuracy) Score 0.9292477611940299\n",
      "Mean Classification Error: 0.07075223880597015\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAARoAAADgCAYAAADczalxAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuNCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8QVMy6AAAACXBIWXMAAAsTAAALEwEAmpwYAAAuJUlEQVR4nO2dd3hVVdaH30UooXdQiEiX0KsKihQVUUDFAiLSLFhgHAcbn6Ayoo4FyyBFURFlFCyIbVAURgUBhaCRKtUIUUoAhdADWd8f+yRcQspNyMm9Sdb7PPfJKfvs87s3ub/ss/bea4uqYhiG4SdFQi3AMIyCjxmNYRi+Y0ZjGIbvmNEYhuE7ZjSGYfiOGY1hGL5jRmMYhu+Y0RRCRCRORA6JyH4R2S4i00SkTJoyHUTkfyKSKCJ7ReRTEWmcpkw5EXlRRLZ4dW309qvk7Tsywh0zmsJLL1UtA7QEWgH/l3JCRNoDXwIfAzWAOsDPwCIRqeuVKQ7MB5oA3YFyQAdgN3CuX6JFpKhfdRv+YUZTyFHV7cBcnOGk8Azwlqr+W1UTVXWPqo4GvgfGeGUGArWA3qq6RlWTVXWnqo5V1Tnp3UtEmojIVyKyR0R2iMhD3vFpIvJ4QLnOIhIfsB8nIg+KyArggIiMFpEP0tT9bxEZ722XF5HXRWSbiPwuIo+LSMTpfVLG6WBGU8gRkSjgcmCjt18K1zJ5P53i7wGXetuXAF+o6v4g71MWmAd8gWsl1ce1iIKlH9ADqABMB64QkXJe3RFAH+Adr+ybwDHvHq2AbsCt2biXkcuY0RRePhKRRGArsBN41DteCfd3sS2da7YBKfGXyhmUyYiewHZVfU5VD3stpR+ycf14Vd2qqodU9TfgR+Bq71xX4KCqfi8i1XHGeY+qHlDVncALwA3ZuJeRy5jRFF6uVtWyQGegEScM5E8gGTgznWvOBHZ527szKJMRZwGbcqTUsTXN/ju4Vg7AjZxozZwNFAO2ichfIvIX8ApQ7TTubZwmZjSFHFX9FpgGjPP2DwBLgOvTKd6HE48784DLRKR0kLfaCtTL4NwBoFTA/hnpSU2z/z7Q2Xv0680Jo9kKHAGqqGoF71VOVZsEqdPwATMaA+BF4FIRaentjwQGicjdIlJWRCp6wdr2wD+9MtNxX+pZItJIRIqISGUReUhErkjnHp8BZ4jIPSJSwqv3PO9cLC7mUklEzgDuyUqwqiYA3wBvAL+q6lrv+DZcj9lzXvd7ERGpJyKdsvmZGLmIGY2R8qV9C3jY2/8OuAy4BheH+Q0XVL1QVTd4ZY7gAsK/AF8B+4CluEewU2IvqpqICyT3ArYDG4Au3unpuO7zOJxJvBuk9Hc8De+kOT4QKA6swT0KfkD2HvOMXEYs8ZVhGH5jLRrDMHzHjMYwDN8xozEMw3fMaAzD8B0zGsMwfCffzYStUqWK1q5dO9QyDKPQsXz58l2qWjUn1+Y7o6lduzYxMTGhlmEYhQ4R+S2n19qjk2EYvmNGYxiG7/hmNCIyVUR2isiqDM6LiIz30j+uEJHWfmkxDCO0+NmimYZL8ZgRlwMNvNdQYLKPWgzDCCG+BYNVdYGI1M6kyFW4dJEKfC8iFUTkTG/2rRFI8nHYvwn+WgF710LykVArMgoZv/1R4bSuD2WvU01OTmYU7x07xWhEZCiu1UOtWrXyRFzQ7N8MK/8Jf630p349Bomb4PhBf+o3jExQhcnz7uS+dx4C7s9xPaE0GknnWLpTyVV1CjAFoG3btnk73fxwAvwxx33hTxYFWz+EbZ/njY5SUVChOZRvCsXK5s09jULNHzvLcPMjvZi7KKN8ZcETSqOJx6V3TCEK+CNEWk4mbiZs9XJzb/0w6/LVLoLGIyGyug9iBErXghKVfajbMNLngw/g9tthzx6oVAleeQWuTy/nYpCE0mg+AYaLyEzgPGBvWMRnkvbDD7ec+qhSqQ1UaHHysYiS0PAuKN8YwygojBwJTz/ttrt3h6lT4czTTBvmm9GIyAxc4usq3ho9j+KSRqOqLwNzgCtwy3wcBIb4pSVbbJ3lTKZIcejgJW4rVROqnB9aXYaRR1x+OUyYAM8+C3fcAZJekCOb+Nnr1C+L8woM8+v+2Wb/Zlh6J2z/0u23ehZqXRtaTYaRBxw5AnPnwpVXuv1OneC336ByLj6t57u5TrmGKmz7EhI3QOJ62PwGHPPWQpMiUPWC0OozjDxgxQq46SZYuRLmzYOLL3bHc9NkoDAbzaqxsPLRk4/V7AVNH4ZStaCkH4FdwwgPjh+HF16AUaPg6FGoXx/K+tiZWXiNZv1497PmlVDuHKh9I1RsGVJJhpEX/PYbDBoE337r9u+4A8aNg9LBrtCVAwqv0SR742LavwnFK4RUimHkFV9/DVdfDfv2QfXqrkfpivRW4cplCq/RGEYhpFkzKFXKxWKmTIEqVbK+JjcofEaTtA/i3oakvaFWYhh5wjffQIcOULy4M5aYGKhRI3e6rYOl8OWj2TgFlt3ltqWoGy9jGAWQAwdg2DDo0gUee+zE8Zo189ZkoDC2aI56LZkqHaDRP6BoqczLG0Y+ZOlSGDAA1q+HYsWgQoXQ6il8RpPCmd2h1nWhVmEYuUpSEjz5JIwd67qwmzaF//wHWrTI+lo/KbxGYxgFjD173PSBpUvd/ogR8MQTEBkZWl1gRmMYBYaKFd1M67POgmnToGvXUCs6QeEymuRjlkDKKFBs2+ZG9p59tgvwvvmm610KdUwmLYWn1yn5GMxpCr88H2olhpErfPihGxdz440uHgNQrVr4mQwU9BbN8SPw80Nw8HdIPgz71rnjpWvDmd1CKs0wcsrevfD3v7vWC0CZMpCYGJ4Gk0LBNJrkY7AnxmXHS9uCKd8YeqwOjS7DOE0WLICBA918pchIN0fprrvyflxMdimYRvPjvScmTQK0fApKne22Lf2DkU95+GHXi6QKbdvC9OnQqFGoVQVHwTKaQ9vh+yGw7QuXU6ZKe6jWCRo/GGplhnHalCvnWi6jRzvTKVYs1IqCp2AZzXd9IGGh225wF7R9KbR6DOM0SE52I3tTWi0jRkC3bqEffJcTCk6vkyafMJnOc6D1iyGVYxinw2+/uRnW7dtDfLw7FhGRP00GCoLR7I+D2TVgRoTbL14Rql8MRSJCKsswcoKqi700b+5mXRcv7kwnv5P/jea3GXDIW6Ulshp0mAERNiPbyH/s3g19+rhepX37XIKqVavgggLQf5F/YzSqcHAL/P6Z27/gXTdJUvK/dxqFj2++cQPvtm1z42LGj4fBg8O/2zpYfP1Wikh3EVknIhtFZGQ658uLyKci8rOIrBaR4Nd2WvUYfFwbdi125nLmpWYyRr6lZEnYuRMuvNCtTDBkSMExGfDRaEQkApgIXA40BvqJSNolHYcBa1S1BW6xuedEJOvnnoPxsPpfJ/ajH3SxGcPIRwTGXs47z7VqvvkG6tQJlSL/8LMJcC6wUVU3q+pRYCZwVZoyCpQVEQHKAHuAY1nWvHU2JB+BMvWg72Fo+WQuSzcM/zh2zGW8q18fPv30xPELL3Q9SwWRoI1GRLK7GENNYGvAfrx3LJAJQDTwB7AS+LuqJmdZs3peVLMnRJTIpizDCB0bNjhDefRRZzg//xxqRXlDlkYjIh1EZA2w1ttvISKTgqg7vSdMTbN/GRAL1ABaAhNEpFw6GoaKSIyIxCQkJGRxC8MIP1ThlVegZUv44QeIioL5890o38JAMC2aF3CGsBtAVX8GLgriunjgrID9KFzLJZAhwIfq2Aj8Cpwye0NVp6hqW1VtW7Vq1SBubRjhQ0IC9OzpFmo7eBD693dL0IZTYiq/CerRSVW3pjl0PIjLlgENRKSOF+C9AfgkTZktwMUAIlIdOAfYnIUa+HFEELc3jPCgeHE3HqZiRZg50+XwDeeUDn4QzDiarSLSAVDPMO7Ge4zKDFU9JiLDgblABDBVVVeLyB3e+ZeBscA0EVmJew56UFV3ZVpx4oYT26XPDkK+YeQ9+/a5SY8lS0L58i5J1RlnuKVOCiOimjZskqaASBXg38AlODP4ErhbVff4L+9U2tYVjXkciLoKOs4uWIMNjALBwoVudO9VV8GLL4ZaTe4hIstVtW1Org2mRXOOqvZPc8MLgEU5uWGucNEnENUrZLc3jPQ4csT1Jj3zjAv+LlrkjpWwjtGgYjTp5VoIbf6FUjVCenvDSMuqVW7Q3dNPn8gZs2iRmUwKGbZoRKQ90AGoKiKB0ddyuJiLYRR6VOGFF+D//s+tRlCvnpt93b59qJWFF5m1aIrjRusWBcoGvPYBtsSjYXgsXuxMZuhQiI01k0mPYILBZ6tq2GTEaFtXNCYmBiq1CbUUo5Ci6nqVypd3+7t2uUF4PXqEVpff+B0MPigizwJNgNTFNVW1EA03MgzHnj1u4N3GjfD9926MTJUqBd9kTpdggsFvA78AdYB/AnG4wXiGUaiYOxeaNoX333dzlgrLPKXcIBijqayqrwNJqvqtqt4MnO+zLsMIGw4ehL/9Dbp3d4mpLrjAmUy7dqFWln8I5tEpyfu5TUR64OYrRfknyTDCh5gYuOkmWLcOihZ16R0eeKDgpnPwi2CM5nERKQ/cixs/Uw64x09RhhEuLFvmTKZxYzdHqVWrUCvKn2RpNKrqJeVlL9AFUkcGG0aB5NAhN0cJXOC3SBE3pSDlmJF9MozRiEiEiPQTkftEpKl3rKeILMYlrDKMAoUqTJniUmlu2uSOicDtt5vJnC6ZBYNfB24FKgPjReQNYBzwjKpaA9IoUGzfDr16OVPZscOlczByj8wendoCzVU1WUQigV1AfVXdnjfSDCNv+OgjuO02N/CuQgWYPBluuCHUqgoWmbVojqbk71XVw8B6MxmjIJGYCLfcAr17O5O55BKX+c5MJvfJrEXTSERWeNsC1PP2BVBVbe67OsPwkbg415MUGelSOwwb5gK/Ru6TmdFE55kKw8gjjh1z42EAmjWDqVNdl3XjtCuOGblKhkYTThMpDSM3WLXKDb4bOfLE41H//plfY+QO1lA0CjzJyS5nTNu2burAuHGuK9vIO8xojALN1q1w6aUwYoRLq3nrrfD115ZqOq8JymhEpKSInOO3GMPILVThnXdcHOZ//4OqVeHjj+HVV6Fs2VCrK3wEs1JlL9xqkl94+y1FJO36TIYRVhw9Cv/8J+zdC1de6eIzV14ZalWFl2BaNGOAc4G/AFQ1FqgdTOUi0l1E1onIRhEZmUGZziISKyKrReTbYOo1jIxI9lZuL1HC5e599VU3IK9atZDKKvQEM3v7mKrulWw+1IpIBDARuBS3PO4yEflEVdcElKkATAK6q+oWEbE/ByNHHDrkepMOHXLzlQDOPde9jNATTItmlYjcCESISAMReQlYHMR15wIbVXWzqh4FZgJXpSlzI27t7S0AqrozG9oNA4Dly6F1axg/HqZNOzEh0ggfgjGav+HyBR8B3sGli7gniOtqAoFrdsd7xwJpCFQUkW9EZLmIDAyiXsMA3OC7J56A88+HX36B6GiXx7devVArM9IS7EqVo4BR2aw7vWettKMXigJtgIuBksASEfleVdefVJHIUGAoQJs62VRhFEg2bYIBA2DJErf/97/Dv/5l6RzClWBaNM+LyC8iMlZEmmSj7njgrID9KFwa0LRlvlDVA6q6C1gAtEhbkapOUdW2OV3qwSh4PP20M5maNeGrr9wa12Yy4UuWRqOqXYDOQAIwRURWisjoIOpeBjQQkToiUhy4AUjbLf4x0FFEiopIKeA8YG123oBReAgczfvssy5h+MqVbta1Ed4ENWBPVber6njgDtyYmkeCuOYYMByYizOP91R1tYjcISJ3eGXW4sbnrACWAq+p6qqcvBGjYPPRR9C5s+tVArd42/jxULFiKFUZwRLMSpXRQF/cMri7cb1Hs0LVQ2QrVRYuEhPhnnvcLGuAl16C4cNDKqnQ4vdKlW8AM4Buqpo2xmIYvvHddy4p+K+/ugF4Tz8Nd90ValVGTghmFQRbLM7IU44ehUcfdcaiCi1bwttvW86Y/EyGRiMi76lqHxFZycnd0pZhz/CVOXPgqadctruRI2HMGLfGtZF/yaxF83fvZ8+8EGIYKVx1FfzjH3DNNXDhhaFWY+QGGfY6qeo2b/MuVf0t8AXYk7KRa2zdCj16wBpvFpwIPP+8mUxBIpju7UvTOXZ5bgsxCh+BOWPmzHGtGKNgklmM5k5cy6VuwGoIAGWBRX4LMwo2e/a4HqR333X7PXvCa6+FVpPhH5nFaN4BPgf+BQTmkklU1T2+qjIKNF99BUOGwO+/Q+nSLp/vrbdaes2CTGZGo6oaJyLD0p4QkUpmNkZO2L3bLdh24ICbdT19OtSvH2pVht9k1aLpCSzHdW8H/r9RoK6PuowCSuXKrgWzY4frui4azJBRI9+T2bpOPb2flpjByDHHjrlVIM84A26+2R277bbQajLyniz/n4jIBUCsqh4QkZuA1sCLKVnxDCMjNm1yUwgWL4YyZeDqq6FSpVCrMkJBMN3bk4GDItICeAD4DZjuqyojX6PqepBatHAmU6MGzJplJlOYCcZojqmb4n0V8G9V/Teui9swTmHHDjey97bbXMC3Tx+XM6Zbt1ArM0JJMKG4RBH5P2AALklVBFDMX1lGfuWmm2DePJcvZtIk6NfPuq2N4Fo0fXGJyW9W1e24BOPP+qrKyLc89xxcfrlrxdx4o5mM4Qgmled24G2gvIj0BA6r6lu+KzPyBYsWwf33n9hv3txNJzjrrIyvMQofwSyJ2weXZvN6oA/wg4hc57cwI7w5ehRGjYKLLoJx4+ATWyTZyIRgYjSjgHYpqTtFpCowD/jAT2FG+LJmjYvF/PSTezQaORIuuyzUqoxwJhijKZImP/BugkxqbhQskpNdzt4HH4QjR6B2bXjrLejYMdTKjHAnGKP5QkTm4vIGgwsOz/FPkhGuTJjgEoWDG+X7wgtQrlxIJRn5hGByBt8vItcAF+LmO01R1dm+KzPCjltugffeg3vvdRMjDSNYMnwEEpEGIvKxiKzCBYKfU9V/ZMdkRKS7iKwTkY0iMjKTcu1E5LgFmcOLP/90yagSE91+6dKwcKGZjJF9Mou1TAU+A67FzeB+KTsVewP7JuKy8TUG+onIKXnsvXJP4xaaM8KEefNc5rsXX3QxmRRsXIyREzIzmrKq+qqqrlPVcUDtbNZ9LrBRVTer6lHcwnNXpVPub8AsICQL0hknc+iQi8NceqlLTHX++TBiRKhVGfmdzGI0kSLSihN5aEoG7qvqj1nUXRPYGrAfj1tbOxURqQn0BroC7bKh2/CBH3903dZr17o8MWPGuNaM5YwxTpfM/oS2Ac8H7G8P2FecOWRGeo3stOvvvgg8qKrHJZM2uYgMBYYCtLHsOL6wYYNrvSQlwTnnwH/+A21ztPipYZxKZomvupxm3fFA4ED0KCDtkrptgZmeyVQBrhCRY6r6URotU4Ap4NbePk1dRjo0aAD9+0PZsm7xtlKlQq3IKEj42SheBjQQkTrA78ANwI2BBQKz94nINOCztCZj+IMqTJ3qlptt08Yde/11tzqkYeQ2vhmNqh4TkeG43qQIYKqqrhaRO7zzL/t1byNzdu6EoUPh44+hUSOIjYUSJcxkDP/wNcynqnNIM4o4I4NR1cF+ajEcn37qljbZudPljBk92ta1NvwnmNnbIiI3icgj3n4tETnXf2lGbrJ/v2vFXHmlM5kuXWDFCheXsbExht8E01ieBLQH+nn7ibiBeEY+ITkZOnWCV191j0jPP+8G5NWqFWplRmEhGKM5T1WHAYcBVPVPwBrb+YgiRdxUghYtICbGbVs8xshLgvlzS/KmCSik5qNJ9lWVcdqsXQszZ57Y798fli2Dpk1Dp8kovARjNOOB2UA1EXkC+A540ldVRo5JTobx46F1axg8GFavdsdFoJillDdCRDBpIt4WkeXAxbjRvler6lrflRnZJj4ehgxx8Rdw25a71wgHglmpshZwEPg08JitVBlevPsu3HEH/PUXVKkCU6ZYOgcjfAhmHM1/cfEZASKBOsA6oImPuoxs8Pjj8PDDbrtHD7dK5BlnhFaTYQQSzHIrzVS1ufezAS79w3f+SzOCpW9fqFYNXn7ZDcgzkzHCjWyPDFbVH0XEUjqEkMOH3ezqW25xQd4GDeDXX20ipBG+BBOjCUx7VARoDST4psjIlJ9+cjlj1qxx+7fe6n6ayRjhTDDd22UDXiVwMZv0MuUZPnL8uEvfcN55zmTOOccNwDOM/ECmLRpvoF4ZVb0/s3KGv2zeDAMHuuVnAYYPh6eftlaMkX/I0GhEpKiX6qF1XgoyTmbpUrj4Yjcp8swz4Y03bFVII/+RWYtmKS4eEysinwDvAwdSTqrqhz5rM3CPR3XrukelyZOhcuVQKzKM7BNMr1Ml3DK4XTkxnkYBMxqfmDPH5e+tVMnNtv72W5c7xtI5GPmVzIymmtfjtIoTBpOC5e31gf373dImr77qxsbMmOHMpUKFUCszjNMjM6OJAMoQ3GoGxmmyZAkMGACbNrmMd+1spJJRgMh0uRVVfSzPlBRSkpLgscfgySfdzOvmzd1gvGbNQq3MMHKPzIzGIgI+c/gwdOzoklGJwAMPONMpUSLUygwjd8nMaC7OMxWFlMhINwAvIQHeegsuuijUigzDHzIcGayqe/JSSGHh99/d8iYpPPOMSxJuJmMUZHzNHCsi3UVknYhsFJGR6ZzvLyIrvNdiESnQg+rfe8/FXq65BhIT3bFSpaBcudDqMgy/8c1ovOkLE4HLgcZAPxFpnKbYr0AnVW0OjMVb9rag8ddfbiJk377w558QHe3iM4ZRWPCzRXMusFFVN6vqUWAmaSZjqupib1UFgO9x63MXKL7+2vUkvf22a728/DJ89hlUrRpqZYaRd/hpNDWBrQH78d6xjLgF+NxHPXnO2LHQtSts3eqCvrGxcPvtNsLXKHz4aTRBD/QTkS44o3kwg/NDRSRGRGJyUZ/vNGkCRYu6LuvvvnMJqgyjMOLn2tvxQGAO/ijgj7SFRKQ58BpwuaruTq8iVZ2CF79pW1fCdlTy8eNutnX79m7/mmtg/XqoUye0ugwj1PjZolkGNBCROiJSHLgB+CSwgLfCwofAAFVd76MW34mLc+tZX3QRLF9+4riZjGH42KLxctkMB+bi5k1NVdXVInKHd/5l4BGgMjBJXODimKq29UuTH6jCm2/C3Xe7LuszzjjRdW0YhkNUw/ZJJF3a1hWNiYmBSm1CLYWEBBfcnT3b7V97LbzyiuWMMQomIrI8pw0BP2M0BZpFi5yx7NjhBtxNmODGyliPkmGcihlNDjnrLDh0CDp1co9OZ58dakWGEb6Y0WSDlStdl3WRIlCrFixeDI0aQUREqJUZRnhjRhMESUlu2dknnoBx4+Cee9zxJtlcFDgpKYn4+HgO2/wDI4yJjIwkKiqKYsWK5VqdZjRZsG6di72k5IzZne5In+CIj4+nbNmy1K5dG7FgjhGGqCq7d+8mPj6eOrk4NsPX2dv5GVWYOBFatXImU6uWm7c0dmzO6zx8+DCVK1c2kzHCFhGhcuXKud7qthZNOvz5J/TrB3Pnuv2BA2H8eLcSweliJmOEO378jZrRpEPZsi61Q6VKblzMddeFWpFh5G/s0clj717YtcttFy0KM2fCqlUFz2QiIiJo2bIlTZs2pVevXvz111+p51avXk3Xrl1p2LAhDRo0YOzYsQQO6Pz8889p27Yt0dHRNGrUiPvuuy8E7yBzfvrpJ2699dZQy8iQI0eO0LdvX+rXr895551HXFxcuuXeffddmjdvTpMmTXjggQdSj2/ZsoUuXbrQqlUrmjdvzpw5c066bt++fdSsWZPhw4enHlNVRo0aRcOGDYmOjmb8+PEAfPbZZzz66KO5/ybTQ1Xz1atNHVR3x2hu8vXXqrVqqfbsqZqcnKtVn8SaNWv8qzxISpcunbo9cOBAffzxx1VV9eDBg1q3bl2dO3euqqoeOHBAu3fvrhMmTFBV1ZUrV2rdunV17dq1qqqalJSkEydOzFVtSUlJp13Hddddp7GxsXl6z+wwceJEvf3221VVdcaMGdqnT59TyuzatUvPOuss3blzp6q639O8efNUVfW2227TSZMmqarq6tWr9eyzzz7p2rvvvlv79eunw4YNSz02depUHTBggB4/flxVVXfs2KGqqsnJydqyZUs9cODAKRrS+1sFYjSH39tC3aI5fBjuu8/ljNmyxY3y3bs3j27+jvjzygbt27fn999/d3LeeYcLLriAbt26AVCqVCkmTJjAU089BcAzzzzDqFGjaNSoEQBFixblrrvuOqXO/fv3M2TIEJo1a0bz5s2ZNWsWAGXKlEkt88EHHzB48GAABg8ezIgRI+jSpQv3338/tWvXPqmVVb9+fXbs2EFCQgLXXnst7dq1o127dixatOiUeycmJrJixQpatHAZYZcuXUqHDh1o1aoVHTp0YN26dQBMmzaN66+/nl69etGtWzcOHDjAzTffTLt27WjVqhUff/wxAHFxcXTs2JHWrVvTunVrFi9enK3PNz0+/vhjBg0aBMB1113H/PnzT2o1AmzevJmGDRtS1cuOdskll6R+jiLCvn37ANi7dy81atRIvW758uXs2LEj9XeYwuTJk3nkkUcoUsR93atVq5ZaV+fOnfnss89O+31lRaGN0axY4bqtV650A+5Gj4ZRoyAXhw6ENcePH2f+/PnccsstgHtsatPm5Plj9erVY//+/ezbt49Vq1Zx7733Zlnv2LFjKV++PCtXrgTgzz//zOIKWL9+PfPmzSMiIoLk5GRmz57NkCFD+OGHH6hduzbVq1fnxhtv5B//+AcXXnghW7Zs4bLLLmPt2rUn1RMTE0PTpk1T9xs1asSCBQsoWrQo8+bN46GHHkr9wi5ZsoQVK1ZQqVIlHnroIbp27crUqVP566+/OPfcc7nkkkuoVq0aX331FZGRkWzYsIF+/foRE3NqSqSOHTuSmM5M2nHjxnHJJZecdOz333/nrLNc9pSiRYtSvnx5du/eTZUqVVLL1K9fn19++YW4uDiioqL46KOPOHr0KABjxoyhW7duvPTSSxw4cIB58+YBkJyczL333sv06dOZP3/+SffctGkT7777LrNnz6Zq1aqMHz+eBl5ypLZt27Jw4UL69OmT5e/pdCiURvPcc/DQQ3D0qEtGNX26y4CXp9wYmsmshw4domXLlsTFxdGmTRsuvfRSwD1CZ9TbkJ1eiHnz5jFz5szU/YoVK2Z5zfXXX0+EN7y6b9++PPbYYwwZMoSZM2fSt2/f1HrXrFmTes2+fftITEykbNmyqce2bduW2goA9x9/0KBBbNiwAREhKSkp9dyll15KpUqVAPjyyy/55JNPGDduHOCGIWzZsoUaNWowfPhwYmNjiYiIYP369DOZLFy4MMv3mELa1guc+vlWrFiRyZMn07dvX4oUKUKHDh3YvHkzADNmzGDw4MHce++9LFmyhAEDBrBq1SomTZrEFVdckWpigRw5coTIyEhiYmL48MMPufnmm1M1V6tWjT/+OCVNVK5TKI1m505nMnfeCc8+C6VLh1pR3lGyZEliY2PZu3cvPXv2ZOLEidx99900adKEBQsWnFR28+bNlClThrJly9KkSROWL1+e+liSERkZVuCxtGM0Sgf8Atq3b8/GjRtJSEjgo48+YvTo0YD7j71kyRJKliyZ6XsLrPvhhx+mS5cuzJ49m7i4ODp37pzuPVWVWbNmcc4555xU35gxY6hevTo///wzycnJREZGpnvf7LRooqKi2Lp1K1FRURw7doy9e/emGl4gvXr1olevXgBMmTIl1Yhff/11vvjii9TP6vDhw+zatYslS5awcOFCJk2axP79+zl69ChlypThqaeeIioqimuvvRaA3r17M2TIkNT7HD58ONPPNLcoFDEaVdi27cT+Y4/B/PkwaVLhMplAypcvz/jx4xk3bhxJSUn079+f7777LrUpfujQIe6+++7UHo/777+fJ598MvW/enJyMs8///wp9Xbr1o0JEyak7qc8OlWvXp21a9emPhplhIjQu3dvRowYQXR0NJW9nBtp640NXBzLIzo6mo0bN6bu7927l5o1XZrqadOmZXjPyy67jJdeeim1tfHTTz+lXn/mmWdSpEgRpk+fzvHjx9O9fuHChcTGxp7ySmsyAFdeeSVvvvkm4GJVXbt2TdeYd+7cCbjPb9KkSak9abVq1Up9NFq7di2HDx+matWqvP3222zZsoW4uDjGjRvHwIEDU+NrV199Nf/73/8A+Pbbb2nYsGHqfdavX3/S46Zv5DSKHKpXdnudEhJUr7lGtUYN1d27g77MF8Kt10lVtWfPnvrWW2+pquqKFSu0U6dO2rBhQ61Xr56OGTNGkwO64T799FNt3bq1NmrUSKOjo/W+++47pf7ExEQdOHCgNmnSRJs3b66zZs1SVdX3339f69atq506ddJhw4bpoEGDVFV10KBB+v77759Ux7JlyxTQadOmpR5LSEjQPn36aLNmzTQ6Ojq15yYtTZs21X379qmq6uLFi7VBgwbaoUMHHT16dGoPzRtvvHFSr8zBgwd16NCh2rRpU23SpIn26NFDVVXXr1+vzZo10/POO09Hjhx5ymeXEw4dOqTXXXed1qtXT9u1a6ebNm1KPdeiRYvU7RtuuEGjo6M1OjpaZ8yYkXp89erV2qFDB23evLm2aNEitZcwkLTv788//9QrrrhCmzZtqueff/5JvXI9evTQFStWnFJHbvc6hdw4svvKjtH897+q1au7d1m2rOvGDiXhYDQFneeff15fffXVUMvIF2zfvl27du2a7jnr3g6CAwdc/KVHD9dl3bGj62UKeEQ3Cih33nknJUqUCLWMfMGWLVt47rnn8uReBS4YvGwZ9O8PGza4ruonnoARIyxnTGEhMjKSAQMGhFpGvqBdu3Z5dq8CZzR79jiTadoU/vMfyKKTJM/RTLqRDSMc0HS64E+XAvHoFDgm7LLL4MMPXcsm3EwmMjKS3bt3+/KLNIzcQNXlo8moKz+n5OsWjSpMngwPPgiffw4XXuiO9+4dWl0ZERUVRXx8PAkJCaGWYhgZkpJhLzfx1WhEpDvwb9y6Tq+p6lNpzot3/grgIDBYVX8Mpu5t2+Dmm8Ebu8R//3vCaMKVYsWK5WrWMsPIL/j26CQiEcBE4HKgMdBPRBqnKXY50MB7DQUmB1P3rE8q0LSpM5lKleC99+Bf/8pF8YZh5Cp+xmjOBTaq6mZVPQrMBK5KU+Yq4C2vm/57oIKInJlZpXEJtbluSD327HHxmJUr4frr/XkDhmHkDn4aTU1ga8B+vHcsu2VOYu/B8pQsmcyECS4uEzBL3jCMMMXPGE16fbhpu1uCKYOIDMU9WgEcOXYoYtXw4RCQRCzcqQLsCrWIbJDf9IJpzgvOybpI+vhpNPFA4Jz1KCDtfPRgyqCqU4ApACISozlc/zdU5DfN+U0vmOa8QEROTcYTJH4+Oi0DGohIHREpDtwAfJKmzCfAQHGcD+xV1W1pKzIMI3/jW4tGVY+JyHBgLq57e6qqrhaRO7zzLwNzcF3bG3Hd20Myqs8wjPyLr+NoVHUOzkwCj70csK3AsGxWOyUXpOU1+U1zftMLpjkvyLFeseHwhmH4TYGY62QYRngTtkYjIt1FZJ2IbBSRkemcFxEZ751fISKtQ6EzQE9Wevt7OleIyGIRCfmUz6w0B5RrJyLHRSTky+kFo1lEOotIrIisFpFv81pjGi1Z/V2UF5FPReRnT2/I45QiMlVEdorIqgzOZ/+7l9OMWX6+cMHjTUBdoDjwM9A4TZkrgM9xY3HOB34Ic70dgIre9uWh1Bus5oBy/8PF2q4Ld81ABWANUMvbrxbmeh8Cnva2qwJ7gOIh/pwvAloDqzI4n+3vXri2aHyZvuAjWepV1cWqmpLQ4nvcmKFQEsxnDPA3YBawMy/FZUAwmm8EPlTVLQCqGkrdwehVoKw3wbgMzmiO5a3MNIJUF3g6MiLb371wNRpfpi/4SHa13IL7jxBKstQsIjWB3sDLhAfBfM4NgYoi8o2ILBeRgXmm7lSC0TsBiMYNVF0J/F1Vk/NGXo7J9ncvXPPR5Nr0hTwiaC0i0gVnNKFOahGM5heBB1X1eJhkBQxGc1GgDXAxUBJYIiLfq2r6q7/5SzB6LwNiga5APeArEVmoqvt81nY6ZPu7F65Gk2vTF/KIoLSISHPgNeByVd2dR9oyIhjNbYGZnslUAa4QkWOq+lGeKDyVYP8udqnqAeCAiCwAWgChMJpg9A4BnlIX/NgoIr8CjYCleSMxR2T/uxfKoFMmwaiiwGagDieCaE3SlOnByQGppWGutxZuBHSHUH++wWpOU34aoQ8GB/M5RwPzvbKlgFVA0zDWOxkY421XB34HqoTB30dtMg4GZ/u7F5YtGs1n0xeC1PsIUBmY5LUQjmkIJ9QFqTmsCEazqq4VkS+AFUAyLrNjut204aAXGAtME5GVuC/ug6oa0hndIjID6AxUEZF44FGgGOT8u2cjgw3D8J1w7XUyDKMAYUZjGIbvmNEYhuE7ZjSGYfiOGY1hGL5jRpNP8WZTxwa8amdSdn8u3G+aiPzq3etHEWmfgzpeS1nbS0QeSnNu8elq9OpJ+VxWebOiK2RRvqWIXJEb9zYyxrq38ykisl9Vy+R22UzqmAZ8pqofiEg3YJyqNj+N+k5bU1b1isibwHpVfSKT8oOBtqqaf9bUyIdYi6aAICJlRGS+19pYKSKnzMQWkTNFZEHAf/yO3vFuIrLEu/Z9EcnKABYA9b1rR3h1rRKRe7xjpUXkv16OlVUi0tc7/o2ItBWRp4CSno63vXP7vZ/vBrYwvJbUtSISISLPisgyLwfK7UF8LEvwJvuJyLni8gD95P08R1zS/MeAvp6Wvp72qd59fkrvczRyQKiHOtsrx0PEj+Mm48UCs3HD3ct556rgRm2mtFj3ez/vBUZ52xFAWa/sAqC0d/xB4JF07jcNbwoCcD3wA27y4kqgNC7FwWqgFXAt8GrAteW9n9/gWg+pmgLKpGjsDbzpbRfHzRIuiVvXa7R3vAQQA9RJR+f+gPf3PtDd2y8HFPW2LwFmeduDgQkB1z8J3ORtV8DNkSod6t93fn+F5RQEIygOqWrLlB0RKQY8KSIX4Ybe18TNndkecM0yYKpX9iNVjRWRTri10Rd5UyOK41oC6fGsiIwGEnAz0C8GZqubwIiIfAh0BL4AxonI07jHrYXZeF+fA+NFpATQHVigqoe8x7XmciLLX3ncmu2/prm+pIjE4ubqLAe+Cij/pog0wM00LpbB/bsBV4rIfd5+JG6e2tpsvAcjDWY0BYf+uAxtbVQ1SUTicF+SVFR1gWdEPYDpIvIs8Cfwlar2C+Ie96vqByk7InJJeoVUdb2ItMHNh/mXiHypqo8F8yZU9bCIfINLn9AXmJFyO+Bvqjo3iyoOqWpLESkPfIZbZWM8bk7R16ra2wucf5PB9QJcq6rrgtFrBIfFaAoO5YGdnsl0Ac5OW0BEzvbKvAq8jkvX+D1wgYikxFxKiUjDIO+5ALjau6Y07rFnoYjUAA6q6n+Acd590pLktazSYyZuol5H3IREvJ93plwjIg29e6aLqu4F7gbu864pj5sZDe5xKYVE3CNkCnOBv4nXvBORVhndwwgeM5qCw9tAW3HLlvYHfkmnTGcgVkR+wsVR/q2qCbgv3gwRWYEznkbB3FBVf8TFbpbiYjavqepPQDNgqfcIMwp4PJ3LpwArUoLBafgSl7d2nroUmODy+KwBfhSXNPsVsmiRe1p+xq2S+gyudbUIF79J4WugcUowGNfyKeZpW+XtG6eJdW8bhuE71qIxDMN3zGgMw/AdMxrDMHzHjMYwDN8xozEMw3fMaAzD8B0zGsMwfMeMxjAM3/l/t73RWtGZkzkAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 288x216 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Area Under Curve: 0.9846211975046626\n"
     ]
    },
    {
     "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": [
    "#Random Forest Classifier\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "model = RandomForestClassifier(n_estimators=100, max_depth=13, max_features= 4, \n",
    "                               class_weight='balanced', random_state=1).fit(X_train, y_train)\n",
    "y_pred = model.predict(X_test)\n",
    "print(\"Accuracy:\",metrics.accuracy_score(y_test, y_pred))\n",
    "print(\"Classification Error:\",1-metrics.accuracy_score(y_test, y_pred))\n",
    "cv = RepeatedKFold(n_splits=5, n_repeats= 100, random_state=1)\n",
