binary_classifier_synthetic_simple.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "31c24d5d",
   "metadata": {},
   "source": [
    "# Overview\n",
    "Train classifier on x acceleration features only. Binary classification. Does not account for balancing data. Strictly reads in data and trains MLP classifier using all tsfel features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "id": "893d77e1",
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import tsfel\n",
    "import numpy as np\n",
    "import pandas as pd"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "50499d64",
   "metadata": {},
   "source": [
    "### Set directory locations"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "id": "d62a23ae",
   "metadata": {},
   "outputs": [],
   "source": [
    "norm_dir = './Data/Normal_data'\n",
    "tremor_4_dir = './Data/Tremor_added/4Hz'\n",
    "tremor_8_dir = './Data/Tremor_added/8Hz'"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e7fdd09f",
   "metadata": {},
   "source": [
    "### Dataframes for all 3 levels. \n",
    "Read in all files"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "id": "6f158e35",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_norm = pd.concat((pd.read_csv(norm_dir+'/'+f, header=0, names=[\"x\", \"y\", \"z\"]) for f in os.listdir(norm_dir)), ignore_index=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "id": "0390ebbd",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_4hz = pd.concat((pd.read_csv(tremor_4_dir+'/'+f, header=0, names=[\"x\", \"y\", \"z\"]) for f in os.listdir(tremor_4_dir)), ignore_index=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "id": "190d5ee0",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_8hz = pd.concat((pd.read_csv(tremor_8_dir+'/'+f, header=0, names=[\"x\", \"y\", \"z\"]) for f in os.listdir(tremor_8_dir)), ignore_index=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "id": "79a433a5",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train_sig_norm = df_norm[\"x\"].to_frame()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7de3aa70",
   "metadata": {},
   "source": [
    "### Extract features for all 3 dataframes. Add labels after feature extraction. 0 for no tremor, 1 for tremor"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "id": "e1ee25d4",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "*** Feature extraction started ***\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "\n",
       "              <p>\n",
       "                  Progress: 100% Complete\n",
       "              <p/>            \n",
       "              <progress\n",
       "                  value='18884'\n",
       "                  max='18884',\n",
       "                  style='width: 25%',\n",
       "              >\n",
       "                  18884\n",
       "              </progress>\n",
       "\n",
       "    "
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "*** Feature extraction finished ***\n",
      "Wall time: 4min 18s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "cfg_file_norm = tsfel.get_features_by_domain()                                                        # If no argument is passed retrieves all available features\n",
    "X_train_norm = tsfel.time_series_features_extractor(cfg_file_norm, X_train_sig_norm, fs=20, window_size=200)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "id": "541bcb4a",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train_norm['label'] = 0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "id": "d2473cad",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train_sig_4hz = df_4hz[\"x\"].to_frame()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "id": "0b7922c9",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "*** Feature extraction started ***\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "\n",
       "              <p>\n",
       "                  Progress: 100% Complete\n",
       "              <p/>            \n",
       "              <progress\n",
       "                  value='8469'\n",
       "                  max='8469',\n",
       "                  style='width: 25%',\n",
       "              >\n",
       "                  8469\n",
       "              </progress>\n",
       "\n",
       "    "
