main.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Strucanti"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Before doing anything, we want to create a variable indicating the initial working directory of the jupyter notebook file."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "/home/nizarmah/Documents/Learning/Capstone/strucanti\n"
     ]
    }
   ],
   "source": [
    "WORKDIR = !pwd\n",
    "WORKDIR = WORKDIR[0]\n",
    "\n",
    "%cd $WORKDIR"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Importing Libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import sys\n",
    "import os, shutil\n",
    "\n",
    "import re\n",
    "\n",
    "import time\n",
    "import subprocess\n",
    "\n",
    "from functools import partial\n",
    "from multiprocessing import Pool\n",
    "\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "\n",
    "import seaborn as sns\n",
    "\n",
    "import matplotlib as mpl\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "\n",
    "from sklearn.preprocessing import LabelEncoder\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "\n",
    "from sklearn.metrics import mean_absolute_error, accuracy_score\n",
    "from sklearn.metrics import confusion_matrix\n",
    "\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.model_selection import cross_val_score, cross_val_predict\n",
    "from sklearn.model_selection import StratifiedKFold\n",
    "\n",
    "from sklearn.decomposition import PCA"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In addition, we need to make some changes to the libraries, such as increasing *dpi* for `matplotlib`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "mpl.rcParams['figure.dpi']= 300"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Most of pandas functions require a `RandomState`. To keep it consistent for anyone who runs the jupyter notebook to get the same results, we will be setting a global `RANDOM_STATE` variable."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "RANDOM_STATE = 42"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Reading Data\n",
    "\n",
    "We have two different datasets, one `bib-40` and the other is `bib-95`.\n",
    "\n",
    "Therefore, we're going to want to make it easy to deal with both."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(29556, 482)"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "list_dataset_filename = ['dataset_astral-scopedom-seqres-gd-sel-gs-bib-40-2.0.7.csv',\n",
    "                            'dataset_astral-scopedom-seqres-gd-sel-gs-bib-95-2.0.7.csv']\n",
    "\n",
    "dataset_filename = list_dataset_filename[1]\n",
    "dataset_name = os.path.splitext(dataset_filename)[0]\n",
    "\n",
    "dataset = pd.read_csv(dataset_filename)\n",
    "dataset.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let us take a small look at our dataset."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>sid</th>\n",
       "      <th>class</th>\n",
       "      <th>folds</th>\n",
       "      <th>superfamilies</th>\n",
       "      <th>families</th>\n",
       "      <th>sequence</th>\n",
       "      <th>3_h</th>\n",
       "      <th>5_e</th>\n",
       "      <th>7_E</th>\n",
       "      <th>9_h</th>\n",
       "      <th>...</th>\n",
       "      <th>4645_H</th>\n",
       "      <th>4657_H</th>\n",
       "      <th>4677_h</th>\n",
       "      <th>4881_H</th>\n",
       "      <th>4899_H</th>\n",
       "      <th>4913_H</th>\n",
       "      <th>5257_h</th>\n",
       "      <th>5273_H</th>\n",
       "      <th>6417_H</th>\n",
       "      <th>6419_H</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>d1ux8a_</td>\n",
       "      <td>a</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>napyeaigeellsqlvdtfyervashpllkpifpsdltetarkqkq...</td>\n",
       "      <td>4.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>d1dlwa_</td>\n",
       "      <td>a</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>slfeqlggqaavqavtaqfyaniqadatvatffngidmpnqtnkta...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>d1dlya_</td>\n",
       "      <td>a</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>slfaklggreaveaavdkfynkivadptvstyfsntdmkvqrskqf...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>d2gkma_</td>\n",
       "      <td>a</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>gllsrlrkrepisiydkiggheaievvvedffvrvladdqlsaffs...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>d1ngka_</td>\n",
       "      <td>a</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>ksfydavggaktfdaivsrfyaqvaedevlrrvypeddlagaeerl...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</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>29551</th>\n",
       "      <td>d5lj3n2</td>\n",
       "      <td>g</td>\n",
       "      <td>100</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>nepppniceqclgdeanirmtkipqgseckictlpftlyhfktskr...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29552</th>\n",
       "      <td>d2k0aa1</td>\n",
       "      <td>g</td>\n",
       "      <td>101</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>srhqfdlimclkqpgvqtgllcekcdgkcpicdsyvrpkrkvrvce...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29553</th>\n",
       "      <td>d5lqwy_</td>\n",
       "      <td>g</td>\n",
       "      <td>101</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>fdlimclkqpgvqtgllcekcdgkcpicdsyvrpkrkvrvcencsf...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29554</th>\n",
       "      <td>d5syba_</td>\n",
       "      <td>g</td>\n",
       "      <td>101</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>akhhpdlifcrkqagvaigrlcekcdgkcvicdsyvrpstlvricd...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29555</th>\n",
       "      <td>d5ifed_</td>\n",
       "      <td>g</td>\n",
       "      <td>101</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>dlifcrkqagvaigrlcekcdgkcvicdsyvrpctlvricdecnyg...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>29556 rows × 482 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "           sid class  folds  superfamilies  families  \\\n",