    "print(\"Mean Cross Val (Accuracy) Score\", np.mean(cross_val_score(model, X, y, cv=cv)))\n",
    "print(\"Mean Classification Error:\", 1 - np.mean(cross_val_score(model, X, y, scoring='accuracy', cv=cv, n_jobs=-1)))\n",
    "y_pred_proba = model.predict_proba(X_test)[:,1]\n",
    "fpr, tpr, _ = metrics.roc_curve(y_test, y_pred_proba)\n",
    "roc_auc = auc(fpr,tpr)\n",
    "plt.figure(figsize=(4,3))\n",
    "plt.plot(fpr, tpr, color='orange',\n",
    "lw=2, label='ROC curve (area = {0:.4f})'.format(roc_auc))\n",
    "plt.plot([0, 1], [0, 1], color='blue', lw=2, linestyle='--')\n",
    "plt.xlim([0.0, 1.0])\n",
    "plt.ylim([0.0, 1.05])\n",
    "plt.xlabel('False Positive Rate')\n",
    "plt.ylabel('True Positive Rate')\n",
    "plt.title('ROC curve')\n",
    "plt.legend(loc=\"lower right\")\n",
    "plt.show()\n",
    "print('Area Under Curve:', roc_auc)\n",
    "\n",
    "import pandas as pd\n",
    "feature_imp = pd.Series(model.feature_importances_,index=feature_cols).sort_values(ascending=False)\n",
    "feature_imp\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "%matplotlib inline\n",
    "sns.barplot(x=feature_imp, y=feature_imp.index, label='Strike price of option')\n",
    "plt.xlabel('Feature Importance Score')\n",
    "plt.ylabel('Features')\n",
    "plt.title(\"Visualizing Important Features\")\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4e7f4418",
   "metadata": {},
   "source": [
    "### Gaussian Naive Bayes Classifier "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 900,
   "id": "1596f81e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best Parameters:  {'var_smoothing': 0.43287612810830584}\n",
      "Accuracy Score:  0.8856149500217297\n"
     ]
    }
   ],
   "source": [
    "# Gaussian Naive Bayes Classifier \n",
    "from sklearn.naive_bayes import GaussianNB \n",
    "model = GaussianNB()\n",
    "forest_params = {'var_smoothing': np.logspace(0,-9, num=100)}\n",
    "clf = GridSearchCV(model, forest_params, cv = 10, scoring='accuracy')\n",
    "clf.fit(X_train, y_train)\n",
    "print('Best Parameters: ', clf.best_params_)\n",
    "print('Accuracy Score: ', clf.best_score_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 112,
   "id": "3f688f54",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.878727634194831\n",
      "Classification Error: 0.12127236580516898\n",
      "Mean Cross Val (Accuracy) Score 0.8757850746268657\n",
      "Mean Classification Error: 0.12421492537313428\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x216 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Area Under Curve: 0.9629236606855746\n"
     ]
    }
   ],
   "source": [
    "# Gaussian Naive Bayes Classifier \n",
    "from sklearn.naive_bayes import GaussianNB \n",
    "model = GaussianNB(var_smoothing= 0.43287612810830584).fit(X_train, y_train)\n",
    "y_pred = model.predict(X_test)\n",
    "print(\"Accuracy:\",metrics.accuracy_score(y_test, y_pred))\n",
    "print(\"Classification Error:\",1-metrics.accuracy_score(y_test, y_pred))\n",
    "cv = RepeatedKFold(n_splits=5, n_repeats= 100, random_state=1)\n",
    "print(\"Mean Cross Val (Accuracy) Score\", np.mean(cross_val_score(model, X, y, cv=cv)))\n",
    "print(\"Mean Classification Error:\", 1 - np.mean(cross_val_score(model, X, y, scoring='accuracy', cv=cv, n_jobs=-1)))\n",
    "y_pred_proba = model.predict_proba(X_test)[:,1]\n",
    "fpr, tpr, _ = metrics.roc_curve(y_test, y_pred_proba)\n",
    "roc_auc = auc(fpr,tpr)\n",
    "plt.figure(figsize=(4,3))\n",
    "plt.plot(fpr, tpr, color='orange',\n",
    "lw=2, label='ROC curve (area = {0:.4f})'.format(roc_auc))\n",
    "plt.plot([0, 1], [0, 1], color='blue', lw=2, linestyle='--')\n",
    "plt.xlim([0.0, 1.0])\n",
    "plt.ylim([0.0, 1.05])\n",
    "plt.xlabel('False Positive Rate')\n",
    "plt.ylabel('True Positive Rate')\n",
    "plt.title('ROC curve')\n",
    "plt.legend(loc=\"lower right\")\n",
    "plt.show()\n",
    "print('Area Under Curve:', roc_auc)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "87a81cb8",
   "metadata": {},
   "source": [
    "### Ada Boost Classifier "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 903,
   "id": "5250fda3",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best Parameters:  {'learning_rate': 0.5, 'n_estimators': 500}\n",
      "Accuracy Score:  0.9232217876285673\n"
     ]
    }
   ],
   "source": [
    "# Ada Boost Classifier \n",
    "from sklearn.ensemble import AdaBoostClassifier\n",
    "model = AdaBoostClassifier()\n",
    "forest_params = {'n_estimators':[100,500],'learning_rate':[0.01,.1,.5,1]}\n",
    "clf = GridSearchCV(model, forest_params, cv = 10, scoring='accuracy')\n",
    "clf.fit(X_train, y_train)\n",
    "print('Best Parameters: ', clf.best_params_)\n",
    "print('Accuracy Score: ', clf.best_score_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 113,
   "id": "6c724403",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.9145129224652088\n",
      "Classification Error: 0.08548707753479123\n",
      "Mean Cross Val (Accuracy) Score 0.9221432835820894\n",
      "Mean Classification Error: 0.07785671641791059\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x216 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Area Under Curve: 0.975311917165091\n"
     ]
    }
   ],
   "source": [
    "# Ada Boost Classifier \n",
    "from sklearn.ensemble import AdaBoostClassifier\n",
    "model = AdaBoostClassifier(n_estimators=500,learning_rate=.5, random_state=1).fit(X_train, y_train)\n",
    "y_pred = model.predict(X_test)\n",
    "print(\"Accuracy:\",metrics.accuracy_score(y_test, y_pred))\n",
    "print(\"Classification Error:\",1-metrics.accuracy_score(y_test, y_pred))\n",
    "cv = RepeatedKFold(n_splits=5, n_repeats= 100, random_state=1)\n",
    "print(\"Mean Cross Val (Accuracy) Score\", np.mean(cross_val_score(model, X, y, cv=cv)))\n",
    "print(\"Mean Classification Error:\", 1 - np.mean(cross_val_score(model, X, y, scoring='accuracy', cv=cv, n_jobs=-1)))\n",
    "y_pred_proba = model.predict_proba(X_test)[:,1]\n",
    "fpr, tpr, _ = metrics.roc_curve(y_test, y_pred_proba)\n",
    "roc_auc = auc(fpr,tpr)\n",
    "plt.figure(figsize=(4,3))\n",
    "plt.plot(fpr, tpr, color='orange',\n",
    "lw=2, label='ROC curve (area = {0:.4f})'.format(roc_auc))\n",
    "plt.plot([0, 1], [0, 1], color='blue', lw=2, linestyle='--')\n",
    "plt.xlim([0.0, 1.0])\n",
    "plt.ylim([0.0, 1.05])\n",
    "plt.xlabel('False Positive Rate')\n",
    "plt.ylabel('True Positive Rate')\n",
    "plt.title('ROC curve')\n",
    "plt.legend(loc=\"lower right\")\n",
    "plt.show()\n",
    "print('Area Under Curve:', roc_auc)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6cb2ca67",
   "metadata": {},
   "source": [
    "### Ridge Classifier "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 913,
   "id": "04551c56",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best Parameters:  {'alpha': 0.0001}\n",
      "Accuracy Score:  0.9087136027813993\n"
     ]
    }
   ],
   "source": [
    "#Ridge Classifier \n",
    "from sklearn.linear_model import RidgeClassifier\n",
    "model = RidgeClassifier()\n",
    "forest_params = {'alpha':np.array([.0001, .001,.01,.1,.5,1])}\n",
    "clf = GridSearchCV(model, forest_params, cv = 10, scoring='accuracy')\n",
    "clf.fit(X_train, y_train)\n",
    "print('Best Parameters: ', clf.best_params_)\n",
    "print('Accuracy Score: ', clf.best_score_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 114,
   "id": "3e2380f6",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.9065606361829026\n",
      "Classification Error: 0.09343936381709739\n",
      "Mean Cross Val (Accuracy) Score 0.9111820895522388\n",
      "Mean Classification Error: 0.08881791044776122\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x216 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Area Under Curve: 0.9757219113769374\n"
     ]
    }
   ],
   "source": [
    "#Ridge Classifier \n",
    "from sklearn.linear_model import RidgeClassifier\n",
    "model = RidgeClassifier(alpha=.0001, random_state=1).fit(X_train, y_train)\n",
    "y_pred = model.predict(X_test)\n",
    "print(\"Accuracy:\",metrics.accuracy_score(y_test, y_pred))\n",
    "print(\"Classification Error:\",1-metrics.accuracy_score(y_test, y_pred))\n",
    "cv = RepeatedKFold(n_splits=5, n_repeats= 100, random_state=1)\n",
