Code.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Diagnosis of Cardiac Arrhythmia with PCA And MLP"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Dataset was downloaded  from here : https://archive.ics.uci.edu/ml/datasets/arrhythmia"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-11-18T19:25:52.546748Z",
     "start_time": "2021-11-18T19:25:52.539145Z"
    }
   },
   "outputs": [],
   "source": [
    "import time\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import seaborn as sns\n",
    "from scipy import stats\n",
    "\n",
    "from sklearn import metrics\n",
    "from sklearn.metrics import *\n",
    "from sklearn.preprocessing import OneHotEncoder\n",
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "from sklearn.decomposition import PCA\n",
    "\n",
    "from IPython.display import clear_output    \n",
    "\n",
    "\n",
    "from matplotlib import pyplot as plt\n",
    "from matplotlib import colors\n",
    "from matplotlib.colors import ListedColormap\n",
    "\n",
    "from sklearn.neural_network import MLPClassifier\n",
    "from sklearn.model_selection import cross_val_score\n",
    "from sklearn.preprocessing import OneHotEncoder"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-11-18T19:18:45.469803Z",
     "start_time": "2021-11-18T19:18:45.463055Z"
    }
   },
   "outputs": [],
   "source": [
    "# set some parameter for show figure in beter style\n",
    "pd.options.display.float_format = '{:.1f}'.format\n",
    "sns.set()\n",
    "plt.style.use('ggplot') # 'Solarize_Light2', '_classic_test_patch', 'bmh', 'classic', 'dark_background', 'fast', 'fivethirtyeight', 'ggplot', 'grayscale', 'seaborn', 'seaborn-bright', 'seaborn-colorblind', 'seaborn-dark', 'seaborn-dark-palette', 'seaborn-darkgrid', 'seaborn-deep', 'seaborn-muted', 'seaborn-notebook', 'seaborn-paper', 'seaborn-pastel', 'seaborn-poster', 'seaborn-talk', 'seaborn-ticks', 'seaborn-white', 'seaborn-whitegrid', 'tableau-colorblind10'\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Read datase"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-11-18T19:18:46.423408Z",
     "start_time": "2021-11-18T19:18:46.362952Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "dataset shape is equal : (452, 280)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "      <th>...</th>\n",
       "      <th>270</th>\n",
       "      <th>271</th>\n",
       "      <th>272</th>\n",
       "      <th>273</th>\n",
       "      <th>274</th>\n",
       "      <th>275</th>\n",
       "      <th>276</th>\n",
       "      <th>277</th>\n",
       "      <th>278</th>\n",
       "      <th>279</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>75</td>\n",
       "      <td>0</td>\n",
       "      <td>190</td>\n",
       "      <td>80</td>\n",
       "      <td>91</td>\n",
       "      <td>193</td>\n",
       "      <td>371</td>\n",
       "      <td>174</td>\n",
       "      <td>121</td>\n",
       "      <td>-16</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>-0.9</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.9</td>\n",
       "      <td>2.9</td>\n",
       "      <td>23.3</td>\n",
       "      <td>49.4</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>56</td>\n",
       "      <td>1</td>\n",
       "      <td>165</td>\n",
       "      <td>64</td>\n",
       "      <td>81</td>\n",
       "      <td>174</td>\n",
       "      <td>401</td>\n",
       "      <td>149</td>\n",
       "      <td>39</td>\n",
       "      <td>25</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>8.5</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.2</td>\n",
       "      <td>2.1</td>\n",
       "      <td>20.4</td>\n",
       "      <td>38.8</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>2 rows × 280 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   0    1    2    3    4    5    6    7    8    9    ... 270 271  272 273 274  \\\n",