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "*** Feature extraction finished ***\n",
      "Wall time: 1min 17s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "cfg_file_4hz = tsfel.get_features_by_domain()                                                        # If no argument is passed retrieves all available features\n",
    "X_train_4hz = tsfel.time_series_features_extractor(cfg_file_4hz, X_train_sig_4hz, fs=20, window_size=200)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "id": "8497863b",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train_4hz['label'] = 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "id": "aeaffc3e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "*** Feature extraction started ***\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "\n",
       "              <p>\n",
       "                  Progress: 100% Complete\n",
       "              <p/>            \n",
       "              <progress\n",
       "                  value='981'\n",
       "                  max='981',\n",
       "                  style='width: 25%',\n",
       "              >\n",
       "                  981\n",
       "              </progress>\n",
       "\n",
       "    "
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "*** Feature extraction finished ***\n",
      "Wall time: 10.2 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "X_train_sig_8hz = df_8hz[\"x\"].to_frame()\n",
    "cfg_file_8hz = tsfel.get_features_by_domain()                                                        # If no argument is passed retrieves all available features\n",
    "X_train_8hz = tsfel.time_series_features_extractor(cfg_file_8hz, X_train_sig_8hz, fs=20, window_size=200)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "id": "36b64c73",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train_8hz['label'] = 1"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "be792503",
   "metadata": {},
   "source": [
    "### Combine all 3 feature dataframes into one."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "id": "fac867d4",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_x_features = pd.concat([X_train_norm, X_train_4hz, X_train_8hz], ignore_index=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "id": "7fa349f7",
   "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>0_Absolute energy</th>\n",
       "      <th>0_Area under the curve</th>\n",
       "      <th>0_Autocorrelation</th>\n",
       "      <th>0_Centroid</th>\n",
       "      <th>0_ECDF Percentile Count_0</th>\n",
       "      <th>0_ECDF Percentile Count_1</th>\n",
       "      <th>0_ECDF Percentile_0</th>\n",
       "      <th>0_ECDF Percentile_1</th>\n",
       "      <th>0_ECDF_0</th>\n",
       "      <th>0_ECDF_1</th>\n",
       "      <th>...</th>\n",
       "      <th>0_Wavelet variance_1</th>\n",
       "      <th>0_Wavelet variance_2</th>\n",
       "      <th>0_Wavelet variance_3</th>\n",
       "      <th>0_Wavelet variance_4</th>\n",
       "      <th>0_Wavelet variance_5</th>\n",
       "      <th>0_Wavelet variance_6</th>\n",
       "      <th>0_Wavelet variance_7</th>\n",
       "      <th>0_Wavelet variance_8</th>\n",
       "      <th>0_Zero crossing rate</th>\n",
       "      <th>label</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>3.270655e+04</td>\n",
       "      <td>120.651156</td>\n",
       "      <td>3.270655e+04</td>\n",
       "      <td>5.258573</td>\n",
       "      <td>40.0</td>\n",
       "      <td>160.0</td>\n",
       "      <td>8.293514</td>\n",
       "      <td>15.732300</td>\n",
       "      <td>0.005</td>\n",
       "      <td>0.01</td>\n",
       "      <td>...</td>\n",
       "      <td>33.549756</td>\n",
       "      <td>54.615097</td>\n",
       "      <td>59.839033</td>\n",
       "      <td>64.151396</td>\n",
       "      <td>58.447217</td>\n",
       "      <td>45.540403</td>\n",
       "      <td>35.081710</td>\n",
       "      <td>31.245857</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>3.126574e+04</td>\n",
       "      <td>120.502865</td>\n",
       "      <td>3.126574e+04</td>\n",
       "      <td>4.855203</td>\n",
       "      <td>40.0</td>\n",
       "      <td>160.0</td>\n",
       "      <td>9.402030</td>\n",
       "      <td>14.460978</td>\n",
       "      <td>0.005</td>\n",