    "print(\"Mean Cross Val (Accuracy) Score\", np.mean(cross_val_score(model, X, y, cv=cv)))\n",
    "print(\"Mean Classification Error:\", 1 - np.mean(cross_val_score(model, X, y, scoring='accuracy', cv=cv, n_jobs=-1)))\n",
    "y_pred_proba = model._predict_proba_lr(X_test)[:,1]\n",
    "fpr, tpr, _ = metrics.roc_curve(y_test, y_pred_proba)\n",
    "roc_auc = auc(fpr,tpr)\n",
    "plt.figure(figsize=(4,3))\n",
    "plt.plot(fpr, tpr, color='orange',\n",
    "lw=2, label='ROC curve (area = {0:.4f})'.format(roc_auc))\n",
    "plt.plot([0, 1], [0, 1], color='blue', lw=2, linestyle='--')\n",
    "plt.xlim([0.0, 1.0])\n",
    "plt.ylim([0.0, 1.05])\n",
    "plt.xlabel('False Positive Rate')\n",
    "plt.ylabel('True Positive Rate')\n",
    "plt.title('ROC curve')\n",
    "plt.legend(loc=\"lower right\")\n",
    "plt.show()\n",
    "print('Area Under Curve:', roc_auc)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "04856f1d",
   "metadata": {},
   "source": [
    "### Gradient Boosting Classifier "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 915,
   "id": "8c30b343",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best Parameters:  {'learning_rate': 1, 'max_depth': 3, 'n_estimators': 100}\n",
      "Accuracy Score:  0.9394248877299726\n"
     ]
    }
   ],
   "source": [
    "# Gradient Boosting Classifier \n",
    "from sklearn.ensemble import GradientBoostingClassifier\n",
    "model = GradientBoostingClassifier()\n",
    "forest_params = { \"learning_rate\": [.001,.01,.1,.5,1], \n",
    "                 \"max_depth\":[1,3,5,7,9,11,13,15],  \n",
    "                 \"n_estimators\":[100,500]}\n",
    "clf = GridSearchCV(model, forest_params, cv = 10, scoring='accuracy')\n",
    "clf.fit(X_train, y_train)\n",
    "print('Best Parameters: ', clf.best_params_)\n",
    "print('Accuracy Score: ', clf.best_score_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 115,
   "id": "a53d0e8b",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.9343936381709742\n",
      "Classification Error: 0.0656063618290258\n",
      "Mean Cross Val (Accuracy) Score 0.9293014925373134\n",
      "Mean Classification Error: 0.07069850746268658\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAARoAAADgCAYAAADczalxAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuNCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8QVMy6AAAACXBIWXMAAAsTAAALEwEAmpwYAAAstklEQVR4nO2dd3hVVdaH30XovaM0qULoVQcUKSKgWBFBQCnqoA7oKDY+xdERu1gGARUVUUfBUWw4KAojgoJSJFKliJQgSihC6ISs7499Ei4h5aac3JL1Ps99cvY5++yz7k3uL3uvvdfaoqoYhmH4SaFQG2AYRvRjQmMYhu+Y0BiG4TsmNIZh+I4JjWEYvmNCYxiG75jQGIbhOyY0BRAR2Swih0XkgIj8LiJTRaR0mjodReR/IpIoIvtEZKaINElTp6yIvCAiW722Nnrlyvn7joxwx4Sm4HKZqpYGWgGtgf9LuSAiHYAvgU+A6kBd4CfgOxGp59UpCswFmgK9gLJAR2A3cI5fRotIYb/aNvzDhKaAo6q/A7NxgpPC08BbqvovVU1U1T2qOgb4HnjYqzMYqA1cpaprVDVZVXeq6lhVnZXes0SkqYh8JSJ7ROQPEbnfOz9VRB4NqNdFROIDyptF5D4RWQEcFJExIvJBmrb/JSLjveNyIvK6iOwQke0i8qiIxOTukzJygwlNAUdEagIXAxu9cklcz+T9dKr/B7jIO+4OfKGqB4J8ThlgDvAFrpfUANcjCpYBQG+gPPA2cImIlPXajgH6Ae96dd8EkrxntAZ6ADdl41lGHmNCU3D5WEQSgW3ATuAh73xF3N/FjnTu2QGk+F8qZVAnIy4FflfVZ1X1iNdT+iEb949X1W2qelhVtwA/Ald617oBh1T1exGphhPOO1T1oKruBJ4Hrs3Gs4w8xoSm4HKlqpYBugCNOSkge4Fk4Mx07jkT2OUd786gTkbUAn7JkaWObWnK7+J6OQADOdmbOQsoAuwQkT9F5E/gFaBqLp5t5BITmgKOqn4DTAXGeeWDwCLgmnSq9+PkcGcO0FNESgX5qG1A/QyuHQRKBpTPSM/UNOX3gS7e0O8qTgrNNuAoUFlVy3uvsqraNEg7DR8woTEAXgAuEpFWXnk0MEREbheRMiJSwXPWdgD+6dV5G/elniEijUWkkIhUEpH7ReSSdJ7xGXCGiNwhIsW8ds/1rsXhfC4VReQM4I6sDFbVBGAe8Abwq6qu9c7vwM2YPetNvxcSkfoi0jmbn4mRh5jQGClf2reAB73yt0BPoA/OD7MF51Q9X1U3eHWO4hzCPwNfAfuBxbgh2Gm+F1VNxDmSLwN+BzYAXb3Lb+OmzzfjROK9IE1/17Ph3TTnBwNFgTW4oeAHZG+YZ+QxYomvDMPwG+vRGIbhOyY0hmH4jgmNYRi+Y0JjGIbvmNAYhuE7ERcJW7lyZa1Tp06ozTCMAseyZct2qWqVnNwbcUJTp04dli5dGmozDKPAISJbcnqvDZ0Mw/AdExrDMHzHN6ERkSkislNEVmVwXURkvJf+cYWItPHLFsMwQoufPZqpuBSPGXEx0NB7DQde8tEWwzBCiG/OYFWdLyJ1MqlyBS5dpALfi0h5ETnTi76NXub1ht/SzXRpGGHLloTaubo/lLNONTg1mVG8d+40oRGR4bheD7Vr5+4NhwQTFyNCUYWX5tzK3e+OA4JNPXQ6oRQaSedcuqHkqjoZmAzQrl278Ag3z414VL8Euvw3b+0xjDzmt9/ghhtg9uzctxVKoYnHpXdMoSbwW4hsOZ287oWYuBgRxAcfwM03w549ULEivPIKXJNezsUgCaXQfAqMFJHpwLnAvrDyzwQjMiYeRhQyejQ89ZQ77tULpkyBM3OZNsw3oRGRabjE15W9PXoewiWNRlVfBmYBl+C2+TgEDPPLllwxMDxGaoaRX1x8MUyYAM88A7fcApKekyOb+DnrNCCL6wqM8Ov5OcKctkYB5OhR54e5/HJX7twZtmyBSpXy7hm2MhicwLwrp4tM9fRybBtG9LBiBbRvD1dcAXMDtvPLS5GBCAyq9IVAgTG/i1EAOHECnn8eHngAjh2DBg2gTBn/nlewhSbtUMn8MUYBYMsWGDIEvvnGlW+5BcaNg1I5XyaTJQVPaDLyw9gwySgAfP01XHkl7N8P1aq5GaVL8uFPv2AJTXoiY0MlowDRvDmULAkXXgiTJ0PlylnfkxcULKFJERkTF6MAMW8edOwIRYs6YVm6FKpXz5tp62ApOLNO83qfPDaRMQoABw/CiBHQtSs88sjJ8zVq5K/IQEHq0QT2Zgwjylm8GK6/HtavhyJFoHz50NoT3UKTnk/GejNGFHP8ODz+OIwd66awmzWDf/8bWrYMrV3RJTRZrey13owRxezZ48IHFi925VGj4LHHoHjx0NoF0SI0mQmMOX6NAkKFCi7SulYtmDoVunULtUUniQ6hsZW9RgFlxw63svess5yD98033exSqH0yaYkOoUnBVvYaBYgPP4Thw6FRI5g/H2JioGrVUFuVPgVnetswooR9+2DoULj6ati9G0qXhsTEUFuVOSY0hhFBzJ/vZpDefNM5eSdMgC++CL+hUloiX2gCF+IZRhTz4IPQpYsLimzXDpYvdwvy8nvxXU6IfKGxhXhGAaFsWScqDz4ICxdC48ahtih4oscZbDNNRpSRnOxW9qYIyqhR0KNH6Bff5YTI79EYRhSyZYuLsO7QAeLj3bmYmMgUGTChMYywQhXefhtatHBR10WLOtGJdCJbaMwRbEQRu3dDv34weLBLTHXllbBqFZx3Xqgtyz2R7aMxR7ARJcybBwMHupW+pUvD+PFurUwkzCgFg689GhHpJSLrRGSjiIxO53o5EZkpIj+JyGoRydneTuYINiKcEiVg5044/3y3M8GwYdEjMuCj0IhIDDARuBhoAgwQkSZpqo0A1qhqS9xmc8+KSNGgHmDDJiPCCfS9nHuu69XMmwd164bKIv/ws0dzDrBRVTep6jFgOnBFmjoKlBERAUoDe4CkoFq3YZMRoSQluYx3DRrAzJknz59/vptZikaCFhoRye5mDDWAbQHleO9cIBOAWOA3YCXwd1VNztZTbNhkRBAbNjhBeeghJzg//RRqi/KHLIVGRDqKyBpgrVduKSKTgmg7vRFm2vDqnkAcUB1oBUwQkbLp2DBcRJaKyNKEhAQbNhkRhyq88gq0agU//AA1a7qdIceMCbVl+UMwPZrncYKwG0BVfwIuCOK+eKBWQLkmrucSyDDgQ3VsBH4FTltYraqTVbWdqrarUqWKDZuMiCIhAS691G3UdugQDBoEK1eGV2Iqvwlq6KSq29KcOhHEbUuAhiJS13PwXgt8mqbOVuBCABGpBjQCNgVjE2DDJiMiKFrUrYepUAGmT3c5fMM92jqvCWYdzTYR6QioJxi34w2jMkNVk0RkJDAbiAGmqOpqEbnFu/4yMBaYKiIrcUOt+1R1Vw7fi2GEDfv3u90HSpSAcuVckqozznBbnRREghGaW4B/4Ry58cCXwN+CaVxVZwGz0px7OeD4N6BHsMYaRiSwYIFb3XvFFfDCC+5c27YhNSnkBDN0aqSqg1S1mqpWVdXrcDNFhmEEcPQojB4NnTvD5s3w3XfunBGc0LwY5DnDKLCsWuUW3T31lFvRO2aME5pixUJtWXiQ4dBJRDoAHYEqIjIq4FJZnM/FMAo8qvD88/B//+d2I6hf30Vfd+gQasvCi8x6NEVxq3ULA2UCXvuBvv6bZhiRwcKFTmSGD4e4OBOZ9MiwR6Oq3wDfiMhUVY2CjBiGkTeoulmlcuXcMOnll10QZG9bR5ohwcw6HRKRZ4CmQOrmmqpagJYbGYZjzx638G7jRvj+e7dGpnJlE5msCMYZ/A7wM1AX+CewGbcYzzAKFLNnQ7Nm8P77LmapoMQp5QXBCE0lVX0dOK6q36jqDcBffLbLMMKGQ4fgttugVy+XmOq885zItG8fassih2CGTse9nztEpDcuXqmmfyYZRviwdClcdx2sWweFC7v0DvfeG73pHPwiGKF5VETKAXfh1s+UBe7w06hMSdwQskcbBY8lS5zINGniYpRatw61RZFJlkKjqp95h/uArgAiErp0ycf3u58WuW34xOHDLkYJnOO3UCEXUpByzsg+GfpoRCRGRAaIyN0i0sw7d6mILMQlrAotFrlt5DGqMHmyS6X5yy/unAjcfLOJTG7JzBn8OnATUAkYLyJvAOOAp1XVOpBGVPH773DZZU5U/vjDpXMw8o7Mhk7tgBaqmiwixYFdQANV/T1/TDOM/OHjj+Gvf4Vdu1yemJdegmuvDbVV0UVmPZpjKfl7VfUIsN5ExogmEhPhxhvhqqucyHTv7jLfmcjkPZn1aBqLyArvWID6XlkAVdUWvltnGD6yebObSSpeHJ5+GkaMcI5fI+/JTGgs54wRdSQlufUwAM2bw5Qpbsq6Sdodx4w8JbOgSgukNKKKVavc4rvRo08OjwYNCq1NBQXrKBpRT3KyyxnTrp0LHRg3zk1lG/mHCY0R1WzbBhddBKNGubSaN90EX38dXftaRwJBCY2IlBCRRn4bYxh5hSq8+67zw/zvf1ClCnzyCbz6KpQpE2rrCh7B7FR5GW43yS+8cisRSbs/k2GEFceOwT//Cfv2weWXO//M5ZeH2qqCSzA9moeBc4A/AVQ1DqgTTOMi0ktE1onIRhEZnUGdLiISJyKrReSbYNo1jIxI9nZuL1bM5e599VW3IK9q1ZCaVeAJJno7SVX3STYHtSISA0wELsLtB7VERD5V1TUBdcoDk4BeqrpVROzPwcgRhw+72aTDh128EsA557iXEXqC6dGsEpGBQIyINBSRF4GFQdx3DrBRVTep6jFgOnBFmjoDcXtvbwVQ1Z3ZsN0wAFi2DNq0gfHjYerUkwGRRvgQjNDchssXfBR4F5cu4o4g7qsBBO7ZHe+dC+RsoIKIzBORZSIyOIh2DQNwi+8eewz+8hf4+WeIjXV5fOvXD7VlRlqCGTo1UtUHgAey2XZ6Y620qxcKA22BC4ESwCIR+V5V15/SkMhwYDhA27rZtMKISn75Ba6/HhYtcuW//x2eeMLSOYQrwfRonhORn0VkrIg0zUbb8UCtgHJNXBrQtHW+UNWDqroLmA+0TNuQqk5W1Xaq2i4bzzeimKeeciJTowZ89ZXb49pEJnzJUmhUtSvQBUgAJovIShEZE0TbS4CGIlJXRIoC1wJpp8U/ATqJSGERKQmcC6zNzhswCg6Bq3mfecYlDF+50kVdG+FNUAv2VPV3VR0P3IJbU/OPIO5JAkYCs3Hi8R9VXS0it4jILV6dtbj1OSuAxcBrqroqS4MsjWeB4+OPoUsXN6sEbvO28eOhQoVQWmUEi2gWQR8iEgv0x22Duxs3ezQjVDNE7eqJLt1kgSoFhcREuOMOF2UN8OKLMHJkSE0qsIjIspy6L4JxBr8BTAN6qGpaH4th+Ma337qk4L/+6hbgPfUU/O1vobbKyAnB7IJgm8UZ+cqxY/DQQ05YVKFVK3jnHcsZE8lkKDQi8h9V7SciKzl1Wtoy7Bm+MmsWPPmky3Y3ejQ8/LDb49qIXDLr0fzd+3lpfhhiGClccQXceSf06QPnnx9qa4y8IMNZJ1Xd4R3+TVW3BL4AGykbeca2bdC7N6zxouBE4LnnTGSiiWCmty9K59zFeW2IUfAIzBkza5brxRjRSWY+mltxPZd6AbshAJQBvvPbMCO62bPHzSC9954rX3opvPZaaG0y/CMzH827wOfAE0BgLplEVd3jq1VGVPPVVzBsGGzfDqVKuXy+N91k6TWjmcyERlV1s4iMSHtBRCqa2Bg5Yfdut2HbwYMu6vrtt6FBg1BbZfhNVj2aS4FluOntwP83CtTz0S4jSqlUyfVg/vjDTV0XDmbJqBHxZBmCEG5YCEJkkZTkdoE84wy44YZQW2PkBl9DEETkPCBOVQ+KyHVAG+CFlKx4hpERv/ziQggWLoTSpeHKK6FixVBbZYSCYKa3XwIOiUhL4F5gC/C2r1YZEY2qm0Fq2dKJTPXqMGOGiUxBJhihSVI3vroC+Jeq/gs3xW0Yp/HHH25l71//6hy+/fq5nDE9eoTaMiOUBOOKSxSR/wOuxyWpigGK+GuWEalcdx3MmePyxUyaBAMG2LS1EVyPpj8uMfkNqvo7LsH4M75aZUQszz4LF1