       "0      d1ux8a_     a      1              1         1   \n",
       "1      d1dlwa_     a      1              1         1   \n",
       "2      d1dlya_     a      1              1         1   \n",
       "3      d2gkma_     a      1              1         1   \n",
       "4      d1ngka_     a      1              1         1   \n",
       "...        ...   ...    ...            ...       ...   \n",
       "29551  d5lj3n2     g    100              1         1   \n",
       "29552  d2k0aa1     g    101              1         1   \n",
       "29553  d5lqwy_     g    101              1         1   \n",
       "29554  d5syba_     g    101              1         1   \n",
       "29555  d5ifed_     g    101              1         1   \n",
       "\n",
       "                                                sequence  3_h  5_e  7_E  9_h  \\\n",
       "0      napyeaigeellsqlvdtfyervashpllkpifpsdltetarkqkq...  4.0  0.0  0.0  0.0   \n",
       "1      slfeqlggqaavqavtaqfyaniqadatvatffngidmpnqtnkta...  0.0  0.0  0.0  0.0   \n",
       "2      slfaklggreaveaavdkfynkivadptvstyfsntdmkvqrskqf...  1.0  1.0  0.0  0.0   \n",
       "3      gllsrlrkrepisiydkiggheaievvvedffvrvladdqlsaffs...  0.0  0.0  0.0  2.0   \n",
       "4      ksfydavggaktfdaivsrfyaqvaedevlrrvypeddlagaeerl...  1.0  1.0  0.0  0.0   \n",
       "...                                                  ...  ...  ...  ...  ...   \n",
       "29551  nepppniceqclgdeanirmtkipqgseckictlpftlyhfktskr...  1.0  0.0  0.0  1.0   \n",
       "29552  srhqfdlimclkqpgvqtgllcekcdgkcpicdsyvrpkrkvrvce...  0.0  1.0  0.0  0.0   \n",
       "29553  fdlimclkqpgvqtgllcekcdgkcpicdsyvrpkrkvrvcencsf...  0.0  1.0  0.0  0.0   \n",
       "29554  akhhpdlifcrkqagvaigrlcekcdgkcvicdsyvrpstlvricd...  1.0  0.0  1.0  1.0   \n",
       "29555  dlifcrkqagvaigrlcekcdgkcvicdsyvrpctlvricdecnyg...  1.0  0.0  0.0  1.0   \n",
       "\n",
       "       ...  4645_H  4657_H  4677_h  4881_H  4899_H  4913_H  5257_h  5273_H  \\\n",
       "0      ...     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0   \n",
       "1      ...     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0   \n",
       "2      ...     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0   \n",
       "3      ...     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0   \n",
       "4      ...     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0   \n",
       "...    ...     ...     ...     ...     ...     ...     ...     ...     ...   \n",
       "29551  ...     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0   \n",
       "29552  ...     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0   \n",
       "29553  ...     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0   \n",
       "29554  ...     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0   \n",
       "29555  ...     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0   \n",
       "\n",
       "       6417_H  6419_H  \n",
       "0         0.0     0.0  \n",
       "1         0.0     0.0  \n",
       "2         0.0     0.0  \n",
       "3         0.0     0.0  \n",
       "4         0.0     0.0  \n",
       "...       ...     ...  \n",
       "29551     0.0     0.0  \n",
       "29552     0.0     0.0  \n",
       "29553     0.0     0.0  \n",
       "29554     0.0     0.0  \n",
       "29555     0.0     0.0  \n",
       "\n",
       "[29556 rows x 482 columns]"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dataset"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As you can see, we have a **sparse dataset**"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "If we further take a look at the dataset, we can see that the dataset is sorted by class, folds, superfamilies, then families. However, this can influence our model's accuracy and variance when predicting our data as the train and test splits will not contain a balanced distribution of the classes, folds, superfamilies and families.\n",
    "\n",
    "Therefore, it is important that we shuffle our dataset before we start working with it.\n",
    "\n",
    "> Please Note that we will be resetting the index after sampling"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "dataset = dataset.sample(frac=1, random_state=RANDOM_STATE).reset_index(drop=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "So, let's take another look now at our dataset."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>sid</th>\n",
       "      <th>class</th>\n",
       "      <th>folds</th>\n",
       "      <th>superfamilies</th>\n",
       "      <th>families</th>\n",
       "      <th>sequence</th>\n",
       "      <th>3_h</th>\n",
       "      <th>5_e</th>\n",
       "      <th>7_E</th>\n",
       "      <th>9_h</th>\n",
       "      <th>...</th>\n",
       "      <th>4645_H</th>\n",
       "      <th>4657_H</th>\n",
       "      <th>4677_h</th>\n",
       "      <th>4881_H</th>\n",
       "      <th>4899_H</th>\n",
       "      <th>4913_H</th>\n",
       "      <th>5257_h</th>\n",
       "      <th>5273_H</th>\n",
       "      <th>6417_H</th>\n",
       "      <th>6419_H</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>d1m9sa3</td>\n",
       "      <td>b</td>\n",
       "      <td>34</td>\n",
       "      <td>11</td>\n",
       "      <td>1</td>\n",
       "      <td>rydkmeydkgvtayarvrnasgnsvwtkpyntagakhvnklsvyqg...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>d2gpca2</td>\n",
       "      <td>d</td>\n",
       "      <td>44</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>ngggeptgkvadeinasfgsfakfkeeftnvavghfgsgwawlvkd...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>d1ekxa2</td>\n",
       "      <td>c</td>\n",
       "      <td>78</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>rldnlhvamvgdlkygrtvhsltqalakfdgnrfyfiapdalampq...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>d4ogna_</td>\n",