       "0   75    0  190   80   91  193  371  174  121  -16  ... 0.0 9.0 -0.9 0.0 0.0   \n",
       "1   56    1  165   64   81  174  401  149   39   25  ... 0.0 8.5  0.0 0.0 0.0   \n",
       "\n",
       "   275  276  277  278  279  \n",
       "0  0.9  2.9 23.3 49.4    8  \n",
       "1  0.2  2.1 20.4 38.8    6  \n",
       "\n",
       "[2 rows x 280 columns]"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = pd.read_csv('./arrhythmia.data',header=None)\n",
    "print(f'dataset shape is equal : {df.shape}')\n",
    "df.head(2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Dataset Preparation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-11-18T19:18:47.676348Z",
     "start_time": "2021-11-18T19:18:47.653194Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "      <th>...</th>\n",
       "      <th>270</th>\n",
       "      <th>271</th>\n",
       "      <th>272</th>\n",
       "      <th>273</th>\n",
       "      <th>274</th>\n",
       "      <th>275</th>\n",
       "      <th>276</th>\n",
       "      <th>277</th>\n",
       "      <th>278</th>\n",
       "      <th>279</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>37</td>\n",
       "      <td>1</td>\n",
       "      <td>159</td>\n",
       "      <td>56</td>\n",
       "      <td>75</td>\n",
       "      <td>162</td>\n",
       "      <td>339</td>\n",
       "      <td>141</td>\n",
       "      <td>96</td>\n",
       "      <td>60</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>-1.3</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.4</td>\n",
       "      <td>1.2</td>\n",
       "      <td>9.4</td>\n",
       "      <td>16.6</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>52</td>\n",
       "      <td>0</td>\n",
       "      <td>165</td>\n",
       "      <td>80</td>\n",
       "      <td>88</td>\n",
       "      <td>202</td>\n",
       "      <td>369</td>\n",
       "      <td>113</td>\n",
       "      <td>97</td>\n",
       "      <td>58</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.5</td>\n",
       "      <td>10.7</td>\n",
       "      <td>-1.7</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.4</td>\n",
       "      <td>-0.3</td>\n",
       "      <td>19.6</td>\n",
       "      <td>17.7</td>\n",
       "      <td>16</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>2 rows × 280 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   0    1    2    3    4    5    6    7    8    9    ...  270  271  272 273  \\\n",
       "0   37    1  159   56   75  162  339  141   96   60  ...  0.0  5.0 -1.3 0.0   \n",
       "1   52    0  165   80   88  202  369  113   97   58  ... -0.5 10.7 -1.7 0.0   \n",
       "\n",
       "  274  275  276  277  278  279  \n",
       "0 0.0  0.4  1.2  9.4 16.6    1  \n",
       "1 0.0  0.4 -0.3 19.6 17.7   16  \n",
       "\n",
       "[2 rows x 280 columns]"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# shuffle \n",
    "df = df.sample(frac=1, random_state=0).reset_index(drop=True)   \n",
    "df.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-11-18T19:18:49.164399Z",
     "start_time": "2021-11-18T19:18:48.231566Z"
    }
   },
   "outputs": [],
   "source": [
    "# replace ? with mean value\n",
    "for c in df.iloc[:,:-1].columns:        \n",
    "    notMissedIndex = df[c].apply(lambda x : str(x) != '?')    \n",
    "    MissedIndex = df[c].apply(lambda x : str(x) == '?')    \n",
    "    mean = np.mean(df[c][notMissedIndex].apply(lambda x : float(x)).unique())          \n",
    "    df[c] = df[c].replace('?', mean)    \n",
    "    df[c] = df[c].apply(lambda x : float(x))    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-11-18T19:18:49.267050Z",
     "start_time": "2021-11-18T19:18:49.242082Z"
    }
   },
   "outputs": [],
   "source": [
    "for i in [0,5,8]: # replace 0 with mean value\n",
    "    notMissedIndex = df.iloc[:,i].apply(lambda x : x != 0)\n",
