part1.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import math\n",
    "import warnings\n",
    "from tqdm import tqdm\n",
    "from sklearn.preprocessing import OneHotEncoder\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.model_selection import train_test_split, cross_val_score, GridSearchCV, KFold\n",
    "from sklearn.metrics import accuracy_score, precision_score, recall_score, f1_score, roc_auc_score, roc_curve\n",
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "from sklearn.tree import DecisionTreeClassifier\n",
    "from sklearn.metrics import ConfusionMatrixDisplay"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Turn off while experimenting\n",
    "warnings.filterwarnings('ignore')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Ensure reproducibility\n",
    "REP = 1\n",
    "np.random.seed(REP)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Group 7\n",
    "### 012205927 - Ana Terović\n",
    "### 012206591 - Fani Sentinella-Jerbić\n",
    "### 012228451 - Jasper De Landsheere"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Import data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "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>a0</th>\n",
       "      <th>a1</th>\n",
       "      <th>a2</th>\n",
       "      <th>a3</th>\n",
       "      <th>a4</th>\n",
       "      <th>a5</th>\n",
       "      <th>a6</th>\n",
       "      <th>a7</th>\n",
       "      <th>a8</th>\n",
       "      <th>a9</th>\n",
       "      <th>...</th>\n",
       "      <th>a36</th>\n",
       "      <th>a37</th>\n",
       "      <th>a38</th>\n",
       "      <th>a39</th>\n",
       "      <th>a40</th>\n",
       "      <th>a41</th>\n",
       "      <th>a42</th>\n",
       "      <th>a43</th>\n",
       "      <th>a44</th>\n",
       "      <th>a45c</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>6</td>\n",
       "      <td>v1493</td>\n",
       "      <td>t</td>\n",
       "      <td>139.2</td>\n",
       "      <td>q</td>\n",
       "      <td>33.2</td>\n",
       "      <td>d</td>\n",
       "      <td>...</td>\n",
       "      <td>134.5</td>\n",
       "      <td>t</td>\n",
       "      <td>24.2</td>\n",
       "      <td>q</td>\n",
       "      <td>14.4</td>\n",
       "      <td>t</td>\n",
       "      <td>42.2</td>\n",
       "      <td>t</td>\n",
       "      <td>18.6</td>\n",
       "      <td>52c</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>6</td>\n",
       "      <td>v1255</td>\n",
       "      <td>d</td>\n",
       "      <td>144.0</td>\n",
       "      <td>q</td>\n",
       "      <td>66.7</td>\n",
       "      <td>d</td>\n",
       "      <td>...</td>\n",
       "      <td>22.4</td>\n",
       "      <td>s</td>\n",
       "      <td>37.7</td>\n",
       "      <td>q</td>\n",
       "      <td>19.7</td>\n",
       "      <td>q</td>\n",
       "      <td>16.5</td>\n",
       "      <td>t</td>\n",
       "      <td>26.6</td>\n",
       "      <td>54c</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>6</td>\n",
       "      <td>v2194</td>\n",
       "      <td>d</td>\n",
       "      <td>30.7</td>\n",
       "      <td>t</td>\n",
       "      <td>26.8</td>\n",
       "      <td>t</td>\n",
       "      <td>...</td>\n",
       "      <td>16.7</td>\n",
       "      <td>t</td>\n",
       "      <td>35.4</td>\n",
       "      <td>q</td>\n",
       "      <td>65.8</td>\n",
       "      <td>s</td>\n",
       "      <td>42.4</td>\n",
       "      <td>d</td>\n",
       "      <td>139.8</td>\n",
       "      <td>54c</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>4</td>\n",
       "      <td>9</td>\n",
       "      <td>4</td>\n",
       "      <td>v2021</td>\n",
       "      <td>t</td>\n",
       "      <td>37.2</td>\n",
       "      <td>t</td>\n",
       "      <td>17.4</td>\n",
       "      <td>q</td>\n",
       "      <td>...</td>\n",
       "      <td>33.5</td>\n",
       "      <td>d</td>\n",
       "      <td>44.9</td>\n",
       "      <td>d</td>\n",
       "      <td>56.6</td>\n",
       "      <td>q</td>\n",
       "      <td>20.8</td>\n",
       "      <td>t</td>\n",
       "      <td>36.5</td>\n",
       "      <td>3c</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>3</td>\n",
       "      <td>4</td>\n",
       "      <td>9</td>\n",
       "      <td>4</td>\n",
       "      <td>v1051</td>\n",
       "      <td>q</td>\n",
       "      <td>12.2</td>\n",
       "      <td>d</td>\n",
       "      <td>154.6</td>\n",
       "      <td>t</td>\n",
       "      <td>...</td>\n",
       "      <td>49.5</td>\n",
       "      <td>t</td>\n",
       "      <td>38.8</td>\n",
       "      <td>t</td>\n",
       "      <td>71.4</td>\n",
       "      <td>s</td>\n",
       "      <td>36.8</td>\n",
       "      <td>s</td>\n",
       "      <td>46.2</td>\n",
       "      <td>3c</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1197</th>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "      <td>8</td>\n",
       "      <td>7</td>\n",
       "      <td>v501</td>\n",
       "      <td>q</td>\n",
       "      <td>18.8</td>\n",
       "      <td>t</td>\n",
       "      <td>32.5</td>\n",
       "      <td>d</td>\n",
       "      <td>...</td>\n",
       "      <td>172.1</td>\n",
       "      <td>q</td>\n",
       "      <td>62.3</td>\n",
       "      <td>d</td>\n",
       "      <td>36.1</td>\n",
       "      <td>t</td>\n",
       "      <td>27.0</td>\n",
       "      <td>s</td>\n",
       "      <td>38.8</td>\n",
       "      <td>54c</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1198</th>\n",
       "      <td>2</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>6</td>\n",
       "      <td>v785</td>\n",
       "      <td>q</td>\n",
       "      <td>60.6</td>\n",
       "      <td>t</td>\n",
       "      <td>29.3</td>\n",
       "      <td>t</td>\n",
       "      <td>...</td>\n",
       "      <td>59.1</td>\n",
       "      <td>s</td>\n",
       "      <td>39.7</td>\n",
       "      <td>d</td>\n",
       "      <td>142.6</td>\n",
       "      <td>q</td>\n",
       "      <td>18.8</td>\n",
       "      <td>d</td>\n",
       "      <td>36.2</td>\n",
       "      <td>54c</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1199</th>\n",
       "      <td>2</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>6</td>\n",
       "      <td>v2174</td>\n",
       "      <td>q</td>\n",
       "      <td>166.1</td>\n",
       "      <td>q</td>\n",
       "      <td>62.5</td>\n",
       "      <td>s</td>\n",
       "      <td>...</td>\n",
       "      <td>44.5</td>\n",
       "      <td>d</td>\n",
       "      <td>55.6</td>\n",
       "      <td>d</td>\n",
       "      <td>147.3</td>\n",
       "      <td>q</td>\n",
       "      <td>107.0</td>\n",
       "      <td>t</td>\n",
       "      <td>47.1</td>\n",
       "      <td>52c</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1200</th>\n",
       "      <td>2</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>6</td>\n",
       "      <td>v1332</td>\n",
       "      <td>t</td>\n",
       "      <td>36.5</td>\n",
       "      <td>d</td>\n",
       "      <td>46.4</td>\n",
       "      <td>t</td>\n",
       "      <td>...</td>\n",
       "      <td>17.3</td>\n",
       "      <td>t</td>\n",
       "      <td>120.4</td>\n",
       "      <td>q</td>\n",
       "      <td>17.8</td>\n",
       "      <td>d</td>\n",
       "      <td>144.4</td>\n",
       "      <td>t</td>\n",
       "      <td>35.9</td>\n",
       "      <td>54c</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1201</th>\n",
       "      <td>4</td>\n",
       "      <td>5</td>\n",
       "      <td>7</td>\n",
       "      <td>4</td>\n",
       "      <td>v1522</td>\n",
       "      <td>s</td>\n",
       "      <td>61.2</td>\n",
       "      <td>d</td>\n",
       "      <td>62.9</td>\n",
       "      <td>s</td>\n",
       "      <td>...</td>\n",
       "      <td>38.7</td>\n",
       "      <td>q</td>\n",
       "      <td>64.7</td>\n",
       "      <td>t</td>\n",
       "      <td>29.8</td>\n",
       "      <td>q</td>\n",
       "      <td>21.9</td>\n",
       "      <td>d</td>\n",
       "      <td>30.0</td>\n",
       "      <td>3c</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1202 rows × 46 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      a0  a1  a2  a3     a4 a5     a6 a7     a8 a9  ...    a36 a37    a38 a39  \\\n",
       "0      2   4   8   6  v1493  t  139.2  q   33.2  d  ...  134.5   t   24.2   q   \n",
       "1      2   4   8   6  v1255  d  144.0  q   66.7  d  ...   22.4   s   37.7   q   \n",
       "2      2   4   8   6  v2194  d   30.7  t   26.8  t  ...   16.7   t   35.4   q   \n",
       "3      3   4   9   4  v2021  t   37.2  t   17.4  q  ...   33.5   d   44.9   d   \n",
       "4      3   4   9   4  v1051  q   12.2  d  154.6  t  ...   49.5   t   38.8   t   \n",
       "...   ..  ..  ..  ..    ... ..    ... ..    ... ..  ...    ...  ..    ...  ..   \n",
       "1197   2   3   8   7   v501  q   18.8  t   32.5  d  ...  172.1   q   62.3   d   \n",
       "1198   2   4   8   6   v785  q   60.6  t   29.3  t  ...   59.1   s   39.7   d   \n",
       "1199   2   4   8   6  v2174  q  166.1  q   62.5  s  ...   44.5   d   55.6   d   \n",
       "1200   2   4   8   6  v1332  t   36.5  d   46.4  t  ...   17.3   t  120.4   q   \n",
       "1201   4   5   7   4  v1522  s   61.2  d   62.9  s  ...   38.7   q   64.7   t   \n",
       "\n",
       "        a40 a41    a42 a43    a44 a45c  \n",
       "0      14.4   t   42.2   t   18.6  52c  \n",
       "1      19.7   q   16.5   t   26.6  54c  \n",
       "2      65.8   s   42.4   d  139.8  54c  \n",
       "3      56.6   q   20.8   t   36.5   3c  \n",
       "4      71.4   s   36.8   s   46.2   3c  \n",
       "...     ...  ..    ...  ..    ...  ...  \n",
       "1197   36.1   t   27.0   s   38.8  54c  \n",
       "1198  142.6   q   18.8   d   36.2  54c  \n",
       "1199  147.3   q  107.0   t   47.1  52c  \n",
       "1200   17.8   d  144.4   t   35.9  54c  \n",
       "1201   29.8   q   21.9   d   30.0   3c  \n",
       "\n",
       "[1202 rows x 46 columns]"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = pd.read_csv(\"diterpene_shuf.csv\")\n",
    "df"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Data exploration and preprocessing"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Short explaination of the dataset\n",
    "\n",
    "*Each diterpene is described with NMR spectroscopy. In this dataset each spectrum is described by the frequency and multiplicity of all peaks. The skeleton of every diterpene contains 20 carbon atoms. So, for every carbon atom we have its frequency and multiplicity. For multiplicity we have s(singulet), d(doublet), t(triplet), q(quartet). If an atom of carbon is connected to no hydrogen atoms its singulet, if its connected to one hydrogen its doublet, two hydrogens triplet and three hydrogens connected to one carbon atom is quartet.*\n",