       "      <td>a</td>\n",
       "      <td>42</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>dgavttsqipaseqetlvrpkplllkllksvgaqkdtytmkevlfy...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>d5awxa1</td>\n",
       "      <td>c</td>\n",
       "      <td>45</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>aipikhfpkhvadlhassgfteefeevqsctvdlgitadssnhpdn...</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</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>29551</th>\n",
       "      <td>d4m1na_</td>\n",
       "      <td>d</td>\n",
       "      <td>20</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>siakkrlaqeraewrkdhpagfsakyspmsdgkgldimkwickipg...</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29552</th>\n",
       "      <td>d1bafl1</td>\n",
       "      <td>b</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>qivltqspaimsaspgekvtmtcsasssvyymywyqqkpgssprll...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29553</th>\n",
       "      <td>d1kyza1</td>\n",
       "      <td>a</td>\n",
       "      <td>4</td>\n",
       "      <td>5</td>\n",
       "      <td>29</td>\n",
       "      <td>hisdeeanlfamqlasasvlpmilksaleldlleiiakagpgaqis...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29554</th>\n",
       "      <td>d4cr2i3</td>\n",
       "      <td>c</td>\n",
       "      <td>37</td>\n",
       "      <td>1</td>\n",
       "      <td>20</td>\n",
       "      <td>lsigtleeiidddhaivtsptmpdyyvsilsfvdkellepgcsvll...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29555</th>\n",
       "      <td>d1hdgo2</td>\n",
       "      <td>d</td>\n",
       "      <td>81</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>cttnsiapivkvlhekfgivsgmlttvhsytndqrvldlphkdlrr...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>29556 rows × 482 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "           sid class  folds  superfamilies  families  \\\n",
       "0      d1m9sa3     b     34             11         1   \n",
       "1      d2gpca2     d     44              1         0   \n",
       "2      d1ekxa2     c     78              1         1   \n",
       "3      d4ogna_     a     42              1         1   \n",
       "4      d5awxa1     c     45              1         0   \n",
       "...        ...   ...    ...            ...       ...   \n",
       "29551  d4m1na_     d     20              1         1   \n",
       "29552  d1bafl1     b      1              1         1   \n",
       "29553  d1kyza1     a      4              5        29   \n",
       "29554  d4cr2i3     c     37              1        20   \n",
       "29555  d1hdgo2     d     81              1         1   \n",
       "\n",
       "                                                sequence  3_h  5_e  7_E  9_h  \\\n",
       "0      rydkmeydkgvtayarvrnasgnsvwtkpyntagakhvnklsvyqg...  1.0  0.0  0.0  0.0   \n",
       "1      ngggeptgkvadeinasfgsfakfkeeftnvavghfgsgwawlvkd...  0.0  0.0  0.0  0.0   \n",
       "2      rldnlhvamvgdlkygrtvhsltqalakfdgnrfyfiapdalampq...  1.0  1.0  1.0  0.0   \n",
       "3      dgavttsqipaseqetlvrpkplllkllksvgaqkdtytmkevlfy...  1.0  1.0  1.0  0.0   \n",
       "4      aipikhfpkhvadlhassgfteefeevqsctvdlgitadssnhpdn...  2.0  2.0  2.0  2.0   \n",
       "...                                                  ...  ...  ...  ...  ...   \n",
       "29551  siakkrlaqeraewrkdhpagfsakyspmsdgkgldimkwickipg...  2.0  1.0  0.0  0.0   \n",
       "29552  qivltqspaimsaspgekvtmtcsasssvyymywyqqkpgssprll...  1.0  2.0  0.0  0.0   \n",
       "29553  hisdeeanlfamqlasasvlpmilksaleldlleiiakagpgaqis...  1.0  0.0  0.0  0.0   \n",
       "29554  lsigtleeiidddhaivtsptmpdyyvsilsfvdkellepgcsvll...  1.0  0.0  1.0  0.0   \n",
       "29555  cttnsiapivkvlhekfgivsgmlttvhsytndqrvldlphkdlrr...  0.0  0.0  0.0  0.0   \n",
       "\n",
       "       ...  4645_H  4657_H  4677_h  4881_H  4899_H  4913_H  5257_h  5273_H  \\\n",
       "0      ...     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0   \n",
       "1      ...     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0   \n",
       "2      ...     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0   \n",
       "3      ...     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0   \n",
       "4      ...     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0   \n",
       "...    ...     ...     ...     ...     ...     ...     ...     ...     ...   \n",
       "29551  ...     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0   \n",
       "29552  ...     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0   \n",
       "29553  ...     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0   \n",
       "29554  ...     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0   \n",
       "29555  ...     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0   \n",
       "\n",
       "       6417_H  6419_H  \n",
       "0         0.0     0.0  \n",
       "1         0.0     0.0  \n",
       "2         0.0     0.0  \n",
       "3         0.0     0.0  \n",
       "4         0.0     0.0  \n",
       "...       ...     ...  \n",
       "29551     0.0     0.0  \n",
       "29552     0.0     0.0  \n",
       "29553     0.0     0.0  \n",
       "29554     0.0     0.0  \n",
       "29555     0.0     0.0  \n",
       "\n",
       "[29556 rows x 482 columns]"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dataset"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Checking for Missing Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of columns that contain null values : 0\n",
      "Names of columns that contain null values : Series([], dtype: bool)\n"
     ]
    }
   ],
   "source": [
    "dataset_nullcheck = dataset.isnull().any()\n",
    "\n",
    "null_columns = dataset_nullcheck[dataset_nullcheck == True]\n",
    "\n",
    "print(\"Number of columns that contain null values :\", len(null_columns))\n",