/sejEDB5rIGI6gZp1EpBrQ3isuDuW2KDbrFF58953LfveM/euJenIz6xTMlrj9cGk2rwH6AT+ISN+cPMyIHo4dgwcegAsugHHj4FPbJNnIhGB8NA8A7VN6MSJSBZgDfOCnYUb4smaN88UsX+6GRqNHQ8+eobbKCGeCEZpCaYZKuwkyqbkRXSQnu5y9990HR49CnTrw1lvQqVOoLTPCnWCE5gsRmY3LGwzOOTzLP5OMcGXCBJcoHNwq3+efh7JlQ2qSESEE6wzuA5yPi3ear6of+W1YRpgzOHQcPOiGSHfd5QIjjYKFL85gEWkoIp+IyCqcI/hZVb0zOyIjIr1EZJ2IbBSR0ZnUay8iJ8zJHF7s3euSUSUmunKpUrBggYmMkX0y87VMAT4DrsZFcL+YnYa9hX0Tcdn4mgADROS0PPZevadwG80ZYcKcOS7z3QsvOJ9MCrYuxsgJmQlNGVV9VVXXqeo4oE422z4H2Kiqm1T1GG7juSvSqXcbMAMI2doc4ySHDzs/zEUXucRUf/kLjBoVaquMSCczZ3BxEWnNyTw0JQLLqvpjFm3XALYFlONxe2unIiI1gKuAbpxcEGiEiB9/dNPWa9e6PDEPP+x6M5Yzxsgtmf0J7QCeCyj/HlBWnDhkRnqd7LRe3BeA+1T1hGTSJxeR4cBwgLZ1s3iqkSM2bHC9l+PHoVEj+Pe/oV2O3H6GcToZCo2qds1l2/FArYByTSDtlrrtgOmeyFQGLhGRJFX9OI0tk4HJ4GadcmmXkQ4NG8KgQVCmjNu8rWTJUFtkRBN+doqXAA1FpC6wHbgWGBhYQVVT+yciMhX4LK3IGP6gClOmuO1m27Z1515/3e0OaRh5jW9Co6pJIjISN5sUA0xR1dUicot3/WW/nm1kzs6dMHw4fPIJNG4McXFQrJiJjOEfvrr5VHUWaVYRZyQwqjrUT1sMx8yZbmuTnTtdzpgxY2xfa8N/goneFhG5TkT+4ZVri8g5/ptm5CUHDrhezOWXO5Hp2hVWrHB+GVsbY/hNMJ3lSUAHYIBXTsQtxDMihORk6NwZXn3VDZGee84tyKtdO9SWGQWFYITmXFUdARwBUNW9gHW2I4hChVwoQcuWsHSpOzZ/jJGfBPPndtwLE1BIzUeT7KtVRq5ZuxamTz9ZHjQIliyBZs1CZ5NRcAlGaMYDHwFVReQx4FvgcV+tMnJMcjKMHw9t2sDQobB6tTsvAkUspbwRIrKcdVLVd0RkGXAhbrXvlaq61nfLjGwTHw/Dhjn/C7jjWrUyv8cw8oNgdqqsDRwCZgaes50qw4v33oNbboE//4TKlWHyZEvnYIQPwayj+S/OPyNAcaAusA5o6qNdRjZ49FF48EF33Lu32yXyjDNCa5NhBJKlj0ZVm6tqC+9nQ1z6h2/9N80Ilv79oWpVePlltyDPRMYIN7K9MlhVfxQRS+kQQo4ccdHVN97onLwNG8Kvv1ogpBG+BOOjCUx7VAhoAyT4ZpGRKcuXu5wxa9a48k03uZ8mMkY4E8z0dpmAVzGczya9THmGj5w44dI3nHuuE5lGjdwCPMOIBDLt0XgL9Uqr6j35ZI+RDps2weDBbvtZgJEj4amnrBdjRA4ZCo2IFPZSPbTJT4OMU1m8GC680AVFnnkmvPGG7QppRB6Z9WgW4/wxcSLyKfA+cDDloqp+6LNtBm54VK+eGyq99BJUqhRqiwwj+wQz61QRtw1uN06up1HAhMYnZs1y+XsrVnTR1t9843LHWDoHI1LJTGiqejNOqzgpMClY3l4fOHDAbW3y6qtubcy0aU5cypcPtWWGkTsyE5oYoDTB7WZg5JJFi+D66+GXX1zGu/a2UsmIIjLdbkVVH8k3Swoox4/DI4/A44+7yOsWLdxivObNQ22ZYeQdmQmNeQR85sgR6NTJJaMSgXvvdaJTrFioLTOMvCUzobkw36wooBQv7hbgJSTAW2/BBReE2iLD8AdRjSx3S7t6oks3RZbNgWzf7oSlVStXPnQIkpKgbNmQmmUYWSIiy1Q1R/uX+po5VkR6icg6EdkoIqPTuT5IRFZ4r4UiEtWL6v/zH+d76dMHEhPduZIlTWSM6Mc3ofHCFyYCFwNNgAEi0iRNtV+BzqraAhiLt+1ttPHnny4Qsn9/2LsXYmOdf8YwCgp+9mjOATaq6iZVPQZMJ00wpqou9HZVAPgetz93VPH1124m6Z13XO/l5Zfhs8+gSpVQW2YY+YefQlMD2BZQjvfOZcSNwOc+2pPvjB0L3brBtm3O6RsXBzffbCt8jYKHn0IT9EI/EemKE5r7Mrg+XESWisjSPLTPd5o2hcKF3ZT1t9+6BFWGURDxc+/teCAwB39N4Le0lUSkBfAacLGq7k6vIVWdjOe/aVdPwnbK6cQJF23doYMr9+kD69dD3bqhtcswQo2fPZolQEMRqSsiRYFrgU8DK3g7LHwIXK+q6320xXc2b3b7WV9wASxbdvK8iYxh+Nij8XLZjARm4+KmpqjqahG5xbv+MvAPoBIwSZzjIimn8/ShQhXefBNuv91NWZ9xxsmpa8MwHLZgLxckJDjn7kcfufLVV8Mrr1jOGCM6yc2CPT99NFHNd985YfnjD7fgbsIEt1bGZpQM43RMaHJIrVpw+DB07uyGTmedFWqLDCN8MaHJBitXuinrQoWgdm1YuBAaN4aYmFBbZhjhjQlNEBw/7radfewxGDcO7rjDnW+azU2Bjx8/Tnx8PEcs/sAIY4oXL07NmjUpUqRInrVpQpMF69Y530tKzpjd6a70CY74+HjKlClDnTp1EHPmGGGIqrJ7927i4+Opm4drM3yN3o5kVGHiRGjd2olM7doubmns2Jy3eeTIESpVqmQiY4QtIkKlSpXyvNdtPZp02LsXBgyA2bNdefBgGD/e7USQW0xkjHDHj79RE5p0KFPGpXaoWNGti+nbN9QWGUZkY0Mnj337YNcud1y4MEyfDqtWRZ/IxMTE0KpVK5o1a8Zll13Gn3/+mXpt9erVdOvWjbPPPpuGDRsyduxYAhd0fv7557Rr147Y2FgaN27M3XffHYJ3kDnLly/npptuCrUZGXL06FH69+9PgwYNOPfcc9m8eXO69d577z1atGhB06ZNuffee1PPb926la5du9K6dWtatGjBrFmzTrnWo0cPYmNjadKkSWrbN954Iy1btqRFixb07duXAwcOAPDZZ5/x0EMP+fZeT0FVI+rVti6a13z9tWrt2qqXXqqanJznzaeyZs0a/xoPklKlSqUeDx48WB999FFVVT106JDWq1dPZ8+eraqqBw8e1F69eumECRNUVXXlypVar149Xbt2raqqHj9+XCdOnJinth0/fjzXbfTt21fj4uLy9ZnZYeLEiXrzzTerquq0adO0X79+p9XZtWuX1qpVS3fu3Kmq7vc0Z84cVVX961//qpMmTVJV1dWrV+tZZ52Vel/nzp31yy+/VFXVxMREPXjwoKqq7tu3L7XOnXfeqU888YSqqiYnJ2urVq1S6wWS3t8qsFRz+L0t0D2aI0fg7rtdzpitW90q33378unh74o/r2zQoUMHtm/f7sx5913OO+88evToAUDJkiWZMGECTz75JABPP/00DzzwAI0bNwagcOHC/O1vfzutzQMHDjBs2DCaN29OixYtmDFjBgClS5dOrfPBBx8wdOhQAIYOHcqoUaPo2rUr99xzD3Xq1Dmll9WgQQP++OMPEhISuPrqq2nfvj3t27fnu+++O+3ZiYmJrFixgpYtXUbYxYsX07FjR1q3bk3Hjh1Zt24dAFOnTuWaa67hsssuo0ePHhw8eJAbbriB9u3b07p1az755BMANm/eTKdOnWjTpg1t2rRh4cKF2fp80+OTTz5hyJAhAPTt25e5c+ee0msE2LRpE2effTZVvOxo3bt3T/0cRYT9+/cDsG/fPqpXrw7AmjVrSEpK4qKLLgLc512yZEkAynq5YlWVw4cPp/pgRIQuXbrw2Wef5fp9ZUWB9dGsWOGmrVeudAvuxoyBBx6APFw6ENacOHGCuXPncuONNwJu2NS2bdtT6tSvX58DBw6wf/9+Vq1axV133ZVlu2PHjqVcuXKsXLkSgL1792ZxB6xfv545c+YQExNDcnIyH330EcOGDeOHH36gTp06VKtWjYEDB3LnnXdy/vnns3XrVnr27MnatWtPaWfp0qU0a9Ystdy4cWPmz59P4cKFmTNnDvfff3/qF3bRokWsWLGCihUrcv/999OtWzemTJnCn3/+yTnnnEP37t2pWrUqX331FcWLF2fDhg0MGDCApUtPT4nUqVMnEtOJpB03bhzdu3c/5dz27dupVctlTylcuDDlypVj9+7dVK5cObVOgwYN+Pnnn9m8eTM1a9bk448/5tixYwA8/PDD9OjRgxdffJGDBw8yZ86c1M+wfPny9OnTh19//ZXu3bvz5JNPEuOtJh02bBizZs2iSZMmPPvss6nPateuHQsWLKBfv35Z/p5yQ4EUmmefhfvvh2PHXDKqt992GfDylYGhCQw9fPgwrVq1YvPmzbRt2zb1P6CqZjjbkJ1ZiDlz5jB9+vTUcoUKFbK855prrkn9QvTv359HHnmEYcOGMX36dPr375/a7po1a1Lv2b9/P4mJiZQpUyb13I4dO1J7AeD+4w8ZMoQNGzYgIhw/fjz12kUXXUTFihUB+PLLL/n0008ZN24c4JYhbN26lerVqzNy5Eji4uKIiYlh/fr0M5ksWLAgy/eYQtreC5z++VaoUIGXXnqJ/v37U6hQITp27MimTZsAmDZtGkOHDuWuu+5i0aJFXH/99axatYqkpCQWLFjA8uXLqV27Nv3792fq1Kmp/0jeeOMNTpw4wW233cZ7773HsGHDAKhatSq//XZamqg8p0AOnXbudCJz662wfHkIRCaElChRgri4OLZs2cKxY8eYOHEiAE2bNj3tv/WmTZsoXbo0ZcqUoWnTpiwLTLSTARkJVuC5tGs0SpUqlXrcoUMHNm7cSEJCAh9//DF9+vQBIDk5mUWLFhEXF0dcXBzbt28/RWRS3ltg2w8++CBdu3Zl1apVzJw585Rrgc9UVWbMmJHa9tatW4mNjeX555+nWrVq/PTTTyxdujS1V5GWTp060apVq9NeKb2NQGrWrMm2bS7DbVJSEvv27UsVvEAuu+wyfvjhBxYtWkSjRo1o6KVnfP3111N7Hx06dODIkSPs2rWLmjVr0rp1a+rVq0fhwoW58sor+fHHH09pMyYmhv79+6f26sD9LkqUKJHu+8pLCoTQqMKOHSfLjzwCc+fCpEkQ8PdWoChXrhzjx49n3LhxHD9+nEGDBvHtt9+mfjkOHz7M7bffnjrjcc899/D444+n/ldPTk7mueeeO63dHj16MGHChNRyytCpWrVqrF27NnVolBEiwlVXXcWoUaOIjY2lkpdzI227cXFxp90bGxvLxo0bU8v79u2jRg2Xpnrq1KkZPrNnz568+OKLqb2N5cuXp95/5plnUqhQId5++21OnDiR7v0LFixIFanAV9phE8Dll1/Om2++CThfVbdu3dIV5p07dwLu85s0aVLqTFrt2rWZO3cuAGvXruXIkSNUqVKF9u3bs3fvXhISEgD43//+R5MmTVDV1M9EVZk5c2aqnw3ckCtwuOkbOfUih+qV3VmnhATVPn1Uq1dX3b07W7fmOeE266Sqeumll+pbb72lqqorVqzQzp0769lnn63169fXhx9+WJMDpuFmzpypbdq00caNG2tsbKzefffdp7WfmJiogwcP1qZNm2qLFi10xowZqqr6/vvva7169bRz5846YsQIHTJkiKqqDhkyRN9///1T2liyZIkCOnXq1NRzCQkJ2q9fP23evLnGxsamztykpVmzZrp//35VVV24cKE2bNhQO3bsqGPGjEmdoXnjjTd0xIgRqfccOnRIhw8frs2aNdOmTZtq7969VVV1/fr12rx5cz333HN19OjRp312OeHw4cPat29frV+/vrZv315/+eWX1GstW7ZMPb722ms1NjZWY2Njddq0aannV69erR07dtQWLVpoy5YtU2cJVVW//PJLbd68uTZr1kyHDBmiR48e1RMnTmjHjh1T39vAgQNPmYXq3bu3rlix4jQ783rWKeTCkd1XdoTmv/9VrVbNvcsyZdw0digJB6GJdp577jl99dVXQ21GRPD7779rt27d0r1m09tBcPCg87/07u2mrDt1crNMXbqE2jLDb2699VaKFSsWajMigq1bt54yA+UnUTfrtGQJDBoEGza4qerHHoNRoyxnTEGhePHiXH/99aE2IyJo3759vj0r6oRmzx4nMs2awb//DS3DbDdvzWQa2TDCAU1nCj63RMXQKXBNWM+e8OGHrmcTbiJTvHhxdu/e7csv0jDyAlWXj6Z48eJ52m5E92hU4aWX4L774PPP4fzz3fmrrgqtXRlRs2ZN4uPjU6cgDSMcScmwl5f4KjQi0gv4F25fp9dU9ck018W7fglwCBiqqj+e1lA67NgBN9wAX3zhyv/970mhCVeKFCmSp1nLDCNS8G3oJCIxwETgYqAJMEBEmqSpdjHQ0HsNB14Kpu0ZM5wP5osvXM6Y//wHnngiD403DCNP8dNHcw6wUVU3qeoxYDpwRZo6VwBvedP03wPlReTMzBrdnFCHvn2d07dnTxcUec01/rwBwzDyBj+FpgawLaAc753Lbp1T2HeoHCVKuA3bPv8cvCh5wzDCGD99NOnN4aadbgmmDiIyHDe0AjiadFhWjRwJI0fm0sL8ozKwK9RGZINIsxfM5vygUU5v9FNo4oFaAeWaQNp49GDqoKqTgckAIrJUc7j/b6iINJsjzV4wm/MDETk9GU+Q+Dl0WgI0FJG6IlIUuBb4NE2dT4HB4vgLsE9Vd6RtyDCMyMa3Ho2qJonISGA2bnp7iqquFpFbvOsvA7NwU9sbcdPbw/yyxzCM0OHrOhpVnYUTk8BzLwccKzAim81OzgPT8ptIsznS7AWzOT/Isb1iy+ENw/CbqIh1MgwjvAlboRGRXiKyTkQ2isjodK6LiIz3rq8QkTahsDPAnqzsHeTZuUJEFopIyEM+s7I5oF57ETkhIiHfTi8Ym0Wki4jEichqEfkmv21MY0tWfxflRGSmiPzk2RtyP6WITBGRnSKyKoPr2f/u5TRjlp8vnPP4F6AeUBT4CWiSps4lwOe4tTh/AX4Ic3s7AhW844tDaW+wNgfU+x/O19Y33G0GygNrgNpeuWqY23s/8JR3XAXYAxQN8ed8AdAGWJXB9Wx/98K1R+NL+IKPZGmvqi5U1ZSEFt/j1gyFkmA+Y4DbgBnAzvw0LgOCsXkg8KGqbgVQ1VDaHYy9CpTxAoxL44QmKX/NTGOQ6nzPjozI9ncvXIXGl/AFH8muLTfi/iOEkixtFpEawFXAy4QHwXzOZwMVRGSeiCwTkcH5Zt3pBGPvBCAWt1B1JfB3VU3OH/NyTLa/e+GajybPwhfyiaBtEZGuOKEJdVKLYGx+AbhPVU+ESVbAYGwuDLQFLgRKAItE5HtVTX/3N38Jxt6eQBzQDagPfCUiC1R1v8+25YZsf/fCVWjyLHwhnwjKFhFpAbwGXKyqu/PJtowIxuZ2wHRPZCoDl4hIkqp+nC8Wnk6wfxe7VPUgcFBE5gMtgVAITTD2DgOeVOf82CgivwKNgcX5Y2KOyP53L5ROp0ycUYWBTUBdTjrRmqap05tTHVKLw9ze2rgV0B1D/fkGa3Oa+lMJvTM4mM85Fpjr1S0JrAKahbG9LwEPe8fVgO1A5TD4+6hDxs7gbH/3wrJHoxEWvhCkvf8AKgGTvB5CkoYwoC5Im8OKYGxW1bUi8gWwAkjGZXZMd5o2HOwFxgJTRWQl7ot7n6qGNKJbRKYBXYDKIhIPPAQUgZx/92xlsGEYvhOus06GYUQRJjSGYfiOCY1hGL5jQmMYhu+Y0BiG4TsmNBGKF00dF/Cqk0ndA3nwvKki8qv3rB9FpEMO2ngtZW8vEbk/zbWFubXRayflc1nlRUWXz6J+KxG5JC+ebWSMTW9HKCJyQFVL53XdTNqYCnymqh+ISA9gnKq2yEV7ubYpq3ZF5E1gvao+lkn9oUA7VY2cPTUiEOvRRAkiUlpE5nq9jZUicloktoicKSLzA/7jd/LO9xCRRd6974tIVgIwH2jg3TvKa2uViNzhnSslIv/1cqysEpH+3vl5ItJORJ4ESnh2vONdO+D9fC+wh+H1pK4WkRgReUZElng5UG4O4mNZhBfsJyLniMsDtNz72Uhc0vxHgP6eLf0926d4z1me3udo5IBQL3W2V46XiJ/ABePFAR/hlruX9a5Vxq3aTOmxHvB+3gU84B3HAGW8uvOBUt75+4B/pPO8qXghCMA1wA+44MWVQClcioPVQGvgauDVgHvLeT/n4XoPqTYF1Emx8SrgTe+4KC5KuARuX68x3vliwFKgbjp2Hgh4f+8DvbxyWaCwd9wdmOEdDwUmBNz/OHCdd1weFyNVKtS/70h/hWUIghEUh1W1VUpBRIoAj4vIBbil9zVwsTO/B9yzBJji1f1YVeNEpDNub/TvvNCIorieQHo8IyJjgARcBPqFwEfqAhgRkQ+BTsAXwDgReQo33FqQjff1OTBeRIoBvYD5qnrYG661kJNZ/srh9mz/Nc39JUQkDherswz4KqD+myLSEBdpXCSD5/cALheRu71ycVyc2tpsvAcjDSY00cMgXIa2tqp6XEQ2474kqajqfE+IegNvi8gzwF7gK1UdEMQz7lHVD1IKItI9vUqqul5E2uLiYZ4QkS9V9ZFg3oSqHhGRebj0Cf2BaSmPA25T1dlZNHFYVVuJSDngM9wuG+NxMUVfq+pVnuN8Xgb3C3C1qq4Lxl4jOMxHEz2UA3Z6ItMVOCttBRE5y6vzKvA6Ll3j98B5IpLicykpImcH+cz5wJXePaVww54FIlIdOKSq/wbGec9Jy3GvZ5Ue03GBep1wAYl4P29NuUdEzvaemS6qug+4Hbjbu6ccLjIa3HAphUTcEDKF2cBt4nXvRKR1Rs8wgseEJnp4B2gnbtvSQcDP6dTpAsSJyHKcH+VfqpqA++JNE5EVOOFpHMwDVfVHnO9mMc5n85qqLgeaA4u9IcwDwKPp3D4ZWJHiDE7Dl7i8tXPUpcAEl8dnDfCjuKTZr5BFj9yz5SfcLqlP43pX3+H8Nyl8DTRJcQbjej5FPNtWeWUjl9j0tmEYvmM9GsMwfMeExjAM3zGhMQzDd0xoDMPwHRMawzB8x4TGMAzfMaExDMN3TGgMw/Cd/wdIEaQvcnzygAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 288x216 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Area Under Curve: 0.9862692134542415\n"