    "\n",
    "- Columns a0 to a3 represent how many times s, d, t, q values have accured.\n",
    "- Column a4 is sample id.\n",
    "- Columns a5 to a44 tells us the frequency and multiplicity of each carbon atom.\n",
    "- Column a45c represents the classification class.\n",
    "\n",
    "Example, for sample 0:\n",
    "In a5 to a44, 's' value has occured 2 times, 'd' 4 times, 't' 8 times and 'q' 6 times. 'v1493' is its ID value. Then for the next 40 columns we have for each of the 20 carbon atoms its multiplicity (s,d,t,q) and frequency. Finally, in a45c we can see that sample 0 belongs to class 52c. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "--------------------\n",
    "### Rename according exercise"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "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>ID</th>\n",
       "      <th>a5</th>\n",
       "      <th>a6</th>\n",
       "      <th>a7</th>\n",
       "      <th>a8</th>\n",
       "      <th>a9</th>\n",
       "      <th>a10</th>\n",
       "      <th>a11</th>\n",
       "      <th>a12</th>\n",
       "      <th>a13</th>\n",
       "      <th>...</th>\n",
       "      <th>a36</th>\n",
       "      <th>a37</th>\n",
       "      <th>a38</th>\n",
       "      <th>a39</th>\n",
       "      <th>a40</th>\n",
       "      <th>a41</th>\n",
       "      <th>a42</th>\n",
       "      <th>a43</th>\n",
       "      <th>a44</th>\n",
       "      <th>Class</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>v1493</td>\n",
       "      <td>t</td>\n",
       "      <td>139.2</td>\n",
       "      <td>q</td>\n",
       "      <td>33.2</td>\n",
       "      <td>d</td>\n",
       "      <td>30.7</td>\n",
       "      <td>d</td>\n",
       "      <td>49.6</td>\n",
       "      <td>t</td>\n",
       "      <td>...</td>\n",
       "      <td>134.5</td>\n",
       "      <td>t</td>\n",
       "      <td>24.2</td>\n",
       "      <td>q</td>\n",
       "      <td>14.4</td>\n",
       "      <td>t</td>\n",
       "      <td>42.2</td>\n",
       "      <td>t</td>\n",
       "      <td>18.6</td>\n",
       "      <td>52c</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>v1255</td>\n",
       "      <td>d</td>\n",
       "      <td>144.0</td>\n",
       "      <td>q</td>\n",
       "      <td>66.7</td>\n",
       "      <td>d</td>\n",
       "      <td>46.2</td>\n",
       "      <td>t</td>\n",
       "      <td>32.7</td>\n",
       "      <td>q</td>\n",
       "      <td>...</td>\n",
       "      <td>22.4</td>\n",
       "      <td>s</td>\n",
       "      <td>37.7</td>\n",
       "      <td>q</td>\n",
       "      <td>19.7</td>\n",
       "      <td>q</td>\n",
       "      <td>16.5</td>\n",
       "      <td>t</td>\n",
       "      <td>26.6</td>\n",
       "      <td>54c</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>v2194</td>\n",
       "      <td>d</td>\n",
       "      <td>30.7</td>\n",
       "      <td>t</td>\n",
       "      <td>26.8</td>\n",
       "      <td>t</td>\n",
       "      <td>41.2</td>\n",
       "      <td>t</td>\n",
       "      <td>139.3</td>\n",
       "      <td>q</td>\n",
       "      <td>...</td>\n",
       "      <td>16.7</td>\n",
       "      <td>t</td>\n",
       "      <td>35.4</td>\n",
       "      <td>q</td>\n",
       "      <td>65.8</td>\n",
       "      <td>s</td>\n",
       "      <td>42.4</td>\n",
       "      <td>d</td>\n",
       "      <td>139.8</td>\n",
       "      <td>54c</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>v2021</td>\n",
       "      <td>t</td>\n",
       "      <td>37.2</td>\n",
       "      <td>t</td>\n",
       "      <td>17.4</td>\n",
       "      <td>q</td>\n",
       "      <td>16.7</td>\n",
       "      <td>s</td>\n",
       "      <td>58.1</td>\n",
       "      <td>t</td>\n",
       "      <td>...</td>\n",
       "      <td>33.5</td>\n",
       "      <td>d</td>\n",
       "      <td>44.9</td>\n",
       "      <td>d</td>\n",
       "      <td>56.6</td>\n",
       "      <td>q</td>\n",
       "      <td>20.8</td>\n",
       "      <td>t</td>\n",
       "      <td>36.5</td>\n",
       "      <td>3c</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>v1051</td>\n",
       "      <td>q</td>\n",
       "      <td>12.2</td>\n",
       "      <td>d</td>\n",
       "      <td>154.6</td>\n",
       "      <td>t</td>\n",
       "      <td>18.7</td>\n",
       "      <td>t</td>\n",
       "      <td>37.1</td>\n",
       "      <td>t</td>\n",
       "      <td>...</td>\n",
       "      <td>49.5</td>\n",
       "      <td>t</td>\n",
       "      <td>38.8</td>\n",
       "      <td>t</td>\n",
       "      <td>71.4</td>\n",
       "      <td>s</td>\n",
       "      <td>36.8</td>\n",
       "      <td>s</td>\n",
       "      <td>46.2</td>\n",
       "      <td>3c</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 42 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      ID a5     a6 a7     a8 a9   a10 a11    a12 a13  ...    a36 a37   a38  \\\n",
       "0  v1493  t  139.2  q   33.2  d  30.7   d   49.6   t  ...  134.5   t  24.2   \n",
       "1  v1255  d  144.0  q   66.7  d  46.2   t   32.7   q  ...   22.4   s  37.7   \n",
       "2  v2194  d   30.7  t   26.8  t  41.2   t  139.3   q  ...   16.7   t  35.4   \n",
       "3  v2021  t   37.2  t   17.4  q  16.7   s   58.1   t  ...   33.5   d  44.9   \n",
       "4  v1051  q   12.2  d  154.6  t  18.7   t   37.1   t  ...   49.5   t  38.8   \n",
       "\n",
       "  a39   a40 a41   a42 a43    a44 Class  \n",
       "0   q  14.4   t  42.2   t   18.6   52c  \n",
       "1   q  19.7   q  16.5   t   26.6   54c  \n",
       "2   q  65.8   s  42.4   d  139.8   54c  \n",
       "3   d  56.6   q  20.8   t   36.5    3c  \n",
       "4   t  71.4   s  36.8   s   46.2    3c  \n",
       "\n",
       "[5 rows x 42 columns]"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Drop first four columns as they are not needed according to the exercise\n",
    "df = df.drop([\"a0\", \"a1\", \"a2\", \"a3\"], axis=1)\n",
    "# Rename column \"a45c\" to \"Class\", \"a4\" to \"ID\"\n",
    "df = df.rename(columns={\"a45c\": \"Class\", \"a4\": \"ID\"})\n",
    "# Inspect\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "-------------------------\n",
    "### Basic exploration"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>a6</th>\n",
       "      <th>a8</th>\n",
       "      <th>a10</th>\n",
       "      <th>a12</th>\n",
       "      <th>a14</th>\n",
       "      <th>a16</th>\n",
       "      <th>a18</th>\n",
       "      <th>a20</th>\n",
       "      <th>a22</th>\n",
       "      <th>a24</th>\n",
       "      <th>a26</th>\n",
       "      <th>a28</th>\n",
       "      <th>a30</th>\n",
       "      <th>a32</th>\n",
       "      <th>a34</th>\n",
       "      <th>a36</th>\n",
       "      <th>a38</th>\n",
       "      <th>a40</th>\n",
       "      <th>a42</th>\n",
       "      <th>a44</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>1202.000000</td>\n",
       "      <td>1202.000000</td>\n",
       "      <td>1202.000000</td>\n",
       "      <td>1202.000000</td>\n",
       "      <td>1202.000000</td>\n",
       "      <td>1202.000000</td>\n",
       "      <td>1202.000000</td>\n",
       "      <td>1202.000000</td>\n",
       "      <td>1202.000000</td>\n",
       "      <td>1202.000000</td>\n",
       "      <td>1202.000000</td>\n",
       "      <td>1202.000000</td>\n",
       "      <td>1202.000000</td>\n",
       "      <td>1202.000000</td>\n",
       "      <td>1202.000000</td>\n",
       "      <td>1202.000000</td>\n",
       "      <td>1202.000000</td>\n",
       "      <td>1202.000000</td>\n",
       "      <td>1202.000000</td>\n",
       "      <td>1202.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>56.387155</td>\n",
       "      <td>59.147687</td>\n",
       "      <td>59.999551</td>\n",
       "      <td>58.874193</td>\n",
       "      <td>59.536389</td>\n",
       "      <td>59.597471</td>\n",
       "      <td>57.309401</td>\n",
       "      <td>58.728378</td>\n",
       "      <td>55.568128</td>\n",
       "      <td>57.278544</td>\n",
       "      <td>58.794725</td>\n",
       "      <td>56.415308</td>\n",
       "      <td>61.136672</td>\n",
       "      <td>59.793536</td>\n",
       "      <td>58.655524</td>\n",
       "      <td>60.627945</td>\n",
       "      <td>58.849384</td>\n",
       "      <td>57.287329</td>\n",
       "      <td>60.231140</td>\n",
       "      <td>57.614401</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>43.868545</td>\n",
       "      <td>45.930797</td>\n",
       "      <td>46.909051</td>\n",
       "      <td>45.465840</td>\n",
       "      <td>46.522244</td>\n",
       "      <td>45.921470</td>\n",
       "      <td>43.398889</td>\n",
       "      <td>44.909447</td>\n",
       "      <td>43.476958</td>\n",
       "      <td>45.148903</td>\n",
       "      <td>46.448512</td>\n",
       "      <td>44.489723</td>\n",
       "      <td>47.351395</td>\n",
       "      <td>44.805028</td>\n",
       "      <td>45.697774</td>\n",
       "      <td>46.396705</td>\n",
       "      <td>46.498059</td>\n",
       "      <td>44.742240</td>\n",
       "      <td>47.950173</td>\n",
       "      <td>44.897128</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>7.300000</td>\n",
       "      <td>7.900000</td>\n",
       "      <td>11.600000</td>\n",
       "      <td>7.500000</td>\n",
       "      <td>7.700000</td>\n",
       "      <td>7.500000</td>\n",
       "      <td>7.600000</td>\n",
       "      <td>9.200000</td>\n",
       "      <td>7.500000</td>\n",
       "      <td>6.800000</td>\n",
       "      <td>11.800000</td>\n",
       "      <td>11.600000</td>\n",
       "      <td>8.000000</td>\n",
       "      <td>7.600000</td>\n",
       "      <td>2.400000</td>\n",
       "      <td>7.800000</td>\n",
       "      <td>7.900000</td>\n",
       "      <td>8.100000</td>\n",
       "      <td>7.900000</td>\n",
       "      <td>7.700000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>27.800000</td>\n",
       "      <td>28.200000</td>\n",
       "      <td>27.725000</td>\n",
       "      <td>28.425000</td>\n",
       "      <td>27.700000</td>\n",
       "      <td>28.725000</td>\n",
       "      <td>27.875000</td>\n",
       "      <td>28.302500</td>\n",
       "      <td>28.050000</td>\n",
       "      <td>28.100000</td>\n",
       "      <td>26.625000</td>\n",
       "      <td>27.025000</td>\n",
       "      <td>28.550000</td>\n",
       "      <td>30.300000</td>\n",
       "      <td>28.700000</td>\n",
       "      <td>30.625000</td>\n",
       "      <td>28.300000</td>\n",
       "      <td>27.302500</td>\n",
       "      <td>28.225000</td>\n",
       "      <td>28.225000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>39.600000</td>\n",
       "      <td>41.400000</td>\n",
       "      <td>41.200000</td>\n",
       "      <td>40.400000</td>\n",
       "      <td>41.300000</td>\n",