       "      <td>0.01</td>\n",
       "      <td>...</td>\n",
       "      <td>21.401843</td>\n",
       "      <td>29.617800</td>\n",
       "      <td>36.056555</td>\n",
       "      <td>45.380055</td>\n",
       "      <td>47.727731</td>\n",
       "      <td>44.125786</td>\n",
       "      <td>40.894024</td>\n",
       "      <td>41.280500</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3.129423e+04</td>\n",
       "      <td>120.875463</td>\n",
       "      <td>3.129423e+04</td>\n",
       "      <td>4.872443</td>\n",
       "      <td>40.0</td>\n",
       "      <td>160.0</td>\n",
       "      <td>9.334992</td>\n",
       "      <td>14.858416</td>\n",
       "      <td>0.005</td>\n",
       "      <td>0.01</td>\n",
       "      <td>...</td>\n",
       "      <td>23.898359</td>\n",
       "      <td>33.588663</td>\n",
       "      <td>38.294527</td>\n",
       "      <td>46.768419</td>\n",
       "      <td>49.055807</td>\n",
       "      <td>45.235522</td>\n",
       "      <td>41.550160</td>\n",
       "      <td>41.482886</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3.071311e+04</td>\n",
       "      <td>119.027738</td>\n",
       "      <td>3.071311e+04</td>\n",
       "      <td>5.104300</td>\n",
       "      <td>40.0</td>\n",
       "      <td>160.0</td>\n",
       "      <td>8.710106</td>\n",
       "      <td>14.329297</td>\n",
       "      <td>0.005</td>\n",
       "      <td>0.01</td>\n",
       "      <td>...</td>\n",
       "      <td>29.582273</td>\n",
       "      <td>48.559945</td>\n",
       "      <td>55.129474</td>\n",
       "      <td>63.237796</td>\n",
       "      <td>61.395971</td>\n",
       "      <td>50.875834</td>\n",
       "      <td>41.364797</td>\n",
       "      <td>37.828333</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>3.057171e+04</td>\n",
       "      <td>119.850445</td>\n",
       "      <td>3.057171e+04</td>\n",
       "      <td>5.272381</td>\n",
       "      <td>40.0</td>\n",
       "      <td>160.0</td>\n",
       "      <td>9.392453</td>\n",
       "      <td>14.489709</td>\n",
       "      <td>0.005</td>\n",
       "      <td>0.01</td>\n",
       "      <td>...</td>\n",
       "      <td>16.350803</td>\n",
       "      <td>31.195999</td>\n",
       "      <td>41.947704</td>\n",
       "      <td>50.059430</td>\n",
       "      <td>48.739978</td>\n",
       "      <td>41.516780</td>\n",
       "      <td>36.194489</td>\n",
       "      <td>35.631017</td>\n",
       "      <td>0.0</td>\n",
       "      <td>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>28329</th>\n",
       "      <td>2.297861e+06</td>\n",
       "      <td>297.383914</td>\n",
       "      <td>2.297861e+06</td>\n",
       "      <td>5.026683</td>\n",
       "      <td>40.0</td>\n",
       "      <td>160.0</td>\n",
       "      <td>-135.368116</td>\n",
       "      <td>95.895163</td>\n",
       "      <td>0.005</td>\n",
       "      <td>0.01</td>\n",
       "      <td>...</td>\n",
       "      <td>47.770458</td>\n",
       "      <td>56.085412</td>\n",
       "      <td>61.519417</td>\n",
       "      <td>67.920564</td>\n",
       "      <td>73.575220</td>\n",
       "      <td>79.988524</td>\n",
       "      <td>89.915356</td>\n",
       "      <td>102.937291</td>\n",
       "      <td>159.0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28330</th>\n",
       "      <td>2.311677e+06</td>\n",
       "      <td>299.636214</td>\n",
       "      <td>2.311677e+06</td>\n",
       "      <td>5.014196</td>\n",
       "      <td>40.0</td>\n",
       "      <td>160.0</td>\n",
       "      <td>-138.878017</td>\n",
       "      <td>97.607017</td>\n",
       "      <td>0.005</td>\n",
       "      <td>0.01</td>\n",
       "      <td>...</td>\n",
       "      <td>32.423965</td>\n",
       "      <td>39.788846</td>\n",
       "      <td>53.548670</td>\n",
       "      <td>69.733096</td>\n",
       "      <td>83.903553</td>\n",
       "      <td>95.549151</td>\n",
       "      <td>105.680943</td>\n",
       "      <td>115.117285</td>\n",
       "      <td>159.0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28331</th>\n",
       "      <td>2.305055e+06</td>\n",
       "      <td>297.014713</td>\n",
       "      <td>2.305055e+06</td>\n",
       "      <td>5.016476</td>\n",