    "print(\"Names of columns that contain null values :\", null_columns)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As you can see, there are no _column names_ that contain any missing data.\n",
    "\n",
    "This shows that the columns we have don't have any missing data."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Checking for Categorical Values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1800x1200 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "categorical_cols = [ col for col in dataset.columns if dataset[col].dtype == 'object' ]\n",
    "\n",
    "sns.barplot(x=categorical_cols, y=dataset[categorical_cols].nunique())\n",
    "\n",
    "plt.title('Unique Values in Categorical Columns')\n",
    "plt.xlabel('Column Name')\n",
    "plt.ylabel('Unique Values')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As you can tell, `sid` and `sequence` columns are as frequent as the entries that we have. In addition, they are the sequence id, and the sequence itself.\n",
    "\n",
    "Therefore, we will disregard them.\n",
    "\n",
    "So, we have only one categorical column, which is the target or the label `class`.\n",
    "\n",
    "So, let us check what are the _unique_ values we have in `class`. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['b', 'd', 'c', 'a', 'e', 'f', 'g'], dtype=object)"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dataset['class'].unique()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Here are the percentage frequencies of each class."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>c</th>\n",
       "      <th>b</th>\n",
       "      <th>d</th>\n",
       "      <th>a</th>\n",
       "      <th>g</th>\n",
       "      <th>e</th>\n",
       "      <th>f</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Percentage Occurance</th>\n",
       "      <td>26.904859</td>\n",
       "      <td>25.216538</td>\n",
       "      <td>22.327108</td>\n",
       "      <td>16.703884</td>\n",
       "      <td>5.484504</td>\n",
       "      <td>1.752605</td>\n",
       "      <td>1.610502</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                              c          b          d          a         g  \\\n",
       "Percentage Occurance  26.904859  25.216538  22.327108  16.703884  5.484504   \n",
       "\n",
       "                             e         f  \n",
       "Percentage Occurance  1.752605  1.610502  "
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def class_percentage_occurance():\n",
    "    return (dataset['class'].value_counts() * 100 / dataset['class'].size).to_frame(name='Percentage Occurance').T\n",
    "\n",
    "class_percentage_occurance()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "It is also important to see the frequency of each class."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1800x1200 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.barplot(data=class_percentage_occurance())\n",
    "\n",
    "plt.title('Value Counts of Dataset Classes')\n",
    "plt.xlabel('Class')\n",
    "plt.ylabel('Count')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Disregarding Some Columns\n",
    "\n",
    "We will be disregarding some columns at this point.\n",
    "\n",
    "The following categorical columns are not important in our machine learning model:\n",
    "- `sid`\n",
    "- `sequence`"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The following columns are going to be disregarded for now:\n",
    "- `folds`\n",
    "- `superfamilies`\n",
    "- `families`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(29556, 477)"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "unwanted_cols = ['sid', 'sequence', 'folds', 'superfamilies', 'families']\n",
    "\n",
    "dataset = dataset.drop(unwanted_cols, axis=1)\n",
    "dataset.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>class</th>\n",
       "      <th>3_h</th>\n",
       "      <th>5_e</th>\n",
       "      <th>7_E</th>\n",
       "      <th>9_h</th>\n",
       "      <th>11_E</th>\n",
       "      <th>13_e</th>\n",
       "      <th>15_E</th>\n",
       "      <th>17_h</th>\n",
       "      <th>19_H</th>\n",
       "      <th>...</th>\n",
       "      <th>4645_H</th>\n",
       "      <th>4657_H</th>\n",
       "      <th>4677_h</th>\n",
       "      <th>4881_H</th>\n",
       "      <th>4899_H</th>\n",
       "      <th>4913_H</th>\n",
       "      <th>5257_h</th>\n",
       "      <th>5273_H</th>\n",
       "      <th>6417_H</th>\n",
       "      <th>6419_H</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>b</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>d</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>c</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>a</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>c</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>a</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>c</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>a</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>b</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>d</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>10 rows × 477 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "  class  3_h  5_e  7_E  9_h  11_E  13_e  15_E  17_h  19_H  ...  4645_H  \\\n",
       "0     b  1.0  0.0  0.0  0.0   1.0   0.0   0.0   0.0   0.0  ...     0.0   \n",
       "1     d  0.0  0.0  0.0  0.0   1.0   0.0   0.0   0.0   0.0  ...     0.0   \n",
       "2     c  1.0  1.0  1.0  0.0   1.0   0.0   1.0   0.0   0.0  ...     0.0   \n",
       "3     a  1.0  1.0  1.0  0.0   0.0   0.0   0.0   0.0   0.0  ...     0.0   \n",
       "4     c  2.0  2.0  2.0  2.0   0.0   0.0   0.0   0.0   0.0  ...     0.0   \n",
       "5     a  0.0  0.0  0.0  2.0   0.0   0.0   0.0   0.0   0.0  ...     0.0   \n",
       "6     c  0.0  1.0  0.0  0.0   0.0   0.0   0.0   0.0   0.0  ...     0.0   \n",
       "7     a  1.0  0.0  0.0  2.0   1.0   0.0   0.0   0.0   0.0  ...     0.0   \n",
       "8     b  0.0  1.0  1.0  0.0   0.0   1.0   0.0   0.0   0.0  ...     0.0   \n",