       "      <td>41.000000</td>\n",
       "      <td>39.925000</td>\n",
       "      <td>40.500000</td>\n",
       "      <td>39.600000</td>\n",
       "      <td>40.400000</td>\n",
       "      <td>40.215000</td>\n",
       "      <td>39.500000</td>\n",
       "      <td>41.400000</td>\n",
       "      <td>41.200000</td>\n",
       "      <td>40.700000</td>\n",
       "      <td>42.000000</td>\n",
       "      <td>40.050000</td>\n",
       "      <td>39.700000</td>\n",
       "      <td>40.700000</td>\n",
       "      <td>40.400000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>66.700000</td>\n",
       "      <td>71.800000</td>\n",
       "      <td>73.187500</td>\n",
       "      <td>72.700000</td>\n",
       "      <td>72.975000</td>\n",
       "      <td>72.175000</td>\n",
       "      <td>71.975000</td>\n",
       "      <td>72.100000</td>\n",
       "      <td>64.160000</td>\n",
       "      <td>65.775000</td>\n",
       "      <td>71.900000</td>\n",
       "      <td>65.375000</td>\n",
       "      <td>74.600000</td>\n",
       "      <td>72.600000</td>\n",
       "      <td>69.900000</td>\n",
       "      <td>72.575000</td>\n",
       "      <td>71.175000</td>\n",
       "      <td>68.825000</td>\n",
       "      <td>72.075000</td>\n",
       "      <td>70.200000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>216.630000</td>\n",
       "      <td>219.100000</td>\n",
       "      <td>220.700000</td>\n",
       "      <td>216.900000</td>\n",
       "      <td>224.700000</td>\n",
       "      <td>221.250000</td>\n",
       "      <td>223.000000</td>\n",
       "      <td>222.700000</td>\n",
       "      <td>222.100000</td>\n",
       "      <td>218.300000</td>\n",
       "      <td>222.500000</td>\n",
       "      <td>215.270000</td>\n",
       "      <td>220.200000</td>\n",
       "      <td>216.900000</td>\n",
       "      <td>218.000000</td>\n",
       "      <td>216.400000</td>\n",
       "      <td>224.500000</td>\n",
       "      <td>216.100000</td>\n",
       "      <td>217.460000</td>\n",
       "      <td>219.300000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                a6           a8          a10          a12          a14  \\\n",
       "count  1202.000000  1202.000000  1202.000000  1202.000000  1202.000000   \n",
       "mean     56.387155    59.147687    59.999551    58.874193    59.536389   \n",
       "std      43.868545    45.930797    46.909051    45.465840    46.522244   \n",
       "min       7.300000     7.900000    11.600000     7.500000     7.700000   \n",
       "25%      27.800000    28.200000    27.725000    28.425000    27.700000   \n",
       "50%      39.600000    41.400000    41.200000    40.400000    41.300000   \n",
       "75%      66.700000    71.800000    73.187500    72.700000    72.975000   \n",
       "max     216.630000   219.100000   220.700000   216.900000   224.700000   \n",
       "\n",
       "               a16          a18          a20          a22          a24  \\\n",
       "count  1202.000000  1202.000000  1202.000000  1202.000000  1202.000000   \n",
       "mean     59.597471    57.309401    58.728378    55.568128    57.278544   \n",
       "std      45.921470    43.398889    44.909447    43.476958    45.148903   \n",
       "min       7.500000     7.600000     9.200000     7.500000     6.800000   \n",
       "25%      28.725000    27.875000    28.302500    28.050000    28.100000   \n",
       "50%      41.000000    39.925000    40.500000    39.600000    40.400000   \n",
       "75%      72.175000    71.975000    72.100000    64.160000    65.775000   \n",
       "max     221.250000   223.000000   222.700000   222.100000   218.300000   \n",
       "\n",
       "               a26          a28          a30          a32          a34  \\\n",
       "count  1202.000000  1202.000000  1202.000000  1202.000000  1202.000000   \n",
       "mean     58.794725    56.415308    61.136672    59.793536    58.655524   \n",
       "std      46.448512    44.489723    47.351395    44.805028    45.697774   \n",
       "min      11.800000    11.600000     8.000000     7.600000     2.400000   \n",
       "25%      26.625000    27.025000    28.550000    30.300000    28.700000   \n",
       "50%      40.215000    39.500000    41.400000    41.200000    40.700000   \n",
       "75%      71.900000    65.375000    74.600000    72.600000    69.900000   \n",
       "max     222.500000   215.270000   220.200000   216.900000   218.000000   \n",
       "\n",
       "               a36          a38          a40          a42          a44  \n",
       "count  1202.000000  1202.000000  1202.000000  1202.000000  1202.000000  \n",
       "mean     60.627945    58.849384    57.287329    60.231140    57.614401  \n",
       "std      46.396705    46.498059    44.742240    47.950173    44.897128  \n",
       "min       7.800000     7.900000     8.100000     7.900000     7.700000  \n",
       "25%      30.625000    28.300000    27.302500    28.225000    28.225000  \n",
       "50%      42.000000    40.050000    39.700000    40.700000    40.400000  \n",
       "75%      72.575000    71.175000    68.825000    72.075000    70.200000  \n",
       "max     216.400000   224.500000   216.100000   217.460000   219.300000  "
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Basic information about the data\n",
    "df.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "ID       1202\n",
       "a5          4\n",
       "a6        711\n",
       "a7          4\n",
       "a8        728\n",
       "a9          4\n",
       "a10       724\n",
       "a11         4\n",
       "a12       719\n",
       "a13         4\n",
       "a14       725\n",
       "a15         4\n",
       "a16       725\n",
       "a17         4\n",
       "a18       712\n",
       "a19         4\n",
       "a20       721\n",
       "a21         4\n",
       "a22       702\n",
       "a23         4\n",
       "a24       686\n",
       "a25         4\n",
       "a26       726\n",
       "a27         4\n",
       "a28       687\n",
       "a29         4\n",
       "a30       741\n",
       "a31         4\n",
       "a32       729\n",
       "a33         4\n",
       "a34       711\n",
       "a35         4\n",
       "a36       720\n",
       "a37         4\n",
       "a38       720\n",
       "a39         4\n",
       "a40       701\n",
       "a41         4\n",
       "a42       726\n",
       "a43         4\n",
       "a44       710\n",
       "Class      22\n",
       "dtype: int64"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Number of unique values in each column\n",
    "df.nunique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Check for missing values\n",
    "df.isnull().sum().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1202, 42)"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Dimensions of the data\n",
    "df.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "-----------\n",
    "### Data visualization"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Distribution of values in the \"Class\" column\n",
    "ax = sns.countplot(x=\"Class\", data=df)\n",
    "ax.bar_label(ax.containers[0])\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Here, we noticed that the classes are very imbalanced. For now, we keep this in mind. Further preprocessing steps to deal with this might be necessary."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[<AxesSubplot:title={'center':'a6'}>,\n",
       "        <AxesSubplot:title={'center':'a8'}>,\n",
       "        <AxesSubplot:title={'center':'a10'}>,\n",
       "        <AxesSubplot:title={'center':'a12'}>],\n",
       "       [<AxesSubplot:title={'center':'a14'}>,\n",
       "        <AxesSubplot:title={'center':'a16'}>,\n",
       "        <AxesSubplot:title={'center':'a18'}>,\n",
       "        <AxesSubplot:title={'center':'a20'}>],\n",
       "       [<AxesSubplot:title={'center':'a22'}>,\n",
       "        <AxesSubplot:title={'center':'a24'}>,\n",
       "        <AxesSubplot:title={'center':'a26'}>,\n",
       "        <AxesSubplot:title={'center':'a28'}>],\n",
       "       [<AxesSubplot:title={'center':'a30'}>,\n",
       "        <AxesSubplot:title={'center':'a32'}>,\n",
       "        <AxesSubplot:title={'center':'a34'}>,\n",
       "        <AxesSubplot:title={'center':'a36'}>],\n",
       "       [<AxesSubplot:title={'center':'a38'}>,\n",
       "        <AxesSubplot:title={'center':'a40'}>,\n",
       "        <AxesSubplot:title={'center':'a42'}>,\n",
       "        <AxesSubplot:title={'center':'a44'}>]], dtype=object)"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x1440 with 20 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot distribution of each column\n",
    "df.hist(figsize=(20,20))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We notice that all frequencies are very similarly distributed."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "------------------------\n",
    "### Preprocessing"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As decribed in the official paper, some preprocessing can help get better results. For this purpose we will create another dataset df_reduced which will have the reduced multiplicities."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Create a copy\n",
    "df_reduced = df.copy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|█████████████████████████████████████████████████████████████████████████████| 1202/1202 [00:04<00:00, 265.80it/s]\n"
     ]
    }
   ],
   "source": [
    "# For each sample of the dataset\n",
    "col_names = df_reduced.columns\n",
    "for i in tqdm(range(len(df_reduced))):\n",
    "    \n",
    "    for j in range (len(col_names)):\n",
    "\n",
    "        if df_reduced[col_names[j]][i] == 's' and df_reduced[col_names[j+1]][i] >= 64.5 and df_reduced[col_names[j+1]][i] <= 95:\n",
    "            df_reduced[col_names[j]][i] = 'd'\n",
    "\n",
    "        if df_reduced[col_names[j]][i] == 's' and df_reduced[col_names[j+1]][i] >= 96 and df_reduced[col_names[j+1]][i] <= 114:\n",
    "            df_reduced[col_names[j]][i] = 't'\n",
    "\n",
    "        if df_reduced[col_names[j]][i] == 's' and df_reduced[col_names[j+1]][i] >= 115 and df_reduced[col_names[j+1]][i] <= 165:\n",
    "            df_reduced[col_names[j]][i] = 'd'\n",
    "\n",
    "        if df_reduced[col_names[j]][i] == 's' and df_reduced[col_names[j+1]][i] >= 165 and df_reduced[col_names[j+1]][i] <= 188:\n",
    "            df_reduced[col_names[j]][i] = 'q'\n",
    "\n",
    "        if df_reduced[col_names[j]][i] == 's' and df_reduced[col_names[j+1]][i] >= 188 and df_reduced[col_names[j+1]][i] <= math.inf:\n",
    "            df_reduced[col_names[j]][i] = 't'\n",
    "\n",
    "\n",
    "        if df_reduced[col_names[j]][i] == 'd' and df_reduced[col_names[j+1]][i] >= 64.5 and df_reduced[col_names[j+1]][i] <= 95:\n",
    "            df_reduced[col_names[j]][i] = 't'\n",
    "\n",