     ]
    }
   ],
   "source": [
    "# Gradient Boosting Classifier \n",
    "from sklearn.ensemble import GradientBoostingClassifier\n",
    "model = GradientBoostingClassifier(n_estimators=100, learning_rate=1, max_depth=3, random_state=1).fit(X_train, y_train)\n",
    "y_pred = model.predict(X_test)\n",
    "print(\"Accuracy:\",metrics.accuracy_score(y_test, y_pred))\n",
    "print(\"Classification Error:\",1-metrics.accuracy_score(y_test, y_pred))\n",
    "cv = RepeatedKFold(n_splits=5, n_repeats= 100, random_state=1)\n",
    "print(\"Mean Cross Val (Accuracy) Score\", np.mean(cross_val_score(model, X, y, cv=cv)))\n",
    "print(\"Mean Classification Error:\", 1 - np.mean(cross_val_score(model, X, y, scoring='accuracy', cv=cv, n_jobs=-1)))\n",
    "y_pred_proba = model.predict_proba(X_test)[:,1]\n",
    "fpr, tpr, _ = metrics.roc_curve(y_test, y_pred_proba)\n",
    "roc_auc = auc(fpr,tpr)\n",
    "plt.figure(figsize=(4,3))\n",
    "plt.plot(fpr, tpr, color='orange',\n",
    "lw=2, label='ROC curve (area = {0:.4f})'.format(roc_auc))\n",
    "plt.plot([0, 1], [0, 1], color='blue', lw=2, linestyle='--')\n",
    "plt.xlim([0.0, 1.0])\n",
    "plt.ylim([0.0, 1.05])\n",
    "plt.xlabel('False Positive Rate')\n",
    "plt.ylabel('True Positive Rate')\n",
    "plt.title('ROC curve')\n",
    "plt.legend(loc=\"lower right\")\n",
    "plt.show()\n",
    "print('Area Under Curve:', roc_auc)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "67ec0726",
   "metadata": {},
   "source": [
    "### Cat Boost Classifier "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c121c49b",
   "metadata": {},
   "source": [
    "- GridSearchCV was not working, we did trial an error to find the best hyperparameters. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 116,
   "id": "49542dbb",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.9383697813121272\n",
      "Classification Error: 0.06163021868787277\n",
      "Mean Cross Val (Accuracy) Score 0.9312358208955225\n",
      "Mean Classification Error: 0.06876417910447752\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x216 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Area Under Curve: 0.9851839346581773\n"
     ]
    }
   ],
   "source": [
    "#Cat Boost Classifier \n",
    "from catboost import CatBoostClassifier\n",
    "model = CatBoostClassifier(iterations=9,learning_rate=1,depth=10,logging_level='Silent').fit(X_train, y_train)\n",
    "y_pred = model.predict(X_test)\n",
    "print(\"Accuracy:\",metrics.accuracy_score(y_test, y_pred))\n",
    "print(\"Classification Error:\",1-metrics.accuracy_score(y_test, y_pred))\n",
    "cv = RepeatedKFold(n_splits=5, n_repeats= 100, random_state=1)\n",
    "print(\"Mean Cross Val (Accuracy) Score\", np.mean(cross_val_score(model, X, y, cv=cv)))\n",
    "print(\"Mean Classification Error:\", 1 - np.mean(cross_val_score(model, X, y, scoring='accuracy', cv=cv, n_jobs=-1)))\n",
    "y_pred_proba = model.predict_proba(X_test)[:,1]\n",
    "fpr, tpr, _ = metrics.roc_curve(y_test, y_pred_proba)\n",
    "roc_auc = auc(fpr,tpr)\n",
    "plt.figure(figsize=(4,3))\n",
    "plt.plot(fpr, tpr, color='orange',\n",
    "lw=2, label='ROC curve (area = {0:.4f})'.format(roc_auc))\n",
    "plt.plot([0, 1], [0, 1], color='blue', lw=2, linestyle='--')\n",
    "plt.xlim([0.0, 1.0])\n",
    "plt.ylim([0.0, 1.05])\n",
    "plt.xlabel('False Positive Rate')\n",
    "plt.ylabel('True Positive Rate')\n",
    "plt.title('ROC curve')\n",
    "plt.legend(loc=\"lower right\")\n",
    "plt.show()\n",
    "print('Area Under Curve:', roc_auc)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8e453c1d",
   "metadata": {},
   "source": [
    "### XGBoost Classifier"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c4243e44",
   "metadata": {},
   "source": [
    "- GridSearchCV was not working, we did trial an error to find the best hyperparameters. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 117,
   "id": "8e7306c2",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.9383697813121272\n",
      "Classification Error: 0.06163021868787277\n",
      "Mean Cross Val (Accuracy) Score 0.9333313432835821\n",
      "Mean Classification Error: 0.06666865671641786\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x216 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Area Under Curve: 0.9860280403884493\n"
     ]
    }
   ],
   "source": [
    "# XGBoost Classifier \n",
    "from xgboost import XGBClassifier\n",
    "model = XGBClassifier(learning_rate= 1, max_depth= 4, eta=1).fit(X_train, y_train)\n",
    "y_pred = model.predict(X_test)\n",
    "print(\"Accuracy:\",metrics.accuracy_score(y_test, y_pred))\n",
    "print(\"Classification Error:\",1-metrics.accuracy_score(y_test, y_pred))\n",
    "cv = RepeatedKFold(n_splits=5, n_repeats= 100, random_state=1)\n",
    "print(\"Mean Cross Val (Accuracy) Score\", np.mean(cross_val_score(model, X, y, cv=cv)))\n",
    "print(\"Mean Classification Error:\", 1 - np.mean(cross_val_score(model, X, y, scoring='accuracy', cv=cv, n_jobs=-1)))\n",
    "y_pred_proba = model.predict_proba(X_test)[:,1]\n",
    "fpr, tpr, _ = metrics.roc_curve(y_test, y_pred_proba)\n",
    "roc_auc = auc(fpr,tpr)\n",
    "plt.figure(figsize=(4,3))\n",
    "plt.plot(fpr, tpr, color='orange',\n",
    "lw=2, label='ROC curve (area = {0:.4f})'.format(roc_auc))\n",
    "plt.plot([0, 1], [0, 1], color='blue', lw=2, linestyle='--')\n",
    "plt.xlim([0.0, 1.0])\n",
    "plt.ylim([0.0, 1.05])\n",
    "plt.xlabel('False Positive Rate')\n",
    "plt.ylabel('True Positive Rate')\n",
    "plt.title('ROC curve')\n",
    "plt.legend(loc=\"lower right\")\n",
    "plt.show()\n",
    "print('Area Under Curve:', roc_auc)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "76e2606f",
   "metadata": {},
   "source": [
    "### Extra Trees Classifier "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 971,
   "id": "ff49a943",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best Parameters:  {'max_depth': 11, 'n_estimators': 500}\n",
      "Accuracy Score:  0.938548457192525\n"
     ]
    }
   ],
   "source": [
    "#Extra Trees Classifier \n",
    "from sklearn.ensemble import ExtraTreesClassifier\n",
    "model = ExtraTreesClassifier()\n",
    "forest_params = {\"max_depth\":[1,3,5,7,9,11,13,15],  \n",
    "                 \"n_estimators\":[100,500]}\n",
    "clf = GridSearchCV(model, forest_params, cv = 10, scoring='accuracy')\n",
    "clf.fit(X_train, y_train)\n",
    "print('Best Parameters: ', clf.best_params_)\n",
    "print('Accuracy Score: ', clf.best_score_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 118,
   "id": "5b4d25d2",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.9224652087475149\n",
      "Classification Error: 0.07753479125248508\n",
      "Mean Cross Val (Accuracy) Score 0.9319582089552239\n",
      "Mean Classification Error: 0.06804179104477615\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x216 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Area Under Curve: 0.9852723647823011\n"
     ]
    }
   ],
   "source": [
    "#Extra Trees Classifier \n",
    "from sklearn.ensemble import ExtraTreesClassifier \n",
    "model = ExtraTreesClassifier(n_estimators=500, max_depth = 11, random_state=1).fit(X_train, y_train)\n",
    "y_pred = model.predict(X_test)\n",
    "print(\"Accuracy:\",metrics.accuracy_score(y_test, y_pred))\n",