    "    MissedIndex = df.iloc[:,i].apply(lambda x : x == 0)\n",
    "    mean = np.mean(df.iloc[:,i][notMissedIndex].apply(lambda x : x).unique())          \n",
    "    df.iloc[:,i] = df.iloc[:,i].replace(0, mean)    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-11-18T19:18:50.375683Z",
     "start_time": "2021-11-18T19:18:50.207079Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "      <th>...</th>\n",
       "      <th>270</th>\n",
       "      <th>271</th>\n",
       "      <th>272</th>\n",
       "      <th>273</th>\n",
       "      <th>274</th>\n",
       "      <th>275</th>\n",
       "      <th>276</th>\n",
       "      <th>277</th>\n",
       "      <th>278</th>\n",
       "      <th>279</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>-0.6</td>\n",
       "      <td>1.0</td>\n",
       "      <td>-0.2</td>\n",
       "      <td>-0.7</td>\n",
       "      <td>-0.9</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.8</td>\n",
       "      <td>-0.8</td>\n",
       "      <td>0.2</td>\n",
       "      <td>0.6</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-1.2</td>\n",
       "      <td>-1.3</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.3</td>\n",
       "      <td>-0.0</td>\n",
       "      <td>-0.7</td>\n",
       "      <td>-0.7</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.3</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.0</td>\n",
       "      <td>0.7</td>\n",
       "      <td>-0.1</td>\n",
       "      <td>1.3</td>\n",
       "      <td>0.1</td>\n",
       "      <td>-1.6</td>\n",
       "      <td>0.2</td>\n",
       "      <td>0.5</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.5</td>\n",
       "      <td>0.5</td>\n",
       "      <td>-1.7</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.3</td>\n",
       "      <td>-1.1</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.6</td>\n",
       "      <td>16</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>2 rows × 280 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   0    1    2    3    4    5    6    7    8    9    ...  270  271  272  273  \\\n",
       "0 -0.6  1.0 -0.2 -0.7 -0.9  0.0 -0.8 -0.8  0.2  0.6  ...  0.0 -1.2 -1.3  0.0   \n",
       "1  0.3  0.0 -0.0  0.7 -0.1  1.3  0.1 -1.6  0.2  0.5  ... -0.5  0.5 -1.7  0.0   \n",
       "\n",
       "   274  275  276  277  278  279  \n",
       "0  0.0 -0.3 -0.0 -0.7 -0.7    1  \n",
       "1  0.0 -0.3 -1.1  0.0 -0.6   16  \n",
       "\n",
       "[2 rows x 280 columns]"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "for c in df.iloc[:,:-1].columns: # standardization\n",
    "    if(sum(df[c].unique()) > 2):\n",
    "        df[c] = (df[c] - df[c].mean()) / df[c].std()    \n",
    "df.head(2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## In First step, split Feature and Tag from each other"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-11-18T19:18:55.265698Z",
     "start_time": "2021-11-18T19:18:55.246000Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Feature shape : (452, 279)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "      <th>...</th>\n",
       "      <th>269</th>\n",
       "      <th>270</th>\n",
       "      <th>271</th>\n",
       "      <th>272</th>\n",
       "      <th>273</th>\n",
       "      <th>274</th>\n",
       "      <th>275</th>\n",
       "      <th>276</th>\n",
       "      <th>277</th>\n",
       "      <th>278</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>-0.6</td>\n",
       "      <td>1.0</td>\n",
       "      <td>-0.2</td>\n",
       "      <td>-0.7</td>\n",
       "      <td>-0.9</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.8</td>\n",
       "      <td>-0.8</td>\n",
       "      <td>0.2</td>\n",
       "      <td>0.6</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-1.2</td>\n",
       "      <td>-1.3</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.3</td>\n",