       "9     d  1.0  0.0  1.0  0.0   0.0   0.0   0.0   0.0   0.0  ...     0.0   \n",
       "\n",
       "   4657_H  4677_h  4881_H  4899_H  4913_H  5257_h  5273_H  6417_H  6419_H  \n",
       "0     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0  \n",
       "1     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0  \n",
       "2     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0  \n",
       "3     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0  \n",
       "4     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0  \n",
       "5     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0  \n",
       "6     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0  \n",
       "7     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0  \n",
       "8     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0  \n",
       "9     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0  \n",
       "\n",
       "[10 rows x 477 columns]"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dataset.head(10)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Disregarding Some Classes"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "According to the HCADB we are using, the clusters were collected from the following four classes:\n",
    "- `a`\n",
    "- `b`\n",
    "- `c`\n",
    "- `d`\n",
    "\n",
    "Therefore, we will be keeping those classes and disregarding the rest"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "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>c</th>\n",
       "      <th>b</th>\n",
       "      <th>d</th>\n",
       "      <th>a</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Percentage Occurance</th>\n",
       "      <td>29.516351</td>\n",
       "      <td>27.664155</td>\n",
       "      <td>24.494265</td>\n",
       "      <td>18.325229</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                              c          b          d          a\n",
       "Percentage Occurance  29.516351  27.664155  24.494265  18.325229"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def trim_dataset(dataset, classes):\n",
    "    \"\"\"\n",
    "    Reduces the dataset by keeping the classes specified, and dropping the rest\n",
    "    \"\"\"\n",
    "    # using .loc would later on avoid the SettingWithCopyWarning that would appear with LabelEncoding\n",
    "    # basically it helps to avoids returning a copy of the DataFrame\n",
    "    return dataset[dataset['class'].isin(classes)].loc[:, dataset.columns]\n",
    "\n",
    "wanted_classes = ['a', 'b', 'c', 'd']\n",
    "\n",
    "dataset = trim_dataset(dataset, wanted_classes)\n",
    "\n",
    "class_percentage_occurance()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1800x1200 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.barplot(data=class_percentage_occurance())\n",
    "\n",
    "plt.title('Value Counts of Dataset Classes')\n",
    "plt.xlabel('Class')\n",
    "plt.ylabel('Count')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "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>class</th>\n",
       "      <th>3_h</th>\n",
       "      <th>5_e</th>\n",
       "      <th>7_E</th>\n",
       "      <th>9_h</th>\n",
       "      <th>11_E</th>\n",
       "      <th>13_e</th>\n",
       "      <th>15_E</th>\n",
       "      <th>17_h</th>\n",
       "      <th>19_H</th>\n",
       "      <th>...</th>\n",
       "      <th>4645_H</th>\n",
       "      <th>4657_H</th>\n",
       "      <th>4677_h</th>\n",
       "      <th>4881_H</th>\n",
       "      <th>4899_H</th>\n",
       "      <th>4913_H</th>\n",
       "      <th>5257_h</th>\n",
       "      <th>5273_H</th>\n",
       "      <th>6417_H</th>\n",
       "      <th>6419_H</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>b</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>d</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>c</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>a</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>c</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</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>29551</th>\n",
       "      <td>d</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29552</th>\n",
       "      <td>b</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29553</th>\n",
       "      <td>a</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29554</th>\n",
       "      <td>c</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29555</th>\n",
       "      <td>d</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>26941 rows × 477 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      class  3_h  5_e  7_E  9_h  11_E  13_e  15_E  17_h  19_H  ...  4645_H  \\\n",
       "0         b  1.0  0.0  0.0  0.0   1.0   0.0   0.0   0.0   0.0  ...     0.0   \n",
       "1         d  0.0  0.0  0.0  0.0   1.0   0.0   0.0   0.0   0.0  ...     0.0   \n",
       "2         c  1.0  1.0  1.0  0.0   1.0   0.0   1.0   0.0   0.0  ...     0.0   \n",
       "3         a  1.0  1.0  1.0  0.0   0.0   0.0   0.0   0.0   0.0  ...     0.0   \n",
       "4         c  2.0  2.0  2.0  2.0   0.0   0.0   0.0   0.0   0.0  ...     0.0   \n",
       "...     ...  ...  ...  ...  ...   ...   ...   ...   ...   ...  ...     ...   \n",
       "29551     d  2.0  1.0  0.0  0.0   0.0   0.0   1.0   2.0   0.0  ...     0.0   \n",
       "29552     b  1.0  2.0  0.0  0.0   0.0   0.0   1.0   1.0   0.0  ...     0.0   \n",
       "29553     a  1.0  0.0  0.0  0.0   0.0   0.0   0.0   0.0   0.0  ...     0.0   \n",
       "29554     c  1.0  0.0  1.0  0.0   0.0   0.0   0.0   0.0   0.0  ...     0.0   \n",
       "29555     d  0.0  0.0  0.0  0.0   1.0   1.0   0.0   0.0   0.0  ...     0.0   \n",
       "\n",
       "       4657_H  4677_h  4881_H  4899_H  4913_H  5257_h  5273_H  6417_H  6419_H  \n",
       "0         0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0  \n",
       "1         0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0  \n",