       "      <td>40.0</td>\n",
       "      <td>160.0</td>\n",
       "      <td>-131.760055</td>\n",
       "      <td>98.684408</td>\n",
       "      <td>0.005</td>\n",
       "      <td>0.01</td>\n",
       "      <td>...</td>\n",
       "      <td>38.573082</td>\n",
       "      <td>52.580574</td>\n",
       "      <td>66.263598</td>\n",
       "      <td>78.581413</td>\n",
       "      <td>97.636268</td>\n",
       "      <td>122.100078</td>\n",
       "      <td>144.198763</td>\n",
       "      <td>158.076142</td>\n",
       "      <td>159.0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28332</th>\n",
       "      <td>2.304909e+06</td>\n",
       "      <td>295.837664</td>\n",
       "      <td>2.304909e+06</td>\n",
       "      <td>5.020154</td>\n",
       "      <td>40.0</td>\n",
       "      <td>160.0</td>\n",
       "      <td>-136.704081</td>\n",
       "      <td>98.227116</td>\n",
       "      <td>0.005</td>\n",
       "      <td>0.01</td>\n",
       "      <td>...</td>\n",
       "      <td>34.822740</td>\n",
       "      <td>38.273679</td>\n",
       "      <td>45.115673</td>\n",
       "      <td>57.708076</td>\n",
       "      <td>76.654172</td>\n",
       "      <td>101.026351</td>\n",
       "      <td>128.691159</td>\n",
       "      <td>155.703842</td>\n",
       "      <td>159.0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28333</th>\n",
       "      <td>2.302723e+06</td>\n",
       "      <td>298.823832</td>\n",
       "      <td>2.302723e+06</td>\n",
       "      <td>5.017728</td>\n",
       "      <td>40.0</td>\n",
       "      <td>160.0</td>\n",
       "      <td>-131.951591</td>\n",
       "      <td>95.919105</td>\n",
       "      <td>0.005</td>\n",
       "      <td>0.01</td>\n",
       "      <td>...</td>\n",
       "      <td>29.971368</td>\n",
       "      <td>35.332732</td>\n",
       "      <td>44.518774</td>\n",
       "      <td>54.597551</td>\n",
       "      <td>66.197596</td>\n",
       "      <td>80.489521</td>\n",
       "      <td>97.857652</td>\n",
       "      <td>118.868274</td>\n",
       "      <td>159.0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>28334 rows × 236 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       0_Absolute energy  0_Area under the curve  0_Autocorrelation  \\\n",
       "0           3.270655e+04              120.651156       3.270655e+04   \n",
       "1           3.126574e+04              120.502865       3.126574e+04   \n",
       "2           3.129423e+04              120.875463       3.129423e+04   \n",
       "3           3.071311e+04              119.027738       3.071311e+04   \n",
       "4           3.057171e+04              119.850445       3.057171e+04   \n",
       "...                  ...                     ...                ...   \n",
       "28329       2.297861e+06              297.383914       2.297861e+06   \n",
       "28330       2.311677e+06              299.636214       2.311677e+06   \n",
       "28331       2.305055e+06              297.014713       2.305055e+06   \n",
       "28332       2.304909e+06              295.837664       2.304909e+06   \n",
       "28333       2.302723e+06              298.823832       2.302723e+06   \n",
       "\n",
       "       0_Centroid  0_ECDF Percentile Count_0  0_ECDF Percentile Count_1  \\\n",
       "0        5.258573                       40.0                      160.0   \n",
       "1        4.855203                       40.0                      160.0   \n",
       "2        4.872443                       40.0                      160.0   \n",
       "3        5.104300                       40.0                      160.0   \n",
       "4        5.272381                       40.0                      160.0   \n",
       "...           ...                        ...                        ...   \n",
       "28329    5.026683                       40.0                      160.0   \n",
       "28330    5.014196                       40.0                      160.0   \n",
       "28331    5.016476                       40.0                      160.0   \n",
       "28332    5.020154                       40.0                      160.0   \n",
       "28333    5.017728                       40.0                      160.0   \n",
       "\n",