       "      <td>-0.0</td>\n",
       "      <td>-0.7</td>\n",
       "      <td>-0.7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.3</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.0</td>\n",
       "      <td>0.7</td>\n",
       "      <td>-0.1</td>\n",
       "      <td>1.3</td>\n",
       "      <td>0.1</td>\n",
       "      <td>-1.6</td>\n",
       "      <td>0.2</td>\n",
       "      <td>0.5</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.4</td>\n",
       "      <td>-0.5</td>\n",
       "      <td>0.5</td>\n",
       "      <td>-1.7</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.3</td>\n",
       "      <td>-1.1</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.6</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>2 rows × 279 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   0    1    2    3    4    5    6    7    8    9    ...  269  270  271  272  \\\n",
       "0 -0.6  1.0 -0.2 -0.7 -0.9  0.0 -0.8 -0.8  0.2  0.6  ...  0.0  0.0 -1.2 -1.3   \n",
       "1  0.3  0.0 -0.0  0.7 -0.1  1.3  0.1 -1.6  0.2  0.5  ... -0.4 -0.5  0.5 -1.7   \n",
       "\n",
       "   273  274  275  276  277  278  \n",
       "0  0.0  0.0 -0.3 -0.0 -0.7 -0.7  \n",
       "1  0.0  0.0 -0.3 -1.1  0.0 -0.6  \n",
       "\n",
       "[2 rows x 279 columns]"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X = df.iloc[:,:-1]\n",
    "print(f'Feature shape : {X.shape}')\n",
    "X.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-11-18T19:18:56.140017Z",
     "start_time": "2021-11-18T19:18:56.132598Z"
    },
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Target shape : (452,)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "0     1\n",
       "1    16\n",
       "Name: 279, dtype: int64"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Y = df.iloc[:,-1]\n",
    "print(f'Target shape : {Y.shape}')\n",
    "Y.head(2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Use PCA in Dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-11-18T19:19:00.874564Z",
     "start_time": "2021-11-18T19:19:00.437675Z"
    }
   },
   "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": [
    "pca = PCA(n_components=279)\n",
    "principal = pca.fit_transform(X)\n",
    "plt.plot(pca.explained_variance_ratio_)\n",
    "plt.xlabel('number of components')\n",
    "plt.ylabel('variance ratio')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-11-18T19:19:18.957355Z",
     "start_time": "2021-11-18T19:19:09.670393Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "lowest number of components with variances cutoff 0.995 is 43\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "variance_cutoff = 0.995 # save 99.5 % of data ratio\n",
    "\n",
    "variances = list()\n",
    "for i in range(1,X.shape[1]):\n",
    "    pca = PCA(n_components=i)\n",
    "    principal = pca.fit_transform(X)\n",
    "    variances.append(sum(pca.explained_variance_ratio_))\n",
    "d = [i for i in range(0,len(variances)) if variances[i] >= variance_cutoff][0]\n",
    "print(f'lowest number of components with variances cutoff {variance_cutoff} is {d}')\n",
    "\n",
    "plt.plot(variances)\n",
    "plt.plot(d,variances[d],'ys')\n",
    "plt.xlabel('number of components')\n",
    "plt.ylabel('total variances')\n",
    "plt.title('Ratio of components to variation')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-11-18T19:19:19.090655Z",
     "start_time": "2021-11-18T19:19:19.048059Z"
    }
   },
   "outputs": [],
   "source": [
    "pca = PCA(n_components=d)\n",
    "Xpca = pca.fit_transform(X)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Train Test split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-11-18T19:19:19.191824Z",
     "start_time": "2021-11-18T19:19:19.184128Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "X_train shape : (316, 43)\n",
      "y_train shape : (316,)\n",