    "        if df_reduced[col_names[j]][i] == 'd' and df_reduced[col_names[j+1]][i] >= 105 and df_reduced[col_names[j+1]][i] <= 180:\n",
    "            df_reduced[col_names[j]][i] = 't'\n",
    "\n",
    "        if df_reduced[col_names[j]][i] == 'd' and df_reduced[col_names[j+1]][i] >= 96 and df_reduced[col_names[j+1]][i] <= 104:\n",
    "            df_reduced[col_names[j]][i] = 'q'\n",
    "\n",
    "        if df_reduced[col_names[j]][i] == 'd' and df_reduced[col_names[j+1]][i] >= 180 and df_reduced[col_names[j+1]][i] <= math.inf:\n",
    "            df_reduced[col_names[j]][i] = 'q'\n",
    "\n",
    "\n",
    "        if df_reduced[col_names[j]][i] == 't' and df_reduced[col_names[j+1]][i] >= 59 and df_reduced[col_names[j+1]][i] <= 90:\n",
    "            df_reduced[col_names[j]][i] = 'q'\n",
    "\n",
    "        if df_reduced[col_names[j]][i] == 't' and df_reduced[col_names[j+1]][i] >= 90 and df_reduced[col_names[j+1]][i] <= math.inf:\n",
    "            df_reduced[col_names[j]][i] = 'q'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "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>ID</th>\n",
       "      <th>a5</th>\n",
       "      <th>a6</th>\n",
       "      <th>a7</th>\n",
       "      <th>a8</th>\n",
       "      <th>a9</th>\n",
       "      <th>a10</th>\n",
       "      <th>a11</th>\n",
       "      <th>a12</th>\n",
       "      <th>a13</th>\n",
       "      <th>...</th>\n",
       "      <th>a36</th>\n",
       "      <th>a37</th>\n",
       "      <th>a38</th>\n",
       "      <th>a39</th>\n",
       "      <th>a40</th>\n",
       "      <th>a41</th>\n",
       "      <th>a42</th>\n",
       "      <th>a43</th>\n",
       "      <th>a44</th>\n",
       "      <th>Class</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>v1493</td>\n",
       "      <td>q</td>\n",
       "      <td>139.2</td>\n",
       "      <td>q</td>\n",
       "      <td>33.2</td>\n",
       "      <td>d</td>\n",
       "      <td>30.7</td>\n",
       "      <td>d</td>\n",
       "      <td>49.6</td>\n",
       "      <td>t</td>\n",
       "      <td>...</td>\n",
       "      <td>134.5</td>\n",
       "      <td>t</td>\n",
       "      <td>24.2</td>\n",
       "      <td>q</td>\n",
       "      <td>14.4</td>\n",
       "      <td>t</td>\n",
       "      <td>42.2</td>\n",
       "      <td>t</td>\n",
       "      <td>18.6</td>\n",
       "      <td>52c</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>v1255</td>\n",
       "      <td>q</td>\n",
       "      <td>144.0</td>\n",
       "      <td>q</td>\n",
       "      <td>66.7</td>\n",
       "      <td>d</td>\n",
       "      <td>46.2</td>\n",
       "      <td>t</td>\n",
       "      <td>32.7</td>\n",
       "      <td>q</td>\n",
       "      <td>...</td>\n",
       "      <td>22.4</td>\n",
       "      <td>s</td>\n",
       "      <td>37.7</td>\n",
       "      <td>q</td>\n",
       "      <td>19.7</td>\n",
       "      <td>q</td>\n",
       "      <td>16.5</td>\n",
       "      <td>t</td>\n",
       "      <td>26.6</td>\n",
       "      <td>54c</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>v2194</td>\n",
       "      <td>d</td>\n",
       "      <td>30.7</td>\n",
       "      <td>t</td>\n",
       "      <td>26.8</td>\n",
       "      <td>t</td>\n",
       "      <td>41.2</td>\n",
       "      <td>q</td>\n",
       "      <td>139.3</td>\n",
       "      <td>q</td>\n",
       "      <td>...</td>\n",
       "      <td>16.7</td>\n",
       "      <td>t</td>\n",
       "      <td>35.4</td>\n",
       "      <td>q</td>\n",
       "      <td>65.8</td>\n",
       "      <td>s</td>\n",
       "      <td>42.4</td>\n",
       "      <td>q</td>\n",
       "      <td>139.8</td>\n",
       "      <td>54c</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>v2021</td>\n",
       "      <td>t</td>\n",
       "      <td>37.2</td>\n",
       "      <td>t</td>\n",
       "      <td>17.4</td>\n",
       "      <td>q</td>\n",
       "      <td>16.7</td>\n",
       "      <td>s</td>\n",
       "      <td>58.1</td>\n",
       "      <td>t</td>\n",
       "      <td>...</td>\n",
       "      <td>33.5</td>\n",
       "      <td>d</td>\n",
       "      <td>44.9</td>\n",
       "      <td>d</td>\n",
       "      <td>56.6</td>\n",
       "      <td>q</td>\n",
       "      <td>20.8</td>\n",
       "      <td>t</td>\n",
       "      <td>36.5</td>\n",
       "      <td>3c</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>v1051</td>\n",
       "      <td>q</td>\n",
       "      <td>12.2</td>\n",
       "      <td>q</td>\n",
       "      <td>154.6</td>\n",
       "      <td>t</td>\n",
       "      <td>18.7</td>\n",
       "      <td>t</td>\n",
       "      <td>37.1</td>\n",
       "      <td>q</td>\n",
       "      <td>...</td>\n",
       "      <td>49.5</td>\n",
       "      <td>t</td>\n",
       "      <td>38.8</td>\n",
       "      <td>q</td>\n",
       "      <td>71.4</td>\n",
       "      <td>s</td>\n",
       "      <td>36.8</td>\n",
       "      <td>s</td>\n",
       "      <td>46.2</td>\n",
       "      <td>3c</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 42 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      ID a5     a6 a7     a8 a9   a10 a11    a12 a13  ...    a36 a37   a38  \\\n",
       "0  v1493  q  139.2  q   33.2  d  30.7   d   49.6   t  ...  134.5   t  24.2   \n",
       "1  v1255  q  144.0  q   66.7  d  46.2   t   32.7   q  ...   22.4   s  37.7   \n",
       "2  v2194  d   30.7  t   26.8  t  41.2   q  139.3   q  ...   16.7   t  35.4   \n",
       "3  v2021  t   37.2  t   17.4  q  16.7   s   58.1   t  ...   33.5   d  44.9   \n",
       "4  v1051  q   12.2  q  154.6  t  18.7   t   37.1   q  ...   49.5   t  38.8   \n",
       "\n",
       "  a39   a40 a41   a42 a43    a44 Class  \n",
       "0   q  14.4   t  42.2   t   18.6   52c  \n",
       "1   q  19.7   q  16.5   t   26.6   54c  \n",
       "2   q  65.8   s  42.4   q  139.8   54c  \n",
       "3   d  56.6   q  20.8   t   36.5    3c  \n",
       "4   q  71.4   s  36.8   s   46.2    3c  \n",
       "\n",
       "[5 rows x 42 columns]"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_reduced.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "To deal with the categorical features, we opt for one hot encoding over label encoding, as we are not sure yet how the multiplicities work."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "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>a5_s</th>\n",
       "      <th>a5_d</th>\n",
       "      <th>a5_t</th>\n",
       "      <th>a5_q</th>\n",
       "      <th>a6</th>\n",
       "      <th>a7_s</th>\n",
       "      <th>a7_d</th>\n",
       "      <th>a7_t</th>\n",
       "      <th>a7_q</th>\n",
       "      <th>a8</th>\n",
       "      <th>...</th>\n",
       "      <th>a41_s</th>\n",
       "      <th>a41_d</th>\n",
       "      <th>a41_t</th>\n",
       "      <th>a41_q</th>\n",
       "      <th>a42</th>\n",
       "      <th>a43_s</th>\n",
       "      <th>a43_d</th>\n",
       "      <th>a43_t</th>\n",
       "      <th>a43_q</th>\n",
       "      <th>a44</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>139.2</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>33.2</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>42.2</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>18.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>144.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>66.7</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>16.5</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>26.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>30.7</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>26.8</td>\n",
       "      <td>...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>42.4</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>139.8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>37.2</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>17.4</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>20.8</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>36.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>12.2</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>154.6</td>\n",
       "      <td>...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>36.8</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>46.2</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 100 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   a5_s  a5_d  a5_t  a5_q     a6  a7_s  a7_d  a7_t  a7_q     a8  ...  a41_s  \\\n",
       "0   0.0   0.0   0.0   1.0  139.2   0.0   0.0   0.0   1.0   33.2  ...    0.0   \n",
       "1   0.0   0.0   0.0   1.0  144.0   0.0   0.0   0.0   1.0   66.7  ...    0.0   \n",
       "2   0.0   1.0   0.0   0.0   30.7   0.0   0.0   1.0   0.0   26.8  ...    1.0   \n",
       "3   0.0   0.0   1.0   0.0   37.2   0.0   0.0   1.0   0.0   17.4  ...    0.0   \n",
       "4   0.0   0.0   0.0   1.0   12.2   0.0   0.0   0.0   1.0  154.6  ...    1.0   \n",
       "\n",
       "   a41_d  a41_t  a41_q   a42  a43_s  a43_d  a43_t  a43_q    a44  \n",
       "0    0.0    1.0    0.0  42.2    0.0    0.0    1.0    0.0   18.6  \n",
       "1    0.0    0.0    1.0  16.5    0.0    0.0    1.0    0.0   26.6  \n",
       "2    0.0    0.0    0.0  42.4    0.0    0.0    0.0    1.0  139.8  \n",
       "3    0.0    0.0    1.0  20.8    0.0    0.0    1.0    0.0   36.5  \n",
       "4    0.0    0.0    0.0  36.8    1.0    0.0    0.0    0.0   46.2  \n",
       "\n",
       "[5 rows x 100 columns]"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Categorical columns\n",
    "categorical_columns = [f\"a{i}\" for i in range(5, 44, 2)]\n",
    "# Final names\n",
    "final_names = []\n",
    "for i in df_reduced.drop([\"ID\", \"Class\"], axis=1).select_dtypes(include=['object']).columns:\n",
    "    final_names.append(f\"{i}_s\")\n",
    "    final_names.append(f\"{i}_d\")\n",
    "    final_names.append(f\"{i}_t\")\n",
    "    final_names.append(f\"{i}_q\")\n",
    "    final_names.append(f\"a{int(i[1:]) + 1}\")\n",
    "# Drop the \"ID\" column\n",
    "df_reduced = df_reduced.drop([\"ID\"], axis=1)\n",
    "# Split the data into features and labels\n",
    "y = df_reduced[\"Class\"]\n",
    "X = df_reduced.drop([\"Class\"], axis=1)\n",
    "# One hot encoder\n",
    "enc = OneHotEncoder(sparse=False)\n",
    "# Fit the encoder\n",
    "enc.fit(X[categorical_columns])\n",
    "# Transform the data\n",
    "X_enc = enc.transform(X[categorical_columns]) #.toarray()\n",
    "# Create a dataframe from the encoded data\n",
    "X_enc_df = pd.DataFrame(X_enc)\n",
    "# Rename\n",
    "X_enc_df.columns = enc.get_feature_names_out(categorical_columns)\n",
    "# Concatenate the encoded data with the rest of the data\n",
    "X = pd.concat([X, X_enc_df], axis=1)\n",
    "# Drop the categorical columns\n",
    "X = X.drop(categorical_columns, axis=1)\n",
    "# Rename\n",
    "X = X[final_names]\n",
    "# Inspect\n",