       "       0_ECDF Percentile_0  0_ECDF Percentile_1  0_ECDF_0  0_ECDF_1  ...  \\\n",
       "0                 8.293514            15.732300     0.005      0.01  ...   \n",
       "1                 9.402030            14.460978     0.005      0.01  ...   \n",
       "2                 9.334992            14.858416     0.005      0.01  ...   \n",
       "3                 8.710106            14.329297     0.005      0.01  ...   \n",
       "4                 9.392453            14.489709     0.005      0.01  ...   \n",
       "...                    ...                  ...       ...       ...  ...   \n",
       "28329          -135.368116            95.895163     0.005      0.01  ...   \n",
       "28330          -138.878017            97.607017     0.005      0.01  ...   \n",
       "28331          -131.760055            98.684408     0.005      0.01  ...   \n",
       "28332          -136.704081            98.227116     0.005      0.01  ...   \n",
       "28333          -131.951591            95.919105     0.005      0.01  ...   \n",
       "\n",
       "       0_Wavelet variance_1  0_Wavelet variance_2  0_Wavelet variance_3  \\\n",
       "0                 33.549756             54.615097             59.839033   \n",
       "1                 21.401843             29.617800             36.056555   \n",
       "2                 23.898359             33.588663             38.294527   \n",
       "3                 29.582273             48.559945             55.129474   \n",
       "4                 16.350803             31.195999             41.947704   \n",
       "...                     ...                   ...                   ...   \n",
       "28329             47.770458             56.085412             61.519417   \n",
       "28330             32.423965             39.788846             53.548670   \n",
       "28331             38.573082             52.580574             66.263598   \n",
       "28332             34.822740             38.273679             45.115673   \n",
       "28333             29.971368             35.332732             44.518774   \n",
       "\n",
       "       0_Wavelet variance_4  0_Wavelet variance_5  0_Wavelet variance_6  \\\n",
       "0                 64.151396             58.447217             45.540403   \n",
       "1                 45.380055             47.727731             44.125786   \n",
       "2                 46.768419             49.055807             45.235522   \n",
       "3                 63.237796             61.395971             50.875834   \n",
       "4                 50.059430             48.739978             41.516780   \n",
       "...                     ...                   ...                   ...   \n",
       "28329             67.920564             73.575220             79.988524   \n",
       "28330             69.733096             83.903553             95.549151   \n",
       "28331             78.581413             97.636268            122.100078   \n",
       "28332             57.708076             76.654172            101.026351   \n",
       "28333             54.597551             66.197596             80.489521   \n",
       "\n",
       "       0_Wavelet variance_7  0_Wavelet variance_8  0_Zero crossing rate  label  \n",
       "0                 35.081710             31.245857                   0.0      0  \n",
       "1                 40.894024             41.280500                   0.0      0  \n",
       "2                 41.550160             41.482886                   0.0      0  \n",
       "3                 41.364797             37.828333                   0.0      0  \n",
       "4                 36.194489             35.631017                   0.0      0  \n",
       "...                     ...                   ...                   ...    ...  \n",
       "28329             89.915356            102.937291                 159.0      1  \n",
       "28330            105.680943            115.117285                 159.0      1  \n",
       "28331            144.198763            158.076142                 159.0      1  \n",
       "28332            128.691159            155.703842                 159.0      1  \n",
       "28333             97.857652            118.868274                 159.0      1  \n",
       "\n",
       "[28334 rows x 236 columns]"