       "2         0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0  \n",
       "3         0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0  \n",
       "4         0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0  \n",
       "...       ...     ...     ...     ...     ...     ...     ...     ...     ...  \n",
       "29551     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0  \n",
       "29552     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0  \n",
       "29553     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0  \n",
       "29554     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0  \n",
       "29555     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0  \n",
       "\n",
       "[26941 rows x 477 columns]"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dataset"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Running mRMR on Dataset"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> If you already have the `mRMR/queries` necessary, don't run the mRMR feature selection "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "`Python` libraries for `mRMR` are discontinued and unreliable. \n",
    "In addition, after comparing a `C/C++` library for `mRMR` with the one used in `R`, I decided to write my own script to execute `mRMR` on a dataset using `R`.\n",
    "\n",
    "The R script that will be used is in `mRMR/mRMR.R`.\n",
    "It takes two arguments / parameters:\n",
    "1. `dataset_path`\n",
    "2. `number_features`\n",
    "\n",
    "We will be using `subprocesses` to run the `Rscript`. In addition, we will be piping its `stdout` to a directory."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We will be running `mRMR` with `number_features` from 1 to 476.\n",
    "\n",
    "In addition, we will be running `mRMR` including the following *classes*:\n",
    "- Against All (such as: a, not a)\n",
    "- Pairwise (such as: a, b)\n",
    "- All Classes (a, b, c, d)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In order to ease the creation of the dataset for the above `mRMR` runs, we will create a function to do that for us."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "def label_encode_dataset(dataset, classes):\n",
    "    \"\"\"\n",
    "    Label Encodes the specified classes in the Dataset\n",
    "    The classes parameter can be `list` or `str`\n",
    "    \n",
    "    If the type(classes) is `list`\n",
    "    The dataset is first reduced to only match the classes specified\n",
    "    Then it is Label Encoded\n",
    "    \n",
    "    If the type(classes) is `str`\n",
    "    We label encode the classes by `class` and `not class` (1, 0 respectively)\n",
    "    \"\"\"\n",
    "    dataset_clone = dataset.copy(deep=True)\n",
    "    \n",
    "    if type(classes) is list:\n",
    "        dataset_clone = trim_dataset(dataset_clone, classes)\n",
    "        \n",
    "        label_encoder = LabelEncoder()\n",
    "        dataset_clone['class'] = label_encoder.fit_transform(dataset_clone['class'])\n",
    "    elif type(classes) is str:\n",
    "        dataset_clone['class'] = dataset_clone['class'].map(lambda x: int(classes == x))\n",
    "    else:\n",
    "        raise Exception('Unsupported Argument Type \\'classes\\'; Should be \\'list\\' or \\'str\\'')\n",
    "    \n",
    "    return dataset_clone"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Also, in order to ease the running of `mRMR` on the different _label encoded_ datasets, we will be creating a function that queries certain classes to have `mRMR` run on them automatically."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "def subquery_mRMR(task):\n",
    "    \"\"\"\n",
    "    Run the mRMR script on the Task\n",
    "    Task is of form (subquery_dir, num_features)\n",
    "    \"\"\"\n",
    "    subquery_dir, num_features = task\n",
    "    \n",
    "    script_path = 'mRMR.R'\n",
    "    \n",
    "    dataset_path = '{0}/dataset.csv'.format(subquery_dir)\n",
    "    output_file = '{0}/mRMR/out/{1}_features.mrmr'.format(subquery_dir, num_features)\n",
    "    \n",
    "    subprocess.call(['Rscript', script_path, dataset_path, str(num_features), output_file])\n",
    "\n",
    "def query_mRMR(dataset, dataset_name, query_label, subqueries, jobs=1):\n",
    "    \"\"\"\n",
    "    Queries mRMR feature selection on dataset\n",
    "    \n",
    "    Script path is mRMR/mRMR.R\n",
    "    Therefore, we save the query results next to it in mRMR/queries/\n",
    "    \n",
    "    However, we might have different datasets used, so we use\n",
    "    mRMR/queries/dataset_name/query_label\n",
    "    In order to save the output \n",
    "    \n",
    "    If the query_label directory exists, we delete it and its content\n",
    "    We create the query_label directory, then go over every subquery,\n",
    "    We get the subquery path, and the number of features we want to execute\n",
    "    throw that all in a tasks pool, then run the pool on mutliple processors\n",
    "    \"\"\"\n",
    "    start_time = time.time()\n",
    "    \n",
    "    query_dir = 'queries/{0}/{1}'.format(dataset_name, query_label)\n",
    "    \n",
    "    if os.path.exists(query_dir):\n",
    "        shutil.rmtree(query_dir)\n",
    "    \n",
    "    os.makedirs(query_dir)\n",
    "    \n",
    "    tasks = [] \n",
    "    for subquery in subqueries:\n",
    "        subquery_label = '_'.join([ str(i) for i in subquery ]) \\\n",
    "                            if type(subquery) is list else \\\n",
    "                                '{0}_!{0}'.format(subquery)\n",
    "        \n",
    "        subquery_dir = '{0}/{1}'.format(query_dir, subquery_label)\n",
    "        os.makedirs('{0}/mRMR/out'.format(subquery_dir))\n",
    "        \n",
    "        dataset_path = '{0}/dataset.csv'.format(subquery_dir)\n",
    "        label_encode_dataset(dataset, subquery).to_csv(dataset_path, index=False)\n",
    "        \n",