    "X.head()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As it is unsure of what the columns represent, scaling is not done so there is no information loss."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "-----------------------\n",
    "# Classification\n",
    "\n",
    "We preliminarily try 3 different algorithms to gain the initial intuition of our problem. \n",
    "\n",
    "We define a function for faster evaluation of different models. At this stage we are satisfied with hold out for evaluation."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "def evaluate(clfs, X, y, plot=True):\n",
    "    results = []\n",
    "    \n",
    "    for clf in clfs:\n",
    "        # Split the data into training and testing sets\n",
    "        X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2)\n",
    "    \n",
    "        # Fit the model\n",
    "        clf.fit(X_train, y_train)\n",
    "    \n",
    "        # Make predictions\n",
    "        y_pred = clf.predict(X_test)\n",
    "    \n",
    "        # Get name of the classifier\n",
    "        name = type(clf).__name__\n",
    "        \n",
    "        # Calculate different scores\n",
    "        res = {\n",
    "            'Algorithm': name,\n",
    "            'Accuracy': accuracy_score(y_test, y_pred),\n",
    "            'F1 macro': f1_score(y_test, y_pred, average='macro'),\n",
    "            'F1 micro': f1_score(y_test, y_pred, average='micro')\n",
    "        }\n",
    "        \n",
    "        if plot:\n",
    "            ConfusionMatrixDisplay.from_estimator(clf, X_test, y_test,  cmap='rocket_r')\n",
    "        \n",
    "        results.append(res)\n",
    "        \n",
    "    return pd.DataFrame(results)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "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>Algorithm</th>\n",
       "      <th>Accuracy</th>\n",
       "      <th>F1 macro</th>\n",
       "      <th>F1 micro</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>LogisticRegression</td>\n",
       "      <td>0.253112</td>\n",
       "      <td>0.045349</td>\n",
       "      <td>0.253112</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>KNeighborsClassifier</td>\n",
       "      <td>0.290456</td>\n",
       "      <td>0.076810</td>\n",
       "      <td>0.290456</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>DecisionTreeClassifier</td>\n",
       "      <td>0.307054</td>\n",
       "      <td>0.073772</td>\n",
       "      <td>0.307054</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                Algorithm  Accuracy  F1 macro  F1 micro\n",
       "0      LogisticRegression  0.253112  0.045349  0.253112\n",
       "1    KNeighborsClassifier  0.290456  0.076810  0.290456\n",
       "2  DecisionTreeClassifier  0.307054  0.073772  0.307054"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clfs_to_try = [LogisticRegression(), \n",
    "               KNeighborsClassifier(), \n",
    "               DecisionTreeClassifier()]\n",
    "evaluate(clfs_to_try, X, y, plot=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Our premlinirary evaluation doesn't show a great success. We proceed trying to adapt hyperparameters to better fit our problem."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Grid Search"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### K-Nearest Neighbors"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "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>Algorithm</th>\n",
       "      <th>Accuracy</th>\n",
       "      <th>F1 macro</th>\n",
       "      <th>F1 micro</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>KNeighborsClassifier</td>\n",
       "      <td>0.344398</td>\n",
       "      <td>0.092695</td>\n",
       "      <td>0.344398</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              Algorithm  Accuracy  F1 macro  F1 micro\n",
       "0  KNeighborsClassifier  0.344398  0.092695  0.344398"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Do a grid search\n",
    "clf = KNeighborsClassifier()\n",
    "grid_values = {'n_neighbors':[6,8,10,12,14,16,18,20],\n",
    "               'leaf_size':list(range(1,50,5)),\n",
    "               \"metric\":[\"euclidean\", \"manhattan\", \"chebyshev\"],\n",
    "               'weights':['uniform', 'distance'],\n",
    "               'algorithm':['auto', 'ball_tree', 'kd_tree', 'brute']\n",
    "              }\n",
    "grid_clf_acc = GridSearchCV(clf, param_grid = grid_values,scoring = 'accuracy', n_jobs=-1)\n",
    "grid_clf_acc.fit(X_train, y_train)\n",
    "evaluate([KNeighborsClassifier(**grid_clf_acc.best_params_)], X, y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Decision Tree"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "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>Algorithm</th>\n",
       "      <th>Accuracy</th>\n",
       "      <th>F1 macro</th>\n",
       "      <th>F1 micro</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>DecisionTreeClassifier</td>\n",
       "      <td>0.344398</td>\n",
       "      <td>0.073071</td>\n",
       "      <td>0.344398</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                Algorithm  Accuracy  F1 macro  F1 micro\n",
       "0  DecisionTreeClassifier  0.344398  0.073071  0.344398"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Do a grid search\n",
    "clf = DecisionTreeClassifier()\n",
    "grid_values = {'ccp_alpha': [0.1, 0.01, 0.001, 0.0001, 0],\n",
    "               'criterion': ['gini', 'entropy'],\n",
    "               'max_depth': [5, 6, 7, 8, 9, 10, None],\n",
    "               'max_features': ['auto', 'sqrt', 'log2']\n",
    "              }\n",
    "grid_clf_acc = GridSearchCV(clf, param_grid = grid_values,scoring = 'accuracy', n_jobs=-1)\n",
    "grid_clf_acc.fit(X_train, y_train)\n",
    "evaluate([DecisionTreeClassifier(**grid_clf_acc.best_params_)], X, y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Logistic Regression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "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>Algorithm</th>\n",
       "      <th>Accuracy</th>\n",
       "      <th>F1 macro</th>\n",
       "      <th>F1 micro</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>LogisticRegression</td>\n",
       "      <td>0.522822</td>\n",
       "      <td>0.164155</td>\n",
       "      <td>0.522822</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            Algorithm  Accuracy  F1 macro  F1 micro\n",
       "0  LogisticRegression  0.522822  0.164155  0.522822"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Do a grid search\n",
    "clf = LogisticRegression()\n",
    "grid_values = {'penalty': ['l2'], \n",
    "               'C':[0.001,.009,0.01,.09,1,5,10,25], \n",
    "               'solver':['lbfgs'], \n",
    "               'max_iter':[100,1000,2500]\n",
    "              }\n",
    "grid_clf_acc = GridSearchCV(clf, param_grid = grid_values,scoring = 'accuracy', n_jobs=-1)\n",
    "grid_clf_acc.fit(X_train, y_train)\n",
    "evaluate([LogisticRegression(**grid_clf_acc.best_params_)], X, y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Hyperparameter tuning seems to be of great help for our problem."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---\n",
    "\n",
    "# Alternative approach\n",
    "\n",
    "In the previous section we saw how hyperparameter tuning can largly improve performance of our models. However, the accuracy is still quite low compared to the existing literature. We suspect our very agnostic approach with regard to the provided dataset might be too agnostic for these simple models to detect meaningful patterns. \n",
    "\n",
    "In this section we take a more determined approach, making informed guesses on the way the structures of diterpenes work. We must note we are not chemistry professionals so our assumptions might not be fully correct, but they might be good enough."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Multiplicity assumption\n",
    "\n",
    "We know our diterpenes have 20 carbon atoms. The multiplicity of a carbon atom seems to be a very important information. We suspect the amount of carbons with a specific multiplicity may already be a big hint for models to discriminate between different diterpenes."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "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>a0</th>\n",
       "      <th>a1</th>\n",
       "      <th>a2</th>\n",
       "      <th>a3</th>\n",
       "      <th>a45c</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>6</td>\n",
       "      <td>52c</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>6</td>\n",
       "      <td>54c</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>6</td>\n",
       "      <td>54c</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>4</td>\n",
       "      <td>9</td>\n",
       "      <td>4</td>\n",
       "      <td>3c</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>3</td>\n",
       "      <td>4</td>\n",
       "      <td>9</td>\n",
       "      <td>4</td>\n",
       "      <td>3c</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   a0  a1  a2  a3 a45c\n",
       "0   2   4   8   6  52c\n",
       "1   2   4   8   6  54c\n",
       "2   2   4   8   6  54c\n",
       "3   3   4   9   4   3c\n",
       "4   3   4   9   4   3c"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = pd.read_csv(\"diterpene_shuf.csv\")\n",
    "\n",
    "# Consider only multiplicity counts\n",
    "df_simple = df[['a0', 'a1', 'a2', 'a3', 'a45c']]\n",
    "df_simple.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Rename columns for better readability\n",
    "df_simple = df_simple.rename(columns={'a0':'s_cnt', 'a1':'d_cnt', 'a2':'t_cnt', 'a3':'q_cnt', 'a45c':'Class'})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<seaborn.axisgrid.FacetGrid at 0x20b02659700>"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot each multiplicity per class\n",
    "sns.catplot(data=df_simple, y=\"Class\", x='s_cnt')\n",
    "sns.catplot(data=df_simple, y=\"Class\", x='d_cnt')\n",
    "sns.catplot(data=df_simple, y=\"Class\", x='q_cnt')\n",
    "sns.catplot(data=df_simple, y=\"Class\", x='t_cnt')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can see there could be some regularities in this, for example diterpene 52c (our majority class) never has 1 singulet, it always has 2, 3 or 4. This seems promising. Let's check."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "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>Algorithm</th>\n",
       "      <th>Accuracy</th>\n",
       "      <th>F1 macro</th>\n",
       "      <th>F1 micro</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>LogisticRegression</td>\n",
       "      <td>0.726141</td>\n",
       "      <td>0.245911</td>\n",
       "      <td>0.726141</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>KNeighborsClassifier</td>\n",
       "      <td>0.639004</td>\n",
       "      <td>0.300120</td>\n",