      ]
     },
     "execution_count": 87,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_x_features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "id": "eeaa08e7",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import StandardScaler\n",
    "sc = StandardScaler()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "id": "a9d84526",
   "metadata": {},
   "outputs": [],
   "source": [
    "x = df_x_features.drop('label', axis=1).values\n",
    "columns = df_x_features.drop('label', axis=1).columns\n",
    "sc = StandardScaler()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "68601ade",
   "metadata": {},
   "source": [
    "### Use sklearn standard scaler. Apparently better than normalizing for classification.\n",
    "Drops label and then adds back to avoid scaling the label.\n",
    "Note: Normalization is better for regression tasks."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "id": "f4330ee6",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Wall time: 86 ms\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "x_scaled = sc.fit_transform(x)\n",
    "df_x_features_scaled = pd.DataFrame(x_scaled, columns=columns)\n",
    "df_x_features_scaled['label'] = df_x_features['label']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "id": "4108be51",
   "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>0_Absolute energy</th>\n",
       "      <th>0_Area under the curve</th>\n",
       "      <th>0_Autocorrelation</th>\n",
       "      <th>0_Centroid</th>\n",
       "      <th>0_ECDF Percentile Count_0</th>\n",
       "      <th>0_ECDF Percentile Count_1</th>\n",
       "      <th>0_ECDF Percentile_0</th>\n",
       "      <th>0_ECDF Percentile_1</th>\n",
       "      <th>0_ECDF_0</th>\n",
       "      <th>0_ECDF_1</th>\n",
       "      <th>...</th>\n",
       "      <th>0_Wavelet variance_1</th>\n",
       "      <th>0_Wavelet variance_2</th>\n",
       "      <th>0_Wavelet variance_3</th>\n",
       "      <th>0_Wavelet variance_4</th>\n",
       "      <th>0_Wavelet variance_5</th>\n",
       "      <th>0_Wavelet variance_6</th>\n",
       "      <th>0_Wavelet variance_7</th>\n",
       "      <th>0_Wavelet variance_8</th>\n",
       "      <th>0_Zero crossing rate</th>\n",
       "      <th>label</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>-0.163138</td>\n",
       "      <td>0.407456</td>\n",
       "      <td>-0.163138</td>\n",
       "      <td>0.415639</td>\n",
       "      <td>0.035147</td>\n",
       "      <td>0.035635</td>\n",
       "      <td>0.899612</td>\n",
       "      <td>0.409227</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.363412</td>\n",
       "      <td>0.151841</td>\n",
       "      <td>0.174771</td>\n",
       "      <td>0.331125</td>\n",
       "      <td>0.315621</td>\n",
       "      <td>0.098100</td>\n",
       "      <td>-0.074447</td>\n",
       "      <td>-0.154870</td>\n",
       "      <td>-0.797524</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>-0.168711</td>\n",
       "      <td>0.405542</td>\n",
       "      <td>-0.168711</td>\n",
       "      <td>-0.163587</td>\n",
       "      <td>0.035147</td>\n",
       "      <td>0.035635</td>\n",
       "      <td>0.953036</td>\n",
       "      <td>0.330044</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.401844</td>\n",
       "      <td>-0.129688</td>\n",
       "      <td>-0.072401</td>\n",
       "      <td>0.088765</td>\n",
       "      <td>0.165894</td>\n",
       "      <td>0.080628</td>\n",
       "      <td>-0.012129</td>\n",
       "      <td>-0.059523</td>\n",
       "      <td>-0.797524</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>-0.168601</td>\n",
       "      <td>0.410350</td>\n",
       "      <td>-0.168601</td>\n",
       "      <td>-0.138832</td>\n",
       "      <td>0.035147</td>\n",
       "      <td>0.035635</td>\n",
       "      <td>0.949805</td>\n",
       "      <td>0.354798</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.393946</td>\n",
       "      <td>-0.084967</td>\n",
       "      <td>-0.049142</td>\n",
       "      <td>0.106690</td>\n",
       "      <td>0.184444</td>\n",
       "      <td>0.094334</td>\n",
       "      <td>-0.005094</td>\n",
       "      <td>-0.057600</td>\n",
       "      <td>-0.797524</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>-0.170849</td>\n",