    "print(\"Classification Error:\",1-metrics.accuracy_score(y_test, y_pred))\n",
    "cv = RepeatedKFold(n_splits=5, n_repeats= 100, random_state=1)\n",
    "print(\"Mean Cross Val (Accuracy) Score\", np.mean(cross_val_score(model, X, y, cv=cv)))\n",
    "print(\"Mean Classification Error:\", 1 - np.mean(cross_val_score(model, X, y, scoring='accuracy', cv=cv, n_jobs=-1)))\n",
    "y_pred_proba = model.predict_proba(X_test)[:,1]\n",
    "fpr, tpr, _ = metrics.roc_curve(y_test, y_pred_proba)\n",
    "roc_auc = auc(fpr,tpr)\n",
    "plt.figure(figsize=(4,3))\n",
    "plt.plot(fpr, tpr, color='orange',\n",
    "lw=2, label='ROC curve (area = {0:.4f})'.format(roc_auc))\n",
    "plt.plot([0, 1], [0, 1], color='blue', lw=2, linestyle='--')\n",
    "plt.xlim([0.0, 1.0])\n",
    "plt.ylim([0.0, 1.05])\n",
    "plt.xlabel('False Positive Rate')\n",
    "plt.ylabel('True Positive Rate')\n",
    "plt.title('ROC curve')\n",
    "plt.legend(loc=\"lower right\")\n",
    "plt.show()\n",
    "print('Area Under Curve:', roc_auc)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b692ebc4",
   "metadata": {},
   "source": [
    "### Support Vector Classifier (SVC)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 980,
   "id": "d675f777",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best Parameters:  {'C': 10, 'decision_function_shape': 'ovo', 'gamma': 'scale'}\n",
      "Accuracy Score:  0.9377009995654063\n"
     ]
    }
   ],
   "source": [
    "#Support Vector Classifier (SVCs)\n",
    "from sklearn.svm import SVC\n",
    "model = SVC()\n",
    "forest_params = {'C': [0.1, 1, 10, 100, 1000],\n",
    "              'gamma': ['scale', 'auto'],\n",
    "               'decision_function_shape': ['ovo', 'ovr']}\n",
    "clf = GridSearchCV(model, forest_params, cv = 10, scoring='accuracy')\n",
    "clf.fit(X_train, y_train)\n",
    "print('Best Parameters: ', clf.best_params_)\n",
    "print('Accuracy Score: ', clf.best_score_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 119,
   "id": "14e58e87",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.9244532803180915\n",
      "Classification Error: 0.07554671968190851\n",
      "Mean Cross Val (Accuracy) Score 0.9087880597014926\n",
      "Mean Classification Error: 0.09121194029850743\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x216 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Area Under Curve: 0.9817190816129654\n"
     ]
    }
   ],
   "source": [
    "#Support Vector Machines (SVMs)\n",
    "from sklearn import svm\n",
    "model = svm.SVC(decision_function_shape='ovo', gamma = 'scale', C=10, probability=True).fit(X_train, y_train)\n",
    "y_pred = model.predict(X_test)\n",
    "print(\"Accuracy:\",metrics.accuracy_score(y_test, y_pred))\n",
    "print(\"Classification Error:\",1-metrics.accuracy_score(y_test, y_pred))\n",
    "cv = RepeatedKFold(n_splits=5, n_repeats= 100, random_state=1)\n",
    "print(\"Mean Cross Val (Accuracy) Score\", np.mean(cross_val_score(model, X, y, cv=cv)))\n",
    "print(\"Mean Classification Error:\", 1 - np.mean(cross_val_score(model, X, y, scoring='accuracy', cv=cv, n_jobs=-1)))\n",
    "y_pred_proba = model.predict_proba(X_test)[:,1]\n",
    "fpr, tpr, _ = metrics.roc_curve(y_test, y_pred_proba)\n",
    "roc_auc = auc(fpr,tpr)\n",
    "plt.figure(figsize=(4,3))\n",
    "plt.plot(fpr, tpr, color='orange',\n",
    "lw=2, label='ROC curve (area = {0:.4f})'.format(roc_auc))\n",
    "plt.plot([0, 1], [0, 1], color='blue', lw=2, linestyle='--')\n",
    "plt.xlim([0.0, 1.0])\n",
    "plt.ylim([0.0, 1.05])\n",
    "plt.xlabel('False Positive Rate')\n",
    "plt.ylabel('True Positive Rate')\n",
    "plt.title('ROC curve')\n",
    "plt.legend(loc=\"lower right\")\n",
    "plt.show()\n",
    "print('Area Under Curve:', roc_auc)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "285b852b",
   "metadata": {},
   "source": [
    "### Quadratic Discriminant Analysis"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 989,
   "id": "428811fb",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best Parameters:  {'reg_param': 0.001}\n",
      "Accuracy Score:  0.9215196291467478\n"
     ]
    }
   ],
   "source": [
    "# Quadratic Discriminant Analysis\n",
    "from sklearn.discriminant_analysis import QuadraticDiscriminantAnalysis\n",
    "model = QuadraticDiscriminantAnalysis()\n",
    "forest_params = {'reg_param': [.00001, .0001, .001,.01, 0.1, .5, 1]}\n",
    "clf = GridSearchCV(model, forest_params, cv = 10, scoring='accuracy')\n",
    "clf.fit(X_train, y_train)\n",
    "print('Best Parameters: ', clf.best_params_)\n",
    "print('Accuracy Score: ', clf.best_score_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 120,
   "id": "5c484168",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.9244532803180915\n",
      "Classification Error: 0.07554671968190851\n",
      "Mean Cross Val (Accuracy) Score 0.9119402985074627\n",
      "Mean Classification Error: 0.08805970149253728\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x216 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Area Under Curve: 0.9802398868094411\n"
     ]
    }
   ],
   "source": [
    "# Quadratic Discriminant Analysis\n",
    "from sklearn.discriminant_analysis import QuadraticDiscriminantAnalysis\n",
    "model = QuadraticDiscriminantAnalysis(reg_param = .001).fit(X_train, y_train)\n",
    "y_pred = model.predict(X_test)\n",
    "print(\"Accuracy:\",metrics.accuracy_score(y_test, y_pred))\n",
    "print(\"Classification Error:\",1-metrics.accuracy_score(y_test, y_pred))\n",
    "cv = RepeatedKFold(n_splits=5, n_repeats= 100, random_state=1)\n",
    "print(\"Mean Cross Val (Accuracy) Score\", np.mean(cross_val_score(model, X, y, cv=cv)))\n",
    "print(\"Mean Classification Error:\", 1 - np.mean(cross_val_score(model, X, y, scoring='accuracy', cv=cv, n_jobs=-1)))\n",
    "y_pred_proba = model.predict_proba(X_test)[:,1]\n",
    "fpr, tpr, _ = metrics.roc_curve(y_test, y_pred_proba)\n",
    "roc_auc = auc(fpr,tpr)\n",
    "plt.figure(figsize=(4,3))\n",
    "plt.plot(fpr, tpr, color='orange',\n",
    "lw=2, label='ROC curve (area = {0:.4f})'.format(roc_auc))\n",
    "plt.plot([0, 1], [0, 1], color='blue', lw=2, linestyle='--')\n",
    "plt.xlim([0.0, 1.0])\n",
    "plt.ylim([0.0, 1.05])\n",
    "plt.xlabel('False Positive Rate')\n",
    "plt.ylabel('True Positive Rate')\n",
    "plt.title('ROC curve')\n",
    "plt.legend(loc=\"lower right\")\n",
    "plt.show()\n",
    "print('Area Under Curve:', roc_auc)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0eb177da",
   "metadata": {},
   "source": [
    "### MLPClassifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 997,
   "id": "c423e27e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best Parameters:  {'alpha': 0.0001, 'learning_rate': 'constant', 'solver': 'adam'}\n",
      "Accuracy Score:  0.9215196291467478\n"
     ]
    }
   ],
   "source": [
    "# MLP Classifier\n",
    "from sklearn.neural_network import MLPClassifier\n",
    "model = MLPClassifier()\n",
    "forest_params = {\n",
    "    'alpha': [.0001, .001, 0.01, .1, .5, 1],\n",
    "    'learning_rate': ['constant'],\n",
    "    'solver': ['sgd', 'adam']\n",
    "}\n",
    "clf = GridSearchCV(model, forest_params, cv = 10, scoring='accuracy')\n",
    "clf.fit(X_train, y_train)\n",
    "print('Best Parameters: ', clf.best_params_)\n",
    "print('Accuracy Score: ', clf.best_score_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 122,
   "id": "cae5f472",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.9165009940357853\n",
      "Classification Error: 0.08349900596421467\n",
      "Mean Cross Val (Accuracy) Score 0.9013791044776119\n",
      "Mean Classification Error: 0.09847761194029836\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x216 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Area Under Curve: 0.9784552061225803\n"
     ]
    }
   ],
   "source": [
    "# MLP Classifier\n",
    "from sklearn.neural_network import MLPClassifier\n",
    "model = MLPClassifier(alpha= 0.0001, learning_rate= 'constant', solver= 'adam').fit(X_train, y_train)\n",
    "y_pred = model.predict(X_test)\n",
    "print(\"Accuracy:\",metrics.accuracy_score(y_test, y_pred))\n",
    "print(\"Classification Error:\",1-metrics.accuracy_score(y_test, y_pred))\n",
    "cv = RepeatedKFold(n_splits=5, n_repeats= 100, random_state=1)\n",