       "      <td>0.639004</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>DecisionTreeClassifier</td>\n",
       "      <td>0.755187</td>\n",
       "      <td>0.378595</td>\n",
       "      <td>0.755187</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                Algorithm  Accuracy  F1 macro  F1 micro\n",
       "0      LogisticRegression  0.726141  0.245911  0.726141\n",
       "1    KNeighborsClassifier  0.639004  0.300120  0.639004\n",
       "2  DecisionTreeClassifier  0.755187  0.378595  0.755187"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X = df_simple[['s_cnt', 'd_cnt', 't_cnt', 'q_cnt']]\n",
    "y = df_simple['Class']\n",
    "evaluate(clfs_to_try, X, y, plot=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Indeed, it seems like the information about multiplicity brings good discriminatory value to our models."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Insignificance of order assumption\n",
    "\n",
    "Although the multiplicity already brings a lot of information, we want to also somehow incorporate the frequency information into our models."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "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>a5</th>\n",
       "      <th>a6</th>\n",
       "      <th>a7</th>\n",
       "      <th>a8</th>\n",
       "      <th>a9</th>\n",
       "      <th>a10</th>\n",
       "      <th>a11</th>\n",
       "      <th>a12</th>\n",
       "      <th>a13</th>\n",
       "      <th>a14</th>\n",
       "      <th>...</th>\n",
       "      <th>a35</th>\n",
       "      <th>a36</th>\n",
       "      <th>a37</th>\n",
       "      <th>a38</th>\n",
       "      <th>a39</th>\n",
       "      <th>a40</th>\n",
       "      <th>a41</th>\n",
       "      <th>a42</th>\n",
       "      <th>a43</th>\n",
       "      <th>a44</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>t</td>\n",
       "      <td>139.2</td>\n",
       "      <td>q</td>\n",
       "      <td>33.2</td>\n",
       "      <td>d</td>\n",
       "      <td>30.7</td>\n",
       "      <td>d</td>\n",
       "      <td>49.6</td>\n",
       "      <td>t</td>\n",
       "      <td>38.6</td>\n",
       "      <td>...</td>\n",
       "      <td>d</td>\n",
       "      <td>134.5</td>\n",
       "      <td>t</td>\n",
       "      <td>24.2</td>\n",
       "      <td>q</td>\n",
       "      <td>14.4</td>\n",
       "      <td>t</td>\n",
       "      <td>42.2</td>\n",
       "      <td>t</td>\n",
       "      <td>18.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>d</td>\n",
       "      <td>144.0</td>\n",
       "      <td>q</td>\n",
       "      <td>66.7</td>\n",
       "      <td>d</td>\n",
       "      <td>46.2</td>\n",
       "      <td>t</td>\n",
       "      <td>32.7</td>\n",
       "      <td>q</td>\n",
       "      <td>19.1</td>\n",
       "      <td>...</td>\n",
       "      <td>t</td>\n",
       "      <td>22.4</td>\n",
       "      <td>s</td>\n",
       "      <td>37.7</td>\n",
       "      <td>q</td>\n",
       "      <td>19.7</td>\n",
       "      <td>q</td>\n",
       "      <td>16.5</td>\n",
       "      <td>t</td>\n",
       "      <td>26.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>d</td>\n",
       "      <td>30.7</td>\n",
       "      <td>t</td>\n",
       "      <td>26.8</td>\n",
       "      <td>t</td>\n",
       "      <td>41.2</td>\n",
       "      <td>t</td>\n",
       "      <td>139.3</td>\n",
       "      <td>q</td>\n",
       "      <td>19.8</td>\n",
       "      <td>...</td>\n",
       "      <td>t</td>\n",
       "      <td>16.7</td>\n",
       "      <td>t</td>\n",
       "      <td>35.4</td>\n",
       "      <td>q</td>\n",
       "      <td>65.8</td>\n",
       "      <td>s</td>\n",
       "      <td>42.4</td>\n",
       "      <td>d</td>\n",
       "      <td>139.8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>t</td>\n",
       "      <td>37.2</td>\n",
       "      <td>t</td>\n",
       "      <td>17.4</td>\n",
       "      <td>q</td>\n",
       "      <td>16.7</td>\n",
       "      <td>s</td>\n",
       "      <td>58.1</td>\n",
       "      <td>t</td>\n",
       "      <td>41.6</td>\n",
       "      <td>...</td>\n",
       "      <td>s</td>\n",
       "      <td>33.5</td>\n",
       "      <td>d</td>\n",
       "      <td>44.9</td>\n",
       "      <td>d</td>\n",
       "      <td>56.6</td>\n",
       "      <td>q</td>\n",
       "      <td>20.8</td>\n",
       "      <td>t</td>\n",
       "      <td>36.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>q</td>\n",
       "      <td>12.2</td>\n",
       "      <td>d</td>\n",
       "      <td>154.6</td>\n",
       "      <td>t</td>\n",
       "      <td>18.7</td>\n",
       "      <td>t</td>\n",
       "      <td>37.1</td>\n",
       "      <td>t</td>\n",
       "      <td>79.1</td>\n",
       "      <td>...</td>\n",
       "      <td>d</td>\n",
       "      <td>49.5</td>\n",
       "      <td>t</td>\n",
       "      <td>38.8</td>\n",
       "      <td>t</td>\n",
       "      <td>71.4</td>\n",
       "      <td>s</td>\n",
       "      <td>36.8</td>\n",
       "      <td>s</td>\n",
       "      <td>46.2</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 40 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "  a5     a6 a7     a8 a9   a10 a11    a12 a13   a14  ... a35    a36 a37   a38  \\\n",
       "0  t  139.2  q   33.2  d  30.7   d   49.6   t  38.6  ...   d  134.5   t  24.2   \n",
       "1  d  144.0  q   66.7  d  46.2   t   32.7   q  19.1  ...   t   22.4   s  37.7   \n",
       "2  d   30.7  t   26.8  t  41.2   t  139.3   q  19.8  ...   t   16.7   t  35.4   \n",
       "3  t   37.2  t   17.4  q  16.7   s   58.1   t  41.6  ...   s   33.5   d  44.9   \n",
       "4  q   12.2  d  154.6  t  18.7   t   37.1   t  79.1  ...   d   49.5   t  38.8   \n",
       "\n",
       "  a39   a40 a41   a42 a43    a44  \n",
       "0   q  14.4   t  42.2   t   18.6  \n",
       "1   q  19.7   q  16.5   t   26.6  \n",
       "2   q  65.8   s  42.4   d  139.8  \n",
       "3   d  56.6   q  20.8   t   36.5  \n",
       "4   t  71.4   s  36.8   s   46.2  \n",
       "\n",
       "[5 rows x 40 columns]"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Get the rest of features\n",
    "df_rest = df.drop(columns=['a0', 'a1', 'a2', 'a3','a4', 'a45c'])\n",
    "df_rest.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As previously mentioned, we know we have multiple multiplicity-frequency pairs for each diterpene structure. We will assume the order of the pairs is not important, and sum up frequencies for each multiplicity as new features for our models."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [],
   "source": [
    "t = []\n",
    "q = []\n",
    "s = []\n",
    "d = []\n",
    "cols = df_rest.columns.values\n",
    "\n",
    "# Consider each structure individually\n",
    "for index, row in df_rest.iterrows():\n",
    "    # Add up frequencies for each multiplicity\n",
    "    tval = qval = sval = dval = 0\n",
    "    for i in range(1, len(cols), 2):\n",
    "        mul = row[cols[i-1]]\n",
    "        freq = row[cols[i]]\n",
    "        if mul == 't':\n",
    "            tval = tval + freq\n",
    "        elif mul == 'q':\n",
    "            qval = qval + freq\n",
    "        elif mul == 's':\n",
    "            sval = sval + freq\n",
    "        elif mul == 'd':\n",
    "            dval = dval + freq\n",
    "    t.append(tval)\n",
    "    q.append(qval)\n",
    "    s.append(sval)\n",
    "    d.append(dval)\n",
    "    \n",
    "X['t_freq'] = t\n",
    "X['q_freq'] = q\n",
    "X['s_freq'] = s\n",
    "X['d_freq'] = d"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "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>s_cnt</th>\n",
       "      <th>d_cnt</th>\n",
       "      <th>t_cnt</th>\n",
       "      <th>q_cnt</th>\n",
       "      <th>t_freq</th>\n",
       "      <th>q_freq</th>\n",
       "      <th>s_freq</th>\n",
       "      <th>d_freq</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>6</td>\n",
       "      <td>368.0</td>\n",
       "      <td>323.3</td>\n",
       "      <td>69.8</td>\n",
       "      <td>266.1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>6</td>\n",
       "      <td>415.2</td>\n",
       "      <td>199.1</td>\n",
       "      <td>75.7</td>\n",
       "      <td>371.1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>6</td>\n",
       "      <td>347.0</td>\n",
       "      <td>463.2</td>\n",
       "      <td>80.8</td>\n",
       "      <td>252.7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>4</td>\n",
       "      <td>9</td>\n",
       "      <td>4</td>\n",
       "      <td>473.9</td>\n",
       "      <td>139.1</td>\n",
       "      <td>131.7</td>\n",
       "      <td>234.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>3</td>\n",
       "      <td>4</td>\n",
       "      <td>9</td>\n",
       "      <td>4</td>\n",
       "      <td>409.9</td>\n",
       "      <td>203.4</td>\n",
       "      <td>123.0</td>\n",
       "      <td>323.1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   s_cnt  d_cnt  t_cnt  q_cnt  t_freq  q_freq  s_freq  d_freq\n",
       "0      2      4      8      6   368.0   323.3    69.8   266.1\n",
       "1      2      4      8      6   415.2   199.1    75.7   371.1\n",
       "2      2      4      8      6   347.0   463.2    80.8   252.7\n",
       "3      3      4      9      4   473.9   139.1   131.7   234.0\n",
       "4      3      4      9      4   409.9   203.4   123.0   323.1"
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "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>Algorithm</th>\n",
       "      <th>Accuracy</th>\n",
       "      <th>F1 macro</th>\n",
       "      <th>F1 micro</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>LogisticRegression</td>\n",
       "      <td>0.651452</td>\n",
       "      <td>0.237991</td>\n",
       "      <td>0.651452</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>KNeighborsClassifier</td>\n",
       "      <td>0.721992</td>\n",
       "      <td>0.439648</td>\n",
       "      <td>0.721992</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>DecisionTreeClassifier</td>\n",
       "      <td>0.829876</td>\n",
       "      <td>0.600282</td>\n",
       "      <td>0.829876</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                Algorithm  Accuracy  F1 macro  F1 micro\n",
       "0      LogisticRegression  0.651452  0.237991  0.651452\n",
       "1    KNeighborsClassifier  0.721992  0.439648  0.721992\n",
       "2  DecisionTreeClassifier  0.829876  0.600282  0.829876"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "evaluate(clfs_to_try, X, y, plot=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "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>Algorithm</th>\n",
       "      <th>Accuracy</th>\n",
       "      <th>F1 macro</th>\n",
       "      <th>F1 micro</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>LogisticRegression</td>\n",