      "X_test shape : (136, 43)\n",
      "y_test shape : (136,)\n"
     ]
    }
   ],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(Xpca, Y, test_size = 0.3, random_state = 0)\n",
    "print(f'X_train shape : {X_train.shape}')\n",
    "print(f'y_train shape : {y_train.shape}')\n",
    "print(f'X_test shape : {X_test.shape}')\n",
    "print(f'y_test shape : {y_test.shape}')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Run for two class mode :"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-11-18T19:19:19.847763Z",
     "start_time": "2021-11-18T19:19:19.838232Z"
    }
   },
   "outputs": [],
   "source": [
    "y_train2 = y_train.apply(lambda x : int(x != 1))\n",
    "y_test2 = y_test.apply(lambda x : int(x != 1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-11-18T19:19:29.686658Z",
     "start_time": "2021-11-18T19:19:29.471038Z"
    }
   },
   "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": [
    "tag = df.iloc[:,-1]\n",
    "tag = tag != 1\n",
    "ax = tag.value_counts(sort = False).plot(kind='bar',title='show calss histogram')\n",
    "ax.set_xticks(range(0,2))\n",
    "ax.set_xticklabels(['Normal', 'Arrhythmia'])\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Use sklearn MLP"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-11-18T19:19:36.729422Z",
     "start_time": "2021-11-18T19:19:31.390055Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0.6875     0.52380952 0.76190476 0.65079365 0.77777778]\n",
      "0.6803571428571428\n"
     ]
    }
   ],
   "source": [
    "alpha = 0.01\n",
    "clf = MLPClassifier(max_iter=10000, activation='relu', solver='adam', alpha=alpha,learning_rate = 'invscaling'\n",
    "                    ,learning_rate_init = 0.01, power_t = 0.2,\n",
    "                     hidden_layer_sizes=(43,43,21,10,5),momentum=0.9,batch_size='auto', random_state=10, validation_fraction=0.15\n",
    "                   ,early_stopping =True, n_iter_no_change= 50)\n",
    "\n",
    "scores = cross_val_score(clf, X_train, y_train2, cv=5)\n",
    "print(scores)\n",
    "print(scores.mean())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-11-18T19:19:40.627091Z",
     "start_time": "2021-11-18T19:19:38.989313Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "MLPClassifier(alpha=0.01, early_stopping=True,\n",
       "              hidden_layer_sizes=(43, 43, 21, 10, 5),\n",
       "              learning_rate='invscaling', learning_rate_init=0.01,\n",
       "              max_iter=10000, n_iter_no_change=50, power_t=0.2, random_state=10,\n",
       "              validation_fraction=0.15)"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clf.fit(X_train,y_train2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-11-18T19:19:40.813934Z",
     "start_time": "2021-11-18T19:19:40.794502Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train accuracy is : 0.9145569620253164\n",
      "Test accuracy is : 0.6764705882352942\n"
     ]
    }
   ],
   "source": [
    "print(f'Train accuracy is : {metrics.accuracy_score(clf.predict(X_train), y_train2)}')\n",
    "print(f'Test accuracy is : {metrics.accuracy_score(clf.predict(X_test), y_test2)}')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Train data confusion matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-11-18T19:37:26.321555Z",
     "start_time": "2021-11-18T19:37:25.736685Z"
    },
    "run_control": {
     "marked": true
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "matrix = metrics.confusion_matrix(y_train2, clf.predict(X_train))\n",
    "fig, ax = plt.subplots(figsize=(4,4))\n",
    "sns.heatmap(data=matrix, cmap=\"coolwarm\",linecolor='white',linewidths=2,\n",
    "    xticklabels=['Normal', 'Arrhythmia'],yticklabels=['Normal', 'Arrhythmia'],annot=True,fmt=\"d\")    \n",
    "plt.title(\"Confusion Matrix\")\n",
    "plt.ylabel(\"True Label\")\n",
    "plt.xlabel(\"Predicted Label\")\n",