    "print(\"Mean Cross Val (Accuracy) Score\", np.mean(cross_val_score(model, X, y, cv=cv)))\n",
    "print(\"Mean Classification Error:\", 1 - np.mean(cross_val_score(model, X, y, scoring='accuracy', cv=cv, n_jobs=-1)))\n",
    "y_pred_proba = model.predict_proba(X_test)[:,1]\n",
    "fpr, tpr, _ = metrics.roc_curve(y_test, y_pred_proba)\n",
    "roc_auc = auc(fpr,tpr)\n",
    "plt.figure(figsize=(4,3))\n",
    "plt.plot(fpr, tpr, color='orange',\n",
    "lw=2, label='ROC curve (area = {0:.4f})'.format(roc_auc))\n",
    "plt.plot([0, 1], [0, 1], color='blue', lw=2, linestyle='--')\n",
    "plt.xlim([0.0, 1.0])\n",
    "plt.ylim([0.0, 1.05])\n",
    "plt.xlabel('False Positive Rate')\n",
    "plt.ylabel('True Positive Rate')\n",
    "plt.title('ROC curve')\n",
    "plt.legend(loc=\"lower right\")\n",
    "plt.show()\n",
    "print('Area Under Curve:', roc_auc)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1dc43ff2",
   "metadata": {},
   "source": [
    "# Final Results (Predictions on Test data)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ba0a70da",
   "metadata": {},
   "source": [
    "### Scale the Input Test Data "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "869e19b9",
   "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>K</th>\n",
       "      <th>S</th>\n",
       "      <th>r</th>\n",
       "      <th>tau</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>1120.0000</td>\n",
       "      <td>1120.0000</td>\n",
       "      <td>1120.0000</td>\n",
       "      <td>1120.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>0.5233</td>\n",
       "      <td>0.5143</td>\n",
       "      <td>0.3007</td>\n",
       "      <td>0.5133</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>0.1903</td>\n",
       "      <td>0.2509</td>\n",
       "      <td>0.2288</td>\n",
       "      <td>0.2656</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>0.3600</td>\n",
       "      <td>0.2717</td>\n",
       "      <td>0.1308</td>\n",
       "      <td>0.2857</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>0.5200</td>\n",
       "      <td>0.5899</td>\n",
       "      <td>0.2616</td>\n",
       "      <td>0.5204</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>0.6800</td>\n",
       "      <td>0.7195</td>\n",
       "      <td>0.3924</td>\n",
       "      <td>0.7245</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>1.0000</td>\n",
       "      <td>1.0000</td>\n",
       "      <td>1.0000</td>\n",
       "      <td>1.0000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               K          S          r        tau\n",
       "count  1120.0000  1120.0000  1120.0000  1120.0000\n",
       "mean      0.5233     0.5143     0.3007     0.5133\n",
       "std       0.1903     0.2509     0.2288     0.2656\n",
       "min       0.0000     0.0000     0.0000     0.0000\n",
       "25%       0.3600     0.2717     0.1308     0.2857\n",
       "50%       0.5200     0.5899     0.2616     0.5204\n",
       "75%       0.6800     0.7195     0.3924     0.7245\n",
       "max       1.0000     1.0000     1.0000     1.0000"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#All Datasets/Scaling Test Data\n",
    "option_test = pd.read_csv('option_test_wolabel.csv')\n",
    "option_test = option_test[['K','S','r','tau']]\n",
    "scaled = MinMaxScaler().fit(option_test)\n",
    "option_test = scaled.transform(option_test)\n",
    "option_test = pd.DataFrame(option_test, columns=['K','S','r','tau'])\n",
    "option_test = option_test + .0000000000001\n",
    "round(option_test.describe(),4)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "dbac6174",
   "metadata": {},
   "source": [
    "## Best Classifier "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "86295cfa",
   "metadata": {},
   "source": [
    "Our best classifier for this data was the XGBClassifier: \n",
    "\n",
    "- XGBClassifier(learning_rate= 1, max_depth= 4, eta=1)\n",
    "\n",
    "- Mean Cross Val (Accuracy) Score =  0.9333\n",
    "- Mean Classification Error = 0.0666"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a47ae546",
   "metadata": {},
   "source": [
    "### 'BS' Predictions "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "3b38e255",
   "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>BS</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>965</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>294</th>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>686</th>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     BS\n",
       "965   1\n",
       "294   0\n",
       "686   0"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#BS Predictions \n",
    "from xgboost import XGBClassifier\n",
    "model = XGBClassifier(learning_rate= 1, max_depth= 4, eta=1).fit(X_train, y_train)\n",
    "y_pred = model.predict(option_test)\n",
    "y_pred = pd.DataFrame(y_pred, columns=['BS'])\n",
    "y_pred.sample(3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "b82b5518",
   "metadata": {},
   "outputs": [],
   "source": [
    "preds = pd.read_excel('group_10_prediction_.xlsx')\n",
    "preds = preds.drop('BS', 1)\n",
    "preds['BS'] = y_pred\n",
    "preds.to_csv(\"group_10_prediction_.csv\", index=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2a4d9a93",
   "metadata": {},
   "source": [
    "## Best Regression Equation"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2321e0ca",
   "metadata": {},
   "source": [
    "Our best and final regression equation is:\n",
    "- Value = K + S + tau + K*S + K^2 + log(tau) + e^(K)+ 1/r + tau^K "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "bed797cc",
   "metadata": {},
   "source": [
    "### 'Value' Predictions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 412,
   "id": "a0a69a52",
   "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>Value</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>352</th>\n",
       "      <td>9.041225</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>690</th>\n",
       "      <td>18.069283</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>288</th>\n",
       "      <td>33.851223</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         Value\n",
       "352   9.041225\n",
       "690  18.069283\n",
       "288  33.851223"
      ]
     },
     "execution_count": 412,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Value Predictions \n",
    "model = smf.ols('Value ~ K + S + tau + \\\n",
    "K*S + \\\n",
    "np.power(K, 2) + \\\n",
    "np.log(tau) + \\\n",
    "np.exp(K) + \\\n",
    "np.power(r,-1) + \\\n",
    "np.power(tau,K)', data=option_train_train).fit()\n",
    "predictions = model.predict(option_test)\n",
    "predictions = pd.DataFrame(predictions, columns=['Value'])\n",
    "predictions.sample(3)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "87484e1b",
   "metadata": {},
   "source": [
    "## Final Predictions "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 417,
   "id": "db02f830",
   "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>Value</th>\n",
       "      <th>BS</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1038</th>\n",
       "      <td>3.814722</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1019</th>\n",
       "      <td>24.992202</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>147</th>\n",
       "      <td>5.614747</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          Value  BS\n",
       "1038   3.814722   1\n",
       "1019  24.992202   1\n",
       "147    5.614747   0"
      ]
     },
     "execution_count": 417,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "predictions['BS'] = y_pred\n",
    "predictions.sample(3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 418,
   "id": "d8084430",
   "metadata": {},
   "outputs": [],
   "source": [
    "predictions.to_csv(\"group_10_prediction.csv\", index=True)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.8"
  },
  "latex_envs": {
   "LaTeX_envs_menu_present": true,
   "autoclose": true,
   "autocomplete": true,
   "bibliofile": "biblio.bib",
   "cite_by": "apalike",
   "current_citInitial": 1,
   "eqLabelWithNumbers": true,
   "eqNumInitial": 1,
   "hotkeys": {
    "equation": "Ctrl-E",
    "itemize": "Ctrl-I"
   },
   "labels_anchors": false,
   "latex_user_defs": false,
   "report_style_numbering": false,
   "user_envs_cfg": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}