    "        for i in range(1, len(dataset.columns)):\n",
    "            tasks.append((subquery_dir, i))\n",
    "            \n",
    "    with Pool(processes=jobs) as pool:\n",
    "        pool.map(subquery_mRMR, tasks)\n",
    "        \n",
    "    print('Time taken :', (time.time() - start_time))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Against All"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "query_mRMR(dataset, dataset_name, 'against_all', [\n",
    "    'a',\n",
    "    'b',\n",
    "    'c',\n",
    "    'd'\n",
    "], jobs=8)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Pairwise"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "query_mRMR(dataset, dataset_name, 'pairwise', [\n",
    "    ['a', 'b'],\n",
    "    ['a', 'c'],\n",
    "    ['a', 'd'],\n",
    "    ['b', 'c'],\n",
    "    ['b', 'd'],\n",
    "    ['c', 'd']\n",
    "], jobs=6)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### All Classes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "query_mRMR(dataset, dataset_name, 'all_classes', [\n",
    "    ['a', 'b', 'c', 'd']\n",
    "], jobs=4)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Scoring mRMR Selected Features"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "After creating different `mRMR` feature selections with different `num_features` taken everytime, we have to score the results of each `mRMR` run for the different runs in order to identify which run performed the best.\n",
    "\n",
    "Then, we'll take that run's `mRMR` selected features to use them in our model later on."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In addition, we want to run that on the different datasets we have and the different `mRMR` runs, whether *pairwise*, *against all*, or *all classes*. Therefore, we will specify which `query_labels` we want to execute the _cross validation_ on."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "def read_mRMR_features(mrmr_file_path):\n",
    "    \"\"\"\n",
    "    Read the output of the mRMR Rscript\n",
    "    \"\"\"\n",
    "    with open(mrmr_file_path, 'r') as file:\n",
    "        return [ line.rstrip() for line in file.readlines() ]\n",
    "        \n",
    "def score_mRMR_features(model, dataset, mrmr_features, folds=10, jobs=-1):\n",
    "    \"\"\"\n",
    "    Running Cross Validation on a Specific Model\n",
    "    Along with Dataset and Specific mRMR Selected Features\n",
    "    \"\"\"\n",
    "    return cross_val_score(model, dataset[mrmr_features], dataset['class'], cv=folds, n_jobs=jobs)\n",
    "\n",
    "def plot_scores(scores, plot_path, prefix='', xlabel='Number of Features', ci=95):\n",
    "    \"\"\"\n",
    "    After we have the *scores* for each *percentage*, \n",
    "    we want to convert it into a DataFrame of (*n# features*, *score*) \n",
    "    and plot the different scores, along with their mean and error bars.\n",
    "    \"\"\"\n",
    "    scores_df = pd.DataFrame([ [key, score] for key in scores.keys() \n",
    "                                                           for score in scores[key] ],\n",
    "                                    columns=['key', 'score'])\n",
    "    \n",
    "    sns.pointplot(x='key', y='score', data=scores_df, capsize=.2, ci=ci)\n",
    "    sns.stripplot(x='key', y='score', data=scores_df)\n",
    "\n",
    "    plot_title = 'Score'\n",
    "    if len(prefix) > 0:\n",
    "        plot_title = '{0} : {1}'.format(prefix, plot_title)\n",
    "    \n",
    "    plt.title(plot_title)\n",
    "    \n",
    "    plt.xlabel(xlabel, size=8)\n",
    "    plt.xticks(size=8)\n",
    "    \n",
    "    plt.ylabel('Cross Validation Score', size=8)\n",
    "    plt.yticks(size=8)\n",
    "    \n",
    "    plt.savefig(plot_path)\n",
    "    plt.clf()\n",
    "\n",
    "def score_mRMR_queries(model, model_name, dataset_name, query_labels, \n",
    "                            runs_range=range(1, 477), folds=10, jobs=-1):\n",
    "    \"\"\"\n",
    "    For the specified queries under the dataset_name\n",
    "    Run Cross Validation Scoring on the mRMR Rscript Results\n",
    "    According to a certain Range / Ratio\n",
    "    \n",
    "    The ratio can be a normal percentage, or it can be something else\n",
    "    like from 1 to 476 with a step of 1\n",
    "    meaning all number_features are included\n",
    "    \"\"\"\n",
    "    dataset_dir = 'queries/{0}'.format(dataset_name)\n",
    "    \n",
    "    for query in os.listdir(dataset_dir):\n",
    "        if query in query_labels:\n",
    "            query_dir = '{0}/{1}'.format(dataset_dir, query)\n",
    "            \n",
    "            for subquery in os.listdir(query_dir):\n",
    "                subquery_scores = dict()\n",
    "                \n",
    "                subquery_dir = '{0}/{1}'.format(query_dir, subquery)\n",
    "                \n",
    "                dataset_file = '{0}/dataset.csv'.format(subquery_dir)\n",
    "                dataset = pd.read_csv(dataset_file)\n",
    "                \n",
    "                for i in runs_range:\n",
    "                    num_features = int((476 * i) / (runs_range.stop - 1))\n",
    "                    \n",
    "                    mrmr_file = '{0}/mRMR/out/{1}_features.mrmr'.format(subquery_dir, num_features)\n",
    "                    mrmr_features = read_mRMR_features(mrmr_file)\n",
    "                    \n",
    "                    subquery_scores[num_features] = score_mRMR_features(model, dataset, \n",
    "                                                        mrmr_features, folds=folds, jobs=jobs)\n",
    "                \n",
    "                plot_scores(subquery_scores, '{0}/mRMR/{1}_plot.png'.format(subquery_dir, model_name),\n",
    "                                    prefix=subquery, xlabel='Number of Features')\n",
    "                \n",
    "                with open('{0}/mRMR/{1}_scores.txt'.format(subquery_dir, model_name), 'w+') as file:\n",