    "ax.set_ylim([0,2])\n",
    "plt.show()    \n",
    "\n",
    "matrix = metrics.confusion_matrix(y_train2, clf.predict(X_train))\n",
    "matrix =  np.divide(matrix.transpose(),np.sum(matrix, axis=1)).transpose() * 100\n",
    "fig, ax = plt.subplots(figsize=(4,4))\n",
    "ax = sns.heatmap(data=matrix, cmap=\"coolwarm\",linecolor='white',linewidths=2,\n",
    "    xticklabels=['Normal', 'Arrhythmia'],yticklabels=['Normal', 'Arrhythmia'],annot=True,fmt=\".2f\")    \n",
    "for t in ax.texts:t.set_text(t.get_text() + \"%\")\n",
    "plt.title(\"Confusion Matrix\")\n",
    "plt.ylabel(\"True Label\")\n",
    "plt.xlabel(\"Predicted Label\")\n",
    "ax.set_ylim([0,2])\n",
    "plt.show()    "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Test data confusion matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-11-18T19:37:28.123060Z",
     "start_time": "2021-11-18T19:37:27.535753Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "matrix = metrics.confusion_matrix(y_test2, clf.predict(X_test))\n",
    "fig, ax = plt.subplots(figsize=(4,4))\n",
    "sns.heatmap(data=matrix, cmap=\"coolwarm\",linecolor='white',linewidths=2,\n",
    "    xticklabels=['Normal', 'Arrhythmia'],yticklabels=['Normal', 'Arrhythmia'],annot=True,fmt=\"d\")    \n",
    "plt.title(\"Confusion Matrix\")\n",
    "plt.ylabel(\"True Label\")\n",
    "plt.xlabel(\"Predicted Label\")\n",
    "ax.set_ylim([0,2])\n",
    "plt.show()    \n",
    "\n",
    "matrix = metrics.confusion_matrix(y_test2, clf.predict(X_test))\n",
    "matrix =  np.divide(matrix.transpose(),np.sum(matrix, axis=1)).transpose() * 100\n",
    "fig, ax = plt.subplots(figsize=(4,4))\n",
    "ax = sns.heatmap(data=matrix, cmap=\"coolwarm\",linecolor='white',linewidths=2,\n",
    "    xticklabels=['Normal', 'Arrhythmia'],yticklabels=['Normal', 'Arrhythmia'],annot=True,fmt=\".2f\")    \n",
    "for t in ax.texts:t.set_text(t.get_text() + \"%\")\n",
    "plt.title(\"Confusion Matrix\")\n",
    "plt.ylabel(\"True Label\")\n",
    "plt.xlabel(\"Predicted Label\")\n",
    "ax.set_ylim([0,2])\n",
    "plt.show()    "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# All data confusion matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 168,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "matrix = metrics.confusion_matrix(np.concatenate((y_train2, y_test2)), clf.predict(np.concatenate((X_train,X_test))))\n",
    "fig, ax = plt.subplots(figsize=(4,4))\n",
    "sns.heatmap(data=matrix, cmap=\"coolwarm\",linecolor='white',linewidths=2,\n",
    "    xticklabels=['Normal', 'Arrhythmia'],yticklabels=['Normal', 'Arrhythmia'],annot=True,fmt=\"d\")    \n",
    "plt.title(\"Confusion Matrix\")\n",
    "plt.ylabel(\"True Label\")\n",
    "plt.xlabel(\"Predicted Label\")\n",
    "ax.set_ylim([0,2])\n",
    "plt.show()    \n",
    "\n",
    "matrix = metrics.confusion_matrix(y_test2, clf.predict(X_test))\n",
    "matrix =  np.divide(matrix.transpose(),np.sum(matrix, axis=1)).transpose() * 100\n",
    "fig, ax = plt.subplots(figsize=(4,4))\n",
    "ax = sns.heatmap(data=matrix, cmap=\"coolwarm\",linecolor='white',linewidths=2,\n",
    "    xticklabels=['Normal', 'Arrhythmia'],yticklabels=['Normal', 'Arrhythmia'],annot=True,fmt=\".2f\")    \n",
    "for t in ax.texts:t.set_text(t.get_text() + \"%\")\n",
    "plt.title(\"Confusion Matrix\")\n",
    "plt.ylabel(\"True Label\")\n",
    "plt.xlabel(\"Predicted Label\")\n",
    "ax.set_ylim([0,2])\n",
    "plt.show()    "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# for 16 class :"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 186,
   "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": [
    "tag = df.iloc[:,-1]\n",
    "ax = tag.value_counts(sort = False).plot(kind='bar',title='show calss histogram')\n",
    "ax.set_xticks(range(0,16))\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Convert lable to one-hot"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 187,