       "      <td>0.771784</td>\n",
       "      <td>0.382005</td>\n",
       "      <td>0.771784</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            Algorithm  Accuracy  F1 macro  F1 micro\n",
       "0  LogisticRegression  0.771784  0.382005  0.771784"
      ]
     },
     "execution_count": 59,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Do a grid search\n",
    "clf = LogisticRegression(random_state=REP)\n",
    "grid_values = {'penalty': ['l2'], \n",
    "               'C':[0.001,.009,0.01,.09,1,5,10,25], \n",
    "               'solver':['lbfgs'], \n",
    "               'max_iter':[100,1000,2500]\n",
    "              }\n",
    "grid_clf_acc = GridSearchCV(clf, param_grid = grid_values,scoring = 'accuracy', n_jobs=-1)\n",
    "grid_clf_acc.fit(X_train, y_train)\n",
    "evaluate([LogisticRegression(**grid_clf_acc.best_params_)], X, y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "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>Algorithm</th>\n",
       "      <th>Accuracy</th>\n",
       "      <th>F1 macro</th>\n",
       "      <th>F1 micro</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>DecisionTreeClassifier</td>\n",
       "      <td>0.842324</td>\n",
       "      <td>0.500405</td>\n",
       "      <td>0.842324</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                Algorithm  Accuracy  F1 macro  F1 micro\n",
       "0  DecisionTreeClassifier  0.842324  0.500405  0.842324"
      ]
     },
     "execution_count": 58,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Do a grid search\n",
    "clf = DecisionTreeClassifier(random_state=REP)\n",
    "grid_values = {'ccp_alpha': [0.1, 0.01, 0.001, 0.0001, 0],\n",
    "               'criterion': ['gini', 'entropy'],\n",
    "               'max_depth': [5, 6, 7, 8, 9, 10, None],\n",
    "               'max_features': ['auto', 'sqrt', 'log2']\n",
    "              }\n",
    "grid_clf_acc = GridSearchCV(clf, param_grid = grid_values,scoring = 'accuracy', n_jobs=-1)\n",
    "grid_clf_acc.fit(X_train, y_train)\n",
    "evaluate([DecisionTreeClassifier(**grid_clf_acc.best_params_)], X, y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "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>Algorithm</th>\n",
       "      <th>Accuracy</th>\n",
       "      <th>F1 macro</th>\n",
       "      <th>F1 micro</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>KNeighborsClassifier</td>\n",
       "      <td>0.854772</td>\n",
       "      <td>0.552895</td>\n",
       "      <td>0.854772</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              Algorithm  Accuracy  F1 macro  F1 micro\n",
       "0  KNeighborsClassifier  0.854772  0.552895  0.854772"
      ]
     },
     "execution_count": 60,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAT4AAAEHCAYAAAA3TSpVAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjQuMywgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/MnkTPAAAACXBIWXMAAAsTAAALEwEAmpwYAAA0MklEQVR4nO2debwUxbn+v885h0VFRQQMAooKGhcEFbeA4hKJGndcrjFejN4fmutG1CguN5smEndjvFGMRqLodQWNMRFECJqIsgiIopIoKkLY3HBneX9/dI0Mh1l6Znp6pk/Xl09/Zqan3nqqZ4b3VHd1PSUzw+PxeNJEQ60b4PF4PHHjE5/H40kdPvF5PJ7U4ROfx+NJHT7xeTye1OETn8fjSR1NtW5AuXTcvINt3b1rrZvhSQu2pvxYJbd/MWPWnGVm1qmSOtTQ1iDk52crnzKzQ/PWJbUHfg/sAhhwOvA68ADQA5gPnGhmHxSSSWzi27p7V6aMH1vrZnhSgn31edmxar1BhC2Jl9ade75deS1raGoKlztXrVzYsUiRm4G/mtnxkloDGwKXARPMbISk4cBw4JJClST3T5HH40kEAiSF2grWI20C7A/cCWBmX5nZh8DRwChXbBRwTLE2tcjEN27iC+y63yns3P9krv3tvbHFeu10aZ958fVs3e9E+n1naElxUWjXQ3wpRJH4gG2BpcAfJL0k6feSNgK2MLNFAO6xc7GKqpb4JN0laYmkOVn7OkgaL2mee9wsat3Vq1cz7PIbeezea3lp4h95aOwE5r4xv+qxXjt92qcOHsTYu38ZunyU2rWOLxWF/Ad0lDQta8v+q9IE7A78zsx2Az4lOK0tmWr2+O4Gml+kHE5wLt4LmECZjS7E1Jfmsl2Prmyz9Za0bt2KE44+mCeeeq7qsV47fdoD9u5Nh/Ybhy4fpXat40tBiMaGhlAbsMzM+mVtI7OqWgAsMLMX3OuHCRLhYkldANzjkmJtqlriM7PJwPvNduc8F5fUKOk6SS9Lmi3p3HJ1F/57Gd22XNvT7dqlE+/9e2nVY712+rQroZbHHUV8qZTQ48uLmf0beFfSDm7XwcCrwOPAELdvCPBYsfbEPaq7zrm4pMwnPxTYBtjNzFZJ6lCuQC63mRDXDiqO9drp066EWh53FPElIWiIru5zgdFuRPdN4AcEHbgHJZ0BvAOcUKySermd5dvAbWa2CsDMmvcUAXDn+0MBtuq2Zc6KunbpxIKFa3u67y1aypZbFBshrzzWa6dPuxJqedxRxJdKVEnVzGYC/XK8dXAp9cQ9qpvvXFwENyMWxMxGZs79O26eu1PYr+83+edbC5j/zkK++molDz02ge8O6h+qcZXEeu30aVdCLY87ivhSEKIh5BYXcff4MufiI1j3XHwccJakSZlT3Xy9vmI0NTVx41XDOPJ7F7F6zRqGnHQ4O+2wTdVjvXb6tIecdzWTp8xm+Qcf0XPfU7hi2KmcdlLeSQeRatc6vlTiuoQQFlXLgVnS/cABQEdgMfBTYCzwILAV7lzczN6X1ARcQzAKvBK4w8x+W6j+Pfr2Nj9zwxMXKZ65Md3Mcp1ahqZVY1tr327rUGWXffxGxXphqFqPz8xOzvPWeufi7treBW7zeDwtjDhPY8NQL4MbHo+npRJuVkas+MTn8XiqTrF79OLGJz6Px1NVRKT38UWCT3wej6fqNNaZJ6FPfB6Pp6oIf43P4/GkkHob1a2v/mdEpNEbzmt7P76440tBagi1xUXcfnw/k/SepJluOzxq3bR6w3lt78fXQvz4YiFuPz6AG82sr9uejFo0rd5wXtv78dWtH59EQ8gtLuL248uJ9+Pz2knUroRa++nFfeyNagi1xUUtBjfOkfSfwDTgQrcMnPfj89qJ066EWvvpxXnsov5uYI57cON3wHZAX2ARcL3bH9qPL+PFv2x57s5kWr3hvHb82pVQaz+9uI89Nae6uTCzxWa22szWAHcAe7m3vB+f106cdiXU2k8v3mMPO7TRQv34JHXJWM8DxwKZEV/vx+e1E6ft/fjCUY9T1uL24zuA4DTXgPnAmW7tDe/H56lrvB9f+WzYqp312nzXUGVnL36+Rfrx3ZmnrPfj83haMH6ursfjSRWZNTfqCZ/4PB5PdVH9rbnhE18ZfHnjJRXFt/nRryNqSbKo5DoZ1PZaWSXaST7uqKi3+/h84vN4PFVF1J87i098Ho+nyqjuBjfqqzUREadN0dl//Sc9b53Kvn+Y+fW+H/zpDQaMmsWAUbPoPXIGA0bNqnq7K42vpXYt7Z3SetxRxIcl0+OrpwXFq2lL1V3SRElzJb0i6Xy3/1pJrzkzgjGS2kepG7dN0fd27szDx++4zr4/HLk9zw3pw3ND+nBUrw4c2av41OMk2zMl1d4prccdRXypKOQWF9Xs8a0iMCHYEdgHOFvSTsB4YBcz2xV4A7g0StG4bYr6d9+EzdrmvmJgZox9YznH71h8DmSS7ZmSau+U1uOOIr5UopqrK2m+c3GaKWma29dB0nhJ89zjZkXbE8Ex5cTMFpnZDPd8BTAX6Gpm4zJmBMAUoJtrfCTWVPVkU/SPBSvotGErttus+Khcku2ZkmrvlNbjjiK+VCKeq3ug8/PMzPAYDkwws17ABPe6ILEMbkjqAewGvNDsrdOBB9zzSKyp6smm6JHXljH4m+EcL5Jsz5RUe6e0HncU8aUQw6ju0QTTYQFGAZOAgvecVX1wQ1I74BFgmJl9nLX/coLT4dFuV1Frqnq3pcpm1RrjT/Pe57hvbh6LdlKtoSrFH3d52vG2XVEakRowTtJ0SZlRoS0y5ifusXPeaEdVE5+kVgRJb7SZPZq1fwhwBHCKrf3TU9Saqt5tqbKZ9PaH9OrQlq4bt4lFO6nWUJXij7v+254xIg15qtsx07lxW/Mh7/5mtjtwGMG4wf7ltKlqp7oK+s13AnPN7Ias/YcSdEMHmtlnWSGRWFPFbVN0xhNv8Ny7H7P881XsdNt0hvfvxn/23oJHXlvO8SFPcyttd6XxtbY4qpW9U1qPO4r4Uimhh7WskDuLmS10j0skjSHw9FycsbyT1AVYki8+QzVtqQYAzwIvA2vc7suA3wBtgOVu3xQzO6tUa6pa2lL5KWvlkdapW0k+7ihsqTZtvan1/0a43uRf3v1LXj1JGwENZrbCPR8P/AI4GFhuZiMkDQc6mNnFhXSqaUv1HLlvzcm5spq3pvJ4WiYRDm5sAYxxgzBNwH1m9ldJU4EHJZ0BvAOcUKwiP2XN4/FUmWhs5c3sTaBPjv3LCXp9ofGJz+PxVBUBTd6kwOPxpArvx9cyqHRwYs37C8uObeiwZUXatSSpgxOVktbjzqbe3FB84vN4PFWlHhcU94nP4/FUnXozIq23HmgkJNGf7Ysvv2L/ky9n78EX0++Yi7jq1odi06401msnTzuK+LAI0Rhyi4ta+PH1lTQlYysjaa8odZPqz9amdSuevPN/eOGRa3j+oRGM//tMXpw1LxbttHoBplU7ivhSSY0RKfn9+K4Bfm5mfYGfuNeRkVR/Nkm027AtACtXrWblqtWUMhCW1OP22rXx04vbjy81RqT5/PgIjAg2ccU2BRZCy/DjqzR+9eo17HP8JfQYOJSD9unNnrv2ikU7rV6AadWOIr4URHRGpFFRCz++YcBTkq4jSLzfcsUS78dXaXxjYwNTHv41H378KScPu55X5r3Lzr26V107rV6AadWOIr5U6m0woRZ+fD8EfmRm3YEfETi4QAvw44vK46z9Jhux3547Mf7vM2PRTqsXYFq1o4gvlYgdmCumFn58Q4DM84cIbGWgBfjxVRK/9P2P+fDjTwH4/IuvmDjlZXbYJvzNykk9bq9dGz+9eP346m9UN3Y/PoJregMJ7KEPAjJDl4n044sq/t9LP2DoFb9j9eo1rLE1DB60L4cN3CMW7bR6AaZVO4r4Uqm3U91a+PF9DNxMkHS/AP7bzKYnyY+vUtI6Zc2TPKLw4+vcpoMN7vLtUGVve/uhivXCUAs/PoD1ujLej8/jabnU17wNP2XN4/HEQL1NWfOJz+PxVJXAgbm+8ImvBvjrdMmjknUzvC0VsY7YhsEnPo/HU3XqK+35xOfxeKpMhIsNRYZPfB6Pp+rU2zW+emtPJCTVI81rJ0v7zIuvZ+t+J9LvO0NLiotCux7iSyE17iyS2kp6UdIs58f3c7f/BPd6jaTIb1RMqkea106e9qmDBzH27l+GLh+ldq3jS0GIppBbXFSzx/clcJCZ9QH6AodK2geYAxwHTK6GaFI90rx28rQH7N2bDu03Dl0+Su1ax5dKanp8FvCJe9nKbWZmc83s9ebl0+7H57WTp10JtfbTi92Pr84cmKs6uCGpEZgO9ARuNbMXChQv6scnaagrx1bdct8Ll1SPNK+dPO1KqLWfXux+fNWxBCibqg5umNlqZzHfDdhL0i4Fihf14wtjS5VUjzSvnTztSqi1n16cx56ZuRFmC1VfcHb4kqQn3OsOksZLmuceNytWRyyjumb2IYEN1aEFihX14wtDUj3SvHbytCuh1n56cR97xNf4zidYyiLDcGCCmfUCJrjXBammH18nYKWZfShpA4Ie3a8LhKTaj89rJ097yHlXM3nKbJZ/8BE99z2FK4adymknFfrbHp12reNLQRDZiK2kbsB3gV+y1snpaOAA93wUQSfrkoL1VNGPb1fXiEaCnuWDZvYLSccCtwCdgA+BmWb2nTT58XmSR1rn6kbhx9etzeZ2zjcOC1X20ndGF9ST9DBwNbAxcJGZHSHpQzNrn1XmAzMreLpbTT++2QQLDDXfPwYYk2O/9+PzeFooJVxT6yhpWtbrkWY2EkDSEcASZ1x8QCXt8VPWPB5PVZGVNKq7rECPrz9wlKTDgbbAJpLuBRZL6mJmiyR1AZbkif+aFjllzePx1BdRDG6Y2aVm1s3MegD/ATxjZt8HHidYxAz3+Fix9vgeXxlUcr0Hkn3NJ63476wyqtzDGgE8KOkM4B3ghGIBPvF5PJ6qIqAp4jFUM5tEMHqLmS0HDi4l3ic+j8dTderLja+FXuPzNkXpsYby2vVvSxX1zI0oqIUt1ZXOhGCmpHGSIl2AwtsUpcsaymvXvy0VpCjxkd+W6loz29XN4X0C+EmUot6mKF3WUF47IbZUFm6Li1rYUn2cVWwj3PzclmBLVQlJtmfy2snSjiK+FIIpa+G2uKiJLZWkXwL/CXwEHOiKF7WlCoO3KSo93munSzuK+FLxtlTB/svNrDswGjjHFS9qSyVpqKRpkqYtW57bv8DbFJUe77XTpR1FfCmEvXm5LhyYJd0i6Tf5tlJECthS3QcMzkhSxJYqjB+ftylKlzWU106GLVW9DW4UOtWdVuC9ouSzpZLUy8zmuWJHAa+554m3pUqqTZHXTpd2FPGlIOrvVDe0LZWkjczs09AV57elegTYAVgDvA2cZWbvJcmWyk9Z86SFKGyptm29uV3ZMZwt1fcXFbalioqigxuS9gXuBNoBW0nqA5xpZv9dKK6ALdXgHMW9LZXH00KRQVOVfD/LJcxp9U3Ad4DlAGY2C9i/im3yeDwtjCRd4/saM3u32VD36uo0x+PxtETqba5umMT3rqRvASapNXAe6y70kTrSeo3OX9v0lEMwuJG8U92zgLOBrsB7BNPPzq5imzweTwsjcae6ZrYMOCWGtng8nhaIMBorXzk2UoomWUnbSvqTpKWSlkh6TNK2cTTO4/G0DBos3BZbe0KUuQ94EOgCbAk8BNxfzUZVSlI90pKq7X0Ik6cdRXwpCAu1xUWYxCczu8fMVrntXopMLYP8fnzuvXMlve72X1PJATQnqR5pSdb2PoTJ0o4ivhQSZUQqqYNzSJkoabikHpK2lnQx8OcQdef045N0IMHK57ua2c7AdZUfxlqS6pGWZG3vQ5gs7SjiSyUxiY/ATmoacBJwJjCRwGjgh8APilWcz4/PxY8wsy9duSXQMvz40qpdCUk+7qRqRxFfKvV2qpt3VNfMKp6xnMuPT9L2wH7Ok+8L4CIzm0oIPz5JQ105tuqW27E+qR5pSdauhCQfd1K1o4gvBUHdjeqGmrnhfPR2Ili9HAAz+2OxODNbDfSV1B4Y4+ppAjYD9gH2JFgPc1tC+PGZ2UhgJAQmBbk0k+qRlmTtSkjycSdVO4r4Ummos8QX5naWnwK3uO1AAgeVo0oRaebHtwB41J0Kv0jg0tKREH58YUiqR1qStSshycedVO0o4ktBQIPCbXERpsd3PNAHeMnMfiBpC+D3xYLy+fEBnwAHAZPcaW9rYBktwI8vrdrehzBZ2lHEl0a81+/CUNSPT9KLZraXpOkEPb4VwBw3IlsoLp8fX2vgLoKR3q8IrvE9kyQ/vrTi5+qmjyj8+L7ZajO7c/ODQ5UdsPiR+vDjA6a5a3R3EAxUfAK8WCyogB/fV8D3c+z3fnweTwtFca4dGYIwc3UzhqO3SforsIlLah6Px1MUCRojSHyS2gKTgTYEuethM/upuwPkAaAHMB840cw+KFRX3sQnafdC75nZjNKb7vF40khDND2+zKSITyS1Ap6T9BfgOGCCmY2QNBwYDlxSqKJCPb7rC7xnBAMUqWTN+wsrim/okPsexHqn0mt0bbfcr6L4LxY+W3Zskq9PVvJ7q4ffmrBITnUtGJDINSniaOAAt38UwR0k5SU+Mzsw33sej8dTChH1+PJNitjCzBYBmNkiSZ0LVkLIG5g9Ho+nEkqYFNJRUvbStiPdxAUg76SIkolzXnBs1Moq6Isvv2L/ky9n78EX0++Yi7jq1odi0640Pm7tTTdpx313XMmsZ0czc/K97L3Hzhx3xIHMmHQPn703md377BBL22tpqVXL31ql+qUgQWPDmlAbsMzM+mVtI3PV2WxSxGJJXQItdQGW5IrJJvbEV8iuKgpqaRXUpnUrnrzzf3jhkWt4/qERjP/7TF6cNa94YATaSbNIuv7K8xk/8QX67HcKex58Gq/Ne5tXXn+Tk864jOemzKqqdja1stSq5W8tCv1SkcJthetQJ9fTI2tSxGvA48AQV2wI8Fix9oSZsiZJ35f0E/d6K0l7FYsrQE67qgrqW4daWgVJot2GwXTmlatWs3LV6tBd/DRZJG3cbkMG7NOHP9z3BAArV67io48/4fV5bzPvX++G1o2i7bWy1Krlby0K/VKRLNRWhC4ENnmzganAeDN7AhgBHCJpHnCIe12QMD2+/wX2BU52r1cAt4aIy0k+uypJPSU97XqCMyRtV079tbZnWr16Dfscfwk9Bg7loH16s+euvWLRTpJF0jZbb8nS5R9yx02XMWXcXfzuukvYcIO2ectXs+2VkNTfWlT6YRFGg8JthTCz2Wa2m5ntama7mNkv3P7lZnawmfVyj0WnuoZJfHub2dkEFlK4GwNbh4jLi/Pem0lwLj7ezF4ARhOM0vQBvgUsKqfuWtszNTY2MOXhX/PG0//L9Dn/4pV54XowabJIampqZLfe2zNy1Fj2GXQ6n37+BT8+d73JPFXRjpKk/tai0g+NQA0WaouLMIlvpRtCNvjafGBNJaJmttrM+gLdgL0k9Qa6mtkY9/4XZvZZ8zhJQyVNkzRt2fLcSb1e7Jnab7IR++25E+P/PjMW7SRZJL23cCnvLVrK1JdeBWDMExPp23v70HqVaEdJUn9rUeuHIYprfFESJvH9BhgDdHbmoc8Bv4pCPGtk5tiQ5UdmRns6br6eTylQW6ugpe9/zIcffwrA5198xcQpL7PDNuFuIE2TRdLipe+zYOESem3XHYADB/Qr+8J6Ui21avlbi0K/JAQNjWtCbXERZq7uaOfMcjCBtdYxZja3XMECdlULJB1jZmMltQEac/X6ilFLq6B/L/2AoVf8jtWr17DG1jB40L4cNnCPWLSTZpH0o8tv5O5bf0rrVk289c5Chg67mqMO258brhpGp83bM+aea5n9yjyOPPnCqra9VpZatfytRaFfCoEfX32ZFISxpdoq134ze6cswfx2Vb2A2wlMSVcCJ5jZm/nqqaUtVVqnrFWKn7JWHrWcshaFLVXvDTa1sduFu3Gj5yvj6saW6s8E1/dEYD2/DfA6UNCPLx8F7KrmkeL5vx5PyyWaubpREuZUt3f2a+facmbVWuTxeFoWblS3nih5rq6ZzZC0ZzUa4/F4Wh4i3hHbMBRNfJKyHZEbgN2BeO4Q9Xg8yUfQ0BTfiG0YwvT4sufzrCK45vdIdZqTDNI6OFEplQxOQGUDFEle76Ml/N4S1eNzNy63M7Mfx9Qej8fTAknMNT5JTW6Zx7wW9B6Px1MMCVRnBniFmpNZSW2mpMclnSrpuMwWR+PKJUm+dF47uX56lcbXUjuK+PCEc2aJ85aXMHm4A7Cc4B67I4Aj3WMonCHBS5KecK8fkDTTbfOdWUFkJM2Xzmsn00+v0vhaakcRXxKChiYLtcVFocTX2Y3ozgFedo+vuMc5JWicD3w9xc3MTjKzvs6k4BHg0VIbXYgk+dJ57eT66VUaX0vtKOJLRQ3htrgoJNUItHPbxlnPM1tRJHUDvgv8Psd7Ak4E7nevGyVdJ+llSbMlnVvKgWRIki+d106un16l8bXUjiK+ZGThtpgoNKq7KGP0VwE3ARez7i0xGfYDFrupagBDCabD7eYGVdazX5E01JVjq265h/iT5EvntZPrp1dpfC21o4gvhaQNblT0KUg6AlhiZtPzFDkZ19tzfBu4zcxWAeRyUQ1jS5UkXzqvnVw/vUrja6kdRXypJOlU9+AK6+4PHCVpPvB/wEGS7oXgVhmC1c8fyCovnNlpJSTJl85rJ9dPr9L4WmpHEV8Sqr/EV2hB8aK+9YUws0uBSwEkHQBcZGYZf/FvA6+Z2YKskHHAWZImZU51y2lD0nzpvHYy/fQqja+ldhTxpaI6W8G7qB9fJCJrE98R7vXdwBQzuy2rTBNwDcE6mSuBO8zst/nqrKUfn6c2pHXKWi2Jwo+vb/uNbcJ+4eZBdHxict348VWMmU0isJjPvD4tR5lVwAVu83g8LYh6G9yosw6ox+NpcYhwUyVixCc+j8dTVQSoob7sWXzi8ySGSq7TrRz3h4q0Ww36QUXxqcb3+DweTxpRk+/xeTyeNCFBnZ3q1lkHNBqSahXktePR/u87n2Pb8+5n7yvGfL1vzNS32OvyMWx6+h+Y8dayWNpea1up+Gypgmt8YbaCdUjdJU2UNFfSK5LOd/s7SBovaZ573KxYe2qW+JrbVUVFUq2CvHZ82qcM6MmjFxyyzr6dum7G6HMOov/236iqdhSx9RBfEplrfGG2wqwCLjSzHYF9gLMl7QQMByaYWS9ggntdkFr2+Naxq4qKpFoFee34tPvv8A02a9dmnX07bNmeXl02Da1badtrbSsVty0VDQq3FcDMFpnZDPd8BUH+6AocDYxyxUYBxxRtTiXHUi657Kok9ZT0tKRZkmZI2q6cupNqFeS149eulCQfd6yfm0BNDaG20FVKPYDdgBeALcxsEQTJEehcIBSo3eDGTaxvVzUaGGFmYyS1pcyknFSrIK8dv3alJPm44/3cShrc6ChpWtbrkWY2cp3apHYEJsbDzOzjctode+LLtqtyc3iRtDHQ1czGAJjZF3lii/rxJdUqyGvHr10pST7uOD+3YEHx0MlpWaG5upJaESS90WaWcW9fLKmLmS2S1AVYki8+Qy1Oddezq2Lt+XlBwvjxJdUqyGsnwF4pQu1aH3fctlRRXONzru13AnPN7Iastx4HhrjnQ4DHijUp9h5fHruq4yRNkXSMmY2V1AZoNLPPSq0/qVZBXjs+7R/cNonnXvs3yz/5gm9e8ACXHbMbm23Uhh+PnsKyFV9wwk3j6d29A2Mv+k6LOu4o40smmvv4+gOnAi9nLVJ2GTACeFDSGcA7wAnFKorFliqveJZdlaRewO1ARwJbqhPM7M18sd6WylMKfspaeURhS7X7Fu3tbycNCFV2k1v+3HJsqfKRbVfl1t44qJbt8Xg81aGUEds48FPWPB5Pdclc46sjfOLzeDxVpv6WWfOJz+PxVB/f40s+laz9AOld/6GWn1ulgxNtt9yv7NhPX/hdRdqN3XepKL7m+FNdj8eTOiTU1FjrVqyDT3wej6f61Nk1vvpqTUTUyiPtzIuvZ+t+J9LvO0NLiotCu9L4WmrX8nMrJ3bTTdpx3x1XMuvZ0cycfC9777Ezxx1xIDMm3cNn701m9z47hNbf6Zgr2OuUq9j31F+x32kjqt72KONLIoKZG5E2p5qVS5ov6WVJMzMTjyVdK+k1SbMljZHUPkrNWnqknTp4EGPv/mXs7a40vtbebrX63MqNvf7K8xk/8QX67HcKex58Gq/Ne5tXXn+Tk864jOemzCr5GJ68dRjP33MZz95d1Eau4rZHFV8SEjQ0hNtiIg6lA82sb9bd2OOBXcxsV+AN3PS1qKilR9qAvXvTof3GxQtWQTvJ3nC1+tzKid243YYM2KcPf7gv8M9duXIVH338Ca/Pe5t5/3q3rGMoh1p/ZyWTph5fLsxsnFs8HGAK0A2+dmS+zvUQZ0s6t5z6a+mRVgneE6884j7ubbbekqXLP+SOmy5jyri7+N11l7DhBm3LazyBa8nR593CgCFXc9fY8Ikncd+ZGsJtMVFtJQPGSZruLKWaczrwF/d8KLANsJvrDY5uXljSUEnTJE1btvz93II19EirBO+JVx5xH3dTUyO79d6ekaPGss+g0/n08y/48bnfD9fYHDw98kL+/sdLefTGcxj58N947qV5oeIS9Z25Ud0wW1xUO/H1N7PdgcMI/PH3z7wh6XICD/1Mgvs2cFumN2hm62W2MLZUtfRIqwTviVcecR/3ewuX8t6ipUx96VUAxjwxkb69ty+j5QFdOrUHoHOHjTlyYB+mvzo/VFzivrM0neqa2UL3uAQYA+wFIGkIcARwiq390yOCHmJF1NIjrRK8J155xH3ci5e+z4KFS+i1XXcADhzQr+xBgU8//5IVn37x9fNnXpzLTtvmNtiNou1RxpeEVHenulW7j0/SRkCDma1wzwcBv5B0KHAJMLCZ39444CxJk8xslaQOuXp9xailR9qQ865m8pTZLP/gI3ruewpXDDuV0046NBbtJHvD1epzKzf2R5ffyN23/pTWrZp4652FDB12NUcdtj83XDWMTpu3Z8w91zL7lXkcefKFBetZ8v4KTr7kdgBWrV7DiYP6cci+O1e17VHFl0ydzdyomh+fpG0JenkQJNj7zOyXkv4JtAGWu/emmNlZkpqAa4BDCfz47jCz3+arv5Z+fH7KWnkk+XNL65S1SPz4uney5y48LlTZjX40Mtl+fM5EtE+O/T3zlF8FXOA2j8fTUggW3ah1K9bBT1nzeDxVRtBUX6mmvlrj8XhaHt6dpWWQ1mt0lWJf5Vw1NDS1/Nw/GX9V2bGDj769Iu2xM26pKL72eCNSj8eTRmKchxsGn/g8Hk91EajOTnXrKw1HRFLtmdKovWDRUg479XJ2P+xs+n33HG4d9afYtCuNfePdpXzrrFu/3rY85ipuffQfReMaGsQN91/C5TefBUCP7bsyYtSF3PzgZVx+05lssFG4ub/JsaUSNDaF22IidluqrPcukmSSIp0nk1R7prRqNzY28qvhpzPjL7cy8YFrGHnfk8z95zuxaFd63Nt378Q/bjubf9x2Ns/e+kM2aNOKI/vvVDTuiO8dyIK3Fn/9+uyffI97fvMY55/4K6ZMnMWxQw6uetu9LVX1aW5LhaTuwCEEq55HSlLtmdKq3aVzB3bbeTsgsHzaYdtuLFwcfsJOLY87m0kvvck2XTqw1RbtC5bbvHN7+g3YmfFj1vYMu27dmVem/xOAWVNeY9+D+1a97bHbUknhtpio1anujcDFZM3NbQm2VF67vPgMby9YzKy5b7Jnn/CT/uvFhuzhv73MCQf2LlrujB8PZtTNY7E1a2dMvfOvRex1QBD7rUN2p+MWmxWtp16+s1CI1PX41rOlknQU8J6ZNbeqLWpLFUowofZMadXO8Mmnn/O9837NNZf9F5u02zAW7aismb5auYonn3+NY/cvPLWs33678NH7K/jX3HUNS2/52WgOP3F/rh99MRts2IaVK1cX1ayH7yw80ZkUSLpL0hJJc7L2dZA0XtI891j0L0e1ryb2N7OFkjoD4yW9BlxOYFjQnKK2VC55DgXYqltuF4uk2jOlVRsCF+PvnTeCk44cyNGD9g0dV6l2VNZM46bOo2/PLnTerF3Bct/suy17DuzNHgN2plXrVmy4UVuGXfWf3HTFH/nZf98KwJZbdWaP/YobFdT6OyuZ6HpzdwO/Bf6YtW84MMHMRkga7l5fUrA5UbUmFzlsqQYS9OpmSZpP4L48Q9I3CGFLFcaPL6n2TGnVNjN+ePkt7LBtd877wdGhNaPQjsqa6eGJszn+wF2Llrv3lsf5r0P/h6Hf/SnXD/8Ds6e+wU1X/JFNXcKUxAn/7zs89XDxa22Js6VqbAy3FcHMJgPNO0VHA6Pc81HAMcXqid2Wysw6Z5WZD/Qzs2WSEm9L5bVLj39++lzuf2wSO2+/NfscPQyAn13wfQ4dGM6go5bHDfDZF1/xzIx/cfOw0pN2hv0O7cdhJwUevVOemcmEx6YUjUmeLVVV+1hbmNkiADNb5M4wCxK7LVWzMvNZm/gSY0vlKY81n3xQUXxDu+IX/avFqlf+Vnbs8ac+XJF2LaesRWJL1bOr/f2as0OV3XDw5W8Dy7J2jTSzkdllJPUAnjCzXdzrD82sfdb7H5hZwR9L7LZUzcr0yHrubak8nhZJSXN1l5WRaBdL6uJ6e12AJcUCWuTMDY/HU2dU93aWx4Eh7vkQ4LFiAX6ursfjqTLRubNIuh84AOgoaQHwU2AE8KCkMwgmRZxQrB6f+DweT3URkc3DNbOT87xVfJ5fFj7xeWLDPnivsgpqOLjRtPPAsmPHzig/FiobFKrlgNBa5G2pPB5PyshMWasjfOLzeDxVpv4cmOurNRGRRF+6NGsD7HTMFex1ylXse+qv2O+0EbFpJ/Uzr7WPYcnUmUlBTXp87sblFcBqYFWU62hmfMb+fP8NdO3SiQGHD+WIQQPYcfseVY312uXHZ3jy1mF0bF94vmuU2kn+zDM+hrvtvB0rPvmMAYMv5KD+fdix51ZVb3vJSLGajIahlj2+9Xz6oiCpvnRp1a6UpB53pdq19DEsi4jcWaKibk51JfWU9LSkWZJmSNqunHqS6kuXVu0Mkjj6vFsYMORq7hob/j9gUo87Sj+8uH0MS6YOHZhr1f/M+PQZcLubizcaGGFmYyS1pcyknFRfurRqZ3h65IV06dSeJe+v4KjzfsP2W2/BgN16VVU7yZ95hlr4GJaKAKm480qc1CrxNffpex3oamZjAMws5wKs3o+v5Wln6NKpPQCdO2zMkQP7MP3V+aESX1KPO4rPrFY+hqVTf/fx1aQ1eXz6wsR5P74Wpg3w6edfsuLTL75+/syLc9lp29x/2KLUTvJnXksfw5KpQ+v52Ht8+Xz6gMMlHWNmYyW1ARrN7LNS60+qL11atQGWvL+Cky+5HYBVq9dw4qB+HLJvcRfiSrWT/JnX0sewdOpvVLdqfnx5BfP49EnqBdwOdCTw4zvBWVvlxPvxJY/V784pXqgAjd0Lr2nRUqnllLUo/Pj22HE7+/sfrwlVdoO9jq9YLwyxp+F8Pn1mNg84KO72eDyeKiNBgx/c8Hg8aaPOpqz5xOfxeKpM/Y3q+sTniY20XqOzrz6vKL4+rKUqQPgen8fjSRtCja1q3Yh18InP4/FUGW9LFQtptCny2snTPvPi69m634n0+87QknUr1Y4iPjR1eANz1ZQk7SBpZtb2saRhkvpIel7Sy5L+JGmTKHUzdjuP3XstL038Iw+NncDcN+ZXPdZre+1S408dPIixd/+yeMEqaFcaXzJpcWcxs9ed7VRfYA/gM4Ibl38PDDez3u71j6PUTatNkddOljbAgL1706H9xqHLR6kdry2Vu48vzBYTcaXYg4F/mdnbwA7AZLd/PDAYQFKjpOtcT3C2pHPLEUqrTZHXTpZ2pSSq7Rkj0jBbTMSl9B/A/e75HOAogkV/TwC6u/1DgW2A3cxslaTcLgRFSKtNkddOlnalJK3tStvghqTWBInuIbfrdOBsSdOBjYGv3P5vA7eZ2SoAM1vPTlbSUEnTJE1btjy322xabYq8drK0KyVRba9DI9I4lA4DZpjZYgAze83MBpnZHgS9wH+5ciIwKM2Lt6Xy2i1Fu1IS1/Y6G9yI41T3ZNae5iKps5ktUdD3vQK4zb01DjhL0qTMqW6uXl8x0mpT5LWTpQ0w5LyrmTxlNss/+Iie+57CFcNO5bSTDk1E20uj/kwKqmpLJWlD4F1gWzP7yO07HzjbFXkUuNTMTFITcA1wKIEt1R1m9tt8dXtbKk9SqHTKmlpvEFFLSicSW6pdd7bnn/y/UGXbdN81+bZUzkh082b7bgZuzlF2FXCB2zweT0tBQhGN2Eo6lCB/NAK/N7PSFmF2+ClrHo+n+kQwcKFgxaJbgUOABcBUSY+b2aslN6fi1ng8Hk8hpKgGN/YC/mlmb5rZV8D/AaUvOILv8Xk8njiIZnCjK8GYQYYFwN7lVJTYxDdj1pxlrTv3fLtAkY7AsgokKon32unSrjS+nrW3LrPer5kxa85TrTttF/YmwbaSpmW9HunW3YbglrfmlDU6m9jEZ2adCr0vaVolo0OVxHvtdGlXGp9k7TCYWbh7dIqzgLUzvQC6AQvLqchf4/N4PElhKtBL0jZuRth/AI+XU1Fie3wejydduIkN5wBPEdzOcpeZvVJOXS058Y0sXqRq8V47XdqVxidZO1bM7EngyUrriX1BcY/H46k1/hqfx+NJHYlNfJLukrRE0pysfR0kjZc0zz3mXZcvT/zPJL2XZZd/eJ7Y7pImSpor6RU3/xhJ10p6zRmpjpHUvoTYvpKmON1pkvbKo91W0ouSZrn4n7v9J7jXayTlHKUrEHula/NMSeMkbVlKvHvvXEmvu/3X5PnYQ9VVJK5R0kuSnnCvH8j6vuZLmllOPSHKz1dgkjszc7tFmO+7UHzWexdJMklFb/koVE+e8jVZAqLuMbNEbsD+wO7AnKx91xDY2gMMB35dYvzPgItCaHcBdnfPNwbeAHYCBgFNbv+vc+kXiB0HHOb2Hw5MyqMtoJ173gp4AdgH2JHA3XoS0K/E2E2yypxH4ItYSvyBwNNAG/de5xCfYc66QsRdANwHPJHjveuBn4T8/eStJ0/5+UDHZvuKft+F4t3+7gQX69/O9X7YekIeQyPwb4J786YCA93+04Ery6kzqVtie3xmNhloblt1NDDKPR8FHAO5be3zxOekeTxwvJnNcO1YAcwFuprZOHNGqsAUgvuM1okn+JH3bx5LcCNm5q/uprj7k3Jon2Nmn7hyrdxmZjbXzF4v1HZgFvCDHLEfZ4Vs5NoSWhv4ITDCzL50x7Uk3+eeEbGA9eqS1FPS064nOEPSdlnH0g34LsG6Lc2PU8CJOAu0Qtq56imkm48w33dz7RzcCFxM1o24JcaX0vbYloCoe2qdeSvZgB6s22P7sNn7H7jHHwKPsPavc4c88T8j+Is6G7gL2KxQfFYd75DVa3L7/wR8P4T+OwQJb0f3/F3gPWDrfLEEf7lnAp/QrJdBsx5fjviOuWKBXzrtOUCnUrTd658T9Nr+BuxZ7HNzr3PV9QJwrHveFtgwq/zDBAtXHUCznhpBD35agePuUKieQrpu31vADGA6MDTHb7Hg950rnsCZ/Gb3fD6uJ1ek7bnqKdj2rNi7CP54AfwDONo9vwBYEeY7aylbzRtQUePDJ75HgENCxG/h/jM2ECSCu4rEt3M/wOOa7b+cYAU55YtvHgv8Bhjsnp8IPF1I273XHpgI7JK1bxLrJr58bV8v1u2/FPh5KdoEyfI3BKeve7n/nCoUn6eu3sCCPGWOAP7XPT+A9RPf74ALQxz3evUQXHLIqZsVt6V77EzQc96/xO97vXiChLWp2z+ftYmv0OfevJ6BxdruyrcmmJq2hXv9TYLLK9OBnwLLi2m3pC2xp7p5WCypC4B7zCwqUNTWHsDMFpvZajNbA9xB8J84Z7ykVgQ/ktFm9mjW/iEE/7lOMfdLah6fJ3YIgTErBOuT5NXOau+HBImu0JSgnPEFYu/DnfaUoL0AeNQCXgTWEPQsw37umbqOLVCsP3CUpPkErhwHSboXQIGJ7XHAA1nl82mvVw9rL48UauNC97iEIMnt5bSLft954gcSLK41y7WlGzBD0jcKtD1fPWGIbAmIlkBLS3yPEyQQ3ONj7nnG1r4JgtHfXMGZpOk4lqAnky/+TmCumd2QFX8ocAlwlAUmrDSPd9ei7mkeS3BNL/MjPgiYl0e7V2b0UNIGBIs0vVbgM8nW7iSpR/NYSb2yyh+VVV9Y7bGuzUjanrW9i7yfu2tL87pmAAskHeP2t1Hg4o2ZXWpm3cysB8FUpWfM7Puuum8Dr5nZglzHna2dp57j8um61xtJ2jjznGBQY06Y79vFdMsRP9XMOptZD9eWBQSDXv/O1/Y87XixUNuzWG8JCPeYbwmIgv9XEk+tu5zlbgRf4iICm/oFwBkEbs8TCJLGBNZeS2sCbgBeJTg9OCdP/D3AywTX+B4HuuSJv4Hgr+JsgmtUMwlGYv9JcJ0ss++2HPH/zBM7gOC0YxbBKdAeebR/Bbzk4ufgRjEJEvUC4EtgMfBUjvjXXfuaxz7iXs8muFbVtUTt1sC9bt8M4KB8n3vW97drnrp6Ac+4/dMJli1o/t0fQNapLnA3cFazMnm1c9VTSBfY1tUxC3gFuNztD/N9zyI4lVwvvllb5rP2VDdn2wu0o+BnBmwILMedVrt95xPcUfAGMIK1p+lFP7eWsPmZGx6PJ3W0tFNdj8fjKYpPfB6PJ3X4xOfxeFKHT3wejyd1+MTn8XhSh098LRhJqxU4csyR9FCe+7vC1nW3pOPd899L2qlA2QMkfasMjfnK4VCSb3+zMp8Uej9H+Z9JuqjUNnpaBj7xtWw+N7O+ZrYL8BVwVvabChZoLhkz+y8rvIjzAUDJic/jiQuf+NLDs0BP1xubKOk+4GXnxnGtpKnOjeNMCNxOJP1W0quS/kwwNxT33iQ5zz9JhzpHkFmSJriZIWcBP3K9zf3cLI1HnMZUSf1d7OYK/P9eknQ7uZcPXAdJYyVNV+DhN7TZe9e7tkyQ1Mnt207SX13Ms5K+Gcmn6Uk0LXnNDY/DTT86DPir27UXgTnBWy55fGRme0pqA/xd0jhgNwLrot4E5g2vErh7ZNfbiWBO8/6urg5m9r6k24BPzOw6V+4+4EYze07SVgTWXDsSzGh4zsx+Iem7wDqJLA+nO40NgKmSHjGz5QR2WjPM7EJJP3F1n0OwpsRZZjZP0t7A/+Km13nSi098LZsNtNaR+FmC+cXfAl40s7fc/kHArpnrdwRegL0I3EPuN7PVwEJJz+Sofx9gcqYuM8vnb/htYCfp6w7dJm7O6f4E5gKY2Z8lfRDimM6TlDEz6O7aupzAGCFjUnAv8Kikdu54H8rSbhNCw9PC8YmvZfO5mfXN3uESwKfZu4BzzeypZuUOp7hLR1gnjwZgXzP7PEdbQs+ZlHQAQRLd18w+kzSJwH8uF+Z0P2z+GXg8/hqf5ynghwqsspC0vXP+mAz8h7sG2IXAXr45zwMDJW3jYjNOHisIPO4yjCM47cSV6+ueTgZOcfsOA/KukeLYlMBj8TN3rW6frPcagEyv9XsEp9AfA29JOsFpSFKfIhqeFOATn+f3BNfvZihYeOl2gjOBMQQuNy8TmHz+rXmgmS0luC73qKRZrD3V/BNwbGZwg2Adj35u8ORV1o4u/xzYX9IMglPud4q09a9AkwIb/CsJ7N4zfArsLGk6wTW8X7j9pwBnuPa9QrA8gSfleHcWj8eTOnyPz+PxpA6f+DweT+rwic/j8aQOn/g8Hk/q8InP4/GkDp/4PB5P6vCJz+PxpA6f+DweT+r4/y0zdQBe4HUgAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Do a grid search\n",
    "clf = KNeighborsClassifier()\n",
    "grid_values = {'n_neighbors':[6,8,10,12,14,16,18,20],\n",
    "               'leaf_size':list(range(1,50,5)),\n",
    "               \"metric\":[\"euclidean\", \"manhattan\", \"chebyshev\"],\n",
    "               'weights':['uniform', 'distance'],\n",
    "               'algorithm':['auto', 'ball_tree', 'kd_tree', 'brute']\n",
    "              }\n",
    "grid_clf_acc = GridSearchCV(clf, param_grid = grid_values,scoring = 'accuracy', n_jobs=-1)\n",
    "grid_clf_acc.fit(X_train, y_train)\n",
    "evaluate([KNeighborsClassifier(**grid_clf_acc.best_params_)], X, y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Proceeding with KNN"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [],
   "source": [
    "score = cross_val_score(KNeighborsClassifier(**grid_clf_acc.best_params_), X=X, y=y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.8353 +/- 0.02487\n"
     ]
    }
   ],
   "source": [
    "print(\"{:.4} +/- {:.4}\".format(np.mean(score), np.std(score)))"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}