       "      <td>0.386508</td>\n",
       "      <td>-0.170849</td>\n",
       "      <td>0.194108</td>\n",
       "      <td>0.035147</td>\n",
       "      <td>0.035635</td>\n",
       "      <td>0.919690</td>\n",
       "      <td>0.321842</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.375964</td>\n",
       "      <td>0.083646</td>\n",
       "      <td>0.125824</td>\n",
       "      <td>0.319330</td>\n",
       "      <td>0.356808</td>\n",
       "      <td>0.163995</td>\n",
       "      <td>-0.007081</td>\n",
       "      <td>-0.092325</td>\n",
       "      <td>-0.797524</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>-0.171396</td>\n",
       "      <td>0.397124</td>\n",
       "      <td>-0.171396</td>\n",
       "      <td>0.435468</td>\n",
       "      <td>0.035147</td>\n",
       "      <td>0.035635</td>\n",
       "      <td>0.952575</td>\n",
       "      <td>0.331833</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.417824</td>\n",
       "      <td>-0.111914</td>\n",
       "      <td>-0.011174</td>\n",
       "      <td>0.149181</td>\n",
       "      <td>0.180033</td>\n",
       "      <td>0.048406</td>\n",
       "      <td>-0.062516</td>\n",
       "      <td>-0.113203</td>\n",
       "      <td>-0.797524</td>\n",
       "      <td>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>28329</th>\n",
       "      <td>8.599752</td>\n",
       "      <td>2.688004</td>\n",
       "      <td>8.599752</td>\n",
       "      <td>0.082652</td>\n",
       "      <td>0.035147</td>\n",
       "      <td>0.035635</td>\n",
       "      <td>-6.023995</td>\n",
       "      <td>5.402118</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.318423</td>\n",
       "      <td>0.168400</td>\n",
       "      <td>0.192235</td>\n",
       "      <td>0.379790</td>\n",
       "      <td>0.526924</td>\n",
       "      <td>0.523551</td>\n",
       "      <td>0.513470</td>\n",
       "      <td>0.526327</td>\n",
       "      <td>3.423047</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28330</th>\n",
       "      <td>8.653201</td>\n",
       "      <td>2.717068</td>\n",
       "      <td>8.653201</td>\n",
       "      <td>0.064720</td>\n",
       "      <td>0.035147</td>\n",
       "      <td>0.035635</td>\n",
       "      <td>-6.193150</td>\n",
       "      <td>5.508740</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.366974</td>\n",
       "      <td>-0.015138</td>\n",
       "      <td>0.109395</td>\n",
       "      <td>0.403192</td>\n",
       "      <td>0.671188</td>\n",
       "      <td>0.715732</td>\n",
       "      <td>0.682506</td>\n",
       "      <td>0.642059</td>\n",
       "      <td>3.423047</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28331</th>\n",
       "      <td>8.627584</td>\n",
       "      <td>2.683240</td>\n",
       "      <td>8.627584</td>\n",
       "      <td>0.067995</td>\n",
       "      <td>0.035147</td>\n",
       "      <td>0.035635</td>\n",
       "      <td>-5.850108</td>\n",
       "      <td>5.575845</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.347520</td>\n",
       "      <td>0.128928</td>\n",
       "      <td>0.241542</td>\n",
       "      <td>0.517434</td>\n",
       "      <td>0.863002</td>\n",
       "      <td>1.043649</td>\n",
       "      <td>1.095488</td>\n",
       "      <td>1.050245</td>\n",
       "      <td>3.423047</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28332</th>\n",
       "      <td>8.627019</td>\n",
       "      <td>2.668052</td>\n",
       "      <td>8.627019</td>\n",
       "      <td>0.073276</td>\n",
       "      <td>0.035147</td>\n",
       "      <td>0.035635</td>\n",
       "      <td>-6.088380</td>\n",
       "      <td>5.547363</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.359385</td>\n",
       "      <td>-0.032202</td>\n",
       "      <td>0.021750</td>\n",
       "      <td>0.247934</td>\n",
       "      <td>0.569930</td>\n",
       "      <td>0.783378</td>\n",
       "      <td>0.929218</td>\n",
       "      <td>1.027704</td>\n",
       "      <td>3.423047</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28333</th>\n",
       "      <td>8.618563</td>\n",
       "      <td>2.706585</td>\n",
       "      <td>8.618563</td>\n",