    "                    for key, value in subquery_scores.items():\n",
    "                        print('{0} : {1}'.format(key, ','.join([ str(v) for v in value ])), \n",
    "                                  file=file)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "score_mRMR_queries(RandomForestClassifier(random_state=RANDOM_STATE), \"random_forest\", dataset_name,\n",
    "                    [ 'against_all', 'pairwise', 'all_classes' ], runs_range=range(5, 101, 5), jobs=8)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Scoring PCA Features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "def score_pca_subqueries(model, dataset, variance_range, folds=10):\n",
    "    subquery_scores = dict()\n",
    "    \n",
    "    X = dataset.drop('class', axis=1)\n",
    "    y = dataset['class']\n",
    "\n",
    "    splitter = StratifiedKFold(n_splits=folds)\n",
    "    \n",
    "    scaler = StandardScaler()\n",
    "    \n",
    "    for variance in variance_range:\n",
    "        pca = PCA((variance / 100))\n",
    "        subquery_scores[variance] = dict(scores=[], components=[])\n",
    "        \n",
    "        for train_index, test_index in splitter.split(X, y):\n",
    "            X_train, X_test = X.iloc[train_index, :], X.iloc[test_index, :]\n",
    "            y_train, y_test = y[train_index], y[test_index]\n",
    "            scaler.fit(X_train)\n",
    "            X_train, X_test = scaler.transform(X_train), scaler.transform(X_test)\n",
    "            \n",
    "            pca.fit(X_train)\n",
    "            X_train, X_test = pca.transform(X_train), pca.transform(X_test)\n",
    "\n",
    "            model.fit(X_train, y_train)\n",
    "            score = model.score(X_test, y_test)\n",
    "            \n",
    "            subquery_scores[variance]['scores'].append(score)\n",
    "            subquery_scores[variance]['components'].append(pca.n_components_)\n",
    "            \n",
    "    return subquery_scores\n",
    "\n",
    "def score_pca_queries(model, model_name, dataset_name, query_labels, \n",
    "                          variance_range=range(80, 96, 5), folds=10):\n",
    "    \"\"\"\n",
    "    \"\"\"\n",
    "    dataset_dir = 'queries/{0}'.format(dataset_name)\n",
    "\n",
    "    for query in os.listdir(dataset_dir):\n",
    "        if query in query_labels:\n",
    "            query_dir = '{0}/{1}'.format(dataset_dir, query)\n",
    "\n",
    "            for subquery in os.listdir(query_dir):\n",
    "                subquery_dir = '{0}/{1}'.format(query_dir, subquery)\n",
    "                \n",
    "                pca_dir = '{0}/pca'.format(subquery_dir)\n",
    "                if not os.path.exists(pca_dir):\n",
    "                    os.makedirs(pca_dir)\n",
    "                \n",
    "                dataset_file = '{0}/dataset.csv'.format(subquery_dir)\n",
    "                dataset = pd.read_csv(dataset_file)\n",
    "                \n",
    "                subquery_scores = score_pca_subqueries(model, dataset, variance_range, folds=folds)\n",
    "                \n",
    "                plot_scores({ k: v['scores'] for k, v in subquery_scores.items() }, \n",
    "                                '{0}/{1}_plot.png'.format(pca_dir, model_name),\n",
    "                                    prefix=subquery, xlabel='PCA Variance Retain')\n",
    "\n",
    "                with open('{0}/{1}_scores.txt'.format(pca_dir, model_name), 'w+') as file:\n",
    "                    for key, value in subquery_scores.items():\n",
    "                        print('{0} : {1}'.format(key, ','.join([ str(v) for v in value['scores'] ])), \n",
    "                                  file=file)\n",
    "\n",
    "                with open('{0}/{1}_components.txt'.format(pca_dir, model_name), 'w+') as file:\n",
    "                    for key, value in subquery_scores.items():\n",
    "                        print('{0} : {1}'.format(key, ','.join([ str(v) for v in value['components'] ])), \n",
    "                                  file=file)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "score_pca_queries(RandomForestClassifier(random_state=RANDOM_STATE), \"random_forest\", dataset_name,\n",
    "                    [ 'against_all', 'pairwise', 'all_classes' ], variance_range=range(5, 96, 20))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Scoring Other Models"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.ensemble import GradientBoostingClassifier\n",
    "\n",
    "score_mRMR_queries(GradientBoostingClassifier(random_state=RANDOM_STATE), \"gboost\", dataset_name,\n",
    "                    [ 'against_all', 'pairwise', 'all_classes' ], runs_range=range(5, 101, 5), jobs=3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "from sklearn.svm import LinearSVC\n",
    "\n",
    "score_mRMR_queries(LinearSVC(random_state=RANDOM_STATE), \"svm\", dataset_name,\n",
    "                    [ 'against_all', 'pairwise', 'all_classes' ], runs_range=range(5, 101, 5), jobs=3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.linear_model import SGDClassifier\n",
    "\n",
    "score_mRMR_queries(SGDClassifier(random_state=RANDOM_STATE, n_jobs=-1), \"sgd\", dataset_name,\n",
    "                    [ 'against_all', 'pairwise', 'all_classes' ], runs_range=range(5, 101, 5), jobs=6)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "\n",
    "score_mRMR_queries(KNeighborsClassifier(), \"knn\", dataset_name,\n",
    "                    [ 'against_all', 'pairwise', 'all_classes' ], runs_range=range(5, 101, 5), jobs=3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.naive_bayes import GaussianNB\n",
    "\n",
    "score_mRMR_queries(GaussianNB(), \"bayes\", dataset_name,\n",
    "                    [ 'against_all', 'pairwise', 'all_classes' ], runs_range=range(5, 101, 5), jobs=6)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "strucanti",
   "language": "python",
   "name": "strucanti"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}