   "metadata": {},
   "outputs": [],
   "source": [
    "tag = np.zeros((y_train.shape[0],16),dtype=int)\n",
    "for i in range(0,y_train.shape[0]):\n",
    "    tag[i,y_train.iloc[i]-1] = 1\n",
    "y_train = tag\n",
    "\n",
    "\n",
    "tag = np.zeros((y_test.shape[0],16))\n",
    "for i in range(0,y_test.shape[0]):\n",
    "    tag[i,y_test.iloc[i] - 1] = 1\n",
    "y_test = tag"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### after cross validation chose hyperparameter value"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 198,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0.4375     0.61904762 0.46031746 0.6031746  0.49206349]\n",
      "0.5224206349206348\n"
     ]
    }
   ],
   "source": [
    "alpha = 0.01\n",
    "clf = MLPClassifier(max_iter=3000, activation='logistic', solver='adam', alpha=alpha,learning_rate = 'invscaling'\n",
    "                    ,learning_rate_init = 0.01, power_t = 0.2,\n",
    "                     hidden_layer_sizes=(43,43,21,16,),momentum=0.9,batch_size='auto', random_state=10, validation_fraction=0.15\n",
    "                   ,early_stopping =True, n_iter_no_change= 50)\n",
    "scores = cross_val_score(clf, X_train, y_train, cv=5)\n",
    "print(scores)\n",
    "print(scores.mean())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 199,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "MLPClassifier(activation='logistic', alpha=0.01, batch_size='auto', beta_1=0.9,\n",
       "              beta_2=0.999, early_stopping=True, epsilon=1e-08,\n",
       "              hidden_layer_sizes=(43, 43, 21, 16), learning_rate='invscaling',\n",
       "              learning_rate_init=0.01, max_fun=15000, max_iter=3000,\n",
       "              momentum=0.9, n_iter_no_change=50, nesterovs_momentum=True,\n",
       "              power_t=0.2, random_state=10, shuffle=True, solver='adam',\n",
       "              tol=0.0001, validation_fraction=0.15, verbose=False,\n",
       "              warm_start=False)"
      ]
     },
     "execution_count": 199,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clf.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "print(f'Train accuracy is : {metrics.accuracy_score(clf.predict(X_train), y_train)}')\n",
    "print(f'Test accuracy is : {metrics.accuracy_score(clf.predict(X_test), y_test)}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "matrix = create_confusion_matrix(np.argmax(np.concatenate((y_train,y_test)), axis=1), np.argmax(clf.predict(np.concatenate((X_train,X_test))),axis=1))\n",
    "matrix =  np.nan_to_num(np.divide(matrix.transpose(),np.sum(matrix, axis=1)).transpose() * 100 )\n",
    "fig, ax = plt.subplots(figsize=(12,8))\n",
    "sns.heatmap(data=matrix, cmap=\"coolwarm\",linecolor='white',linewidths=2,\n",
    "    xticklabels=[str(i) for i in range(1,17)],yticklabels=[str(i) for i in range(1,17)],annot=True,fmt=\".0f\")    \n",
    "for t in ax.texts:t.set_text(t.get_text() + \"%\")\n",
    "plt.title(\"Confusion Matrix\")\n",
    "plt.ylabel(\"True Label\")\n",
    "plt.xlabel(\"Predicted Label\")\n",
    "ax.set_ylim([0,16])\n",
    "plt.show()    "
   ]
  }
 ],
 "metadata": {
  "hide_input": false,
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.5"
  },
  "varInspector": {
   "cols": {
    "lenName": 16,
    "lenType": 16,
    "lenVar": 40
   },
   "kernels_config": {
    "python": {
     "delete_cmd_postfix": "",
     "delete_cmd_prefix": "del ",
     "library": "var_list.py",
     "varRefreshCmd": "print(var_dic_list())"
    },
    "r": {
     "delete_cmd_postfix": ") ",
     "delete_cmd_prefix": "rm(",
     "library": "var_list.r",
     "varRefreshCmd": "cat(var_dic_list()) "
    }
   },
   "types_to_exclude": [
    "module",
    "function",
    "builtin_function_or_method",
    "instance",
    "_Feature"
   ],
   "window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}