       "      <td>0.069793</td>\n",
       "      <td>0.035147</td>\n",
       "      <td>0.035635</td>\n",
       "      <td>-5.859339</td>\n",
       "      <td>5.403610</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.374733</td>\n",
       "      <td>-0.065324</td>\n",
       "      <td>0.015547</td>\n",
       "      <td>0.207774</td>\n",
       "      <td>0.423876</td>\n",
       "      <td>0.529738</td>\n",
       "      <td>0.598626</td>\n",
       "      <td>0.677700</td>\n",
       "      <td>3.423047</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>28334 rows × 236 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       0_Absolute energy  0_Area under the curve  0_Autocorrelation  \\\n",
       "0              -0.163138                0.407456          -0.163138   \n",
       "1              -0.168711                0.405542          -0.168711   \n",
       "2              -0.168601                0.410350          -0.168601   \n",
       "3              -0.170849                0.386508          -0.170849   \n",
       "4              -0.171396                0.397124          -0.171396   \n",
       "...                  ...                     ...                ...   \n",
       "28329           8.599752                2.688004           8.599752   \n",
       "28330           8.653201                2.717068           8.653201   \n",
       "28331           8.627584                2.683240           8.627584   \n",
       "28332           8.627019                2.668052           8.627019   \n",
       "28333           8.618563                2.706585           8.618563   \n",
       "\n",
       "       0_Centroid  0_ECDF Percentile Count_0  0_ECDF Percentile Count_1  \\\n",
       "0        0.415639                   0.035147                   0.035635   \n",
       "1       -0.163587                   0.035147                   0.035635   \n",
       "2       -0.138832                   0.035147                   0.035635   \n",
       "3        0.194108                   0.035147                   0.035635   \n",
       "4        0.435468                   0.035147                   0.035635   \n",
       "...           ...                        ...                        ...   \n",
       "28329    0.082652                   0.035147                   0.035635   \n",
       "28330    0.064720                   0.035147                   0.035635   \n",
       "28331    0.067995                   0.035147                   0.035635   \n",
       "28332    0.073276                   0.035147                   0.035635   \n",
       "28333    0.069793                   0.035147                   0.035635   \n",
       "\n",
       "       0_ECDF Percentile_0  0_ECDF Percentile_1  0_ECDF_0  0_ECDF_1  ...  \\\n",
       "0                 0.899612             0.409227       1.0       1.0  ...   \n",
       "1                 0.953036             0.330044       1.0       1.0  ...   \n",
       "2                 0.949805             0.354798       1.0       1.0  ...   \n",
       "3                 0.919690             0.321842       1.0       1.0  ...   \n",
       "4                 0.952575             0.331833       1.0       1.0  ...   \n",
       "...                    ...                  ...       ...       ...  ...   \n",
       "28329            -6.023995             5.402118       1.0       1.0  ...   \n",
       "28330            -6.193150             5.508740       1.0       1.0  ...   \n",
       "28331            -5.850108             5.575845       1.0       1.0  ...   \n",
       "28332            -6.088380             5.547363       1.0       1.0  ...   \n",
       "28333            -5.859339             5.403610       1.0       1.0  ...   \n",
       "\n",
       "       0_Wavelet variance_1  0_Wavelet variance_2  0_Wavelet variance_3  \\\n",
       "0                 -0.363412              0.151841              0.174771   \n",
       "1                 -0.401844             -0.129688             -0.072401   \n",
       "2                 -0.393946             -0.084967             -0.049142   \n",
       "3                 -0.375964              0.083646              0.125824   \n",
       "4                 -0.417824             -0.111914             -0.011174   \n",
       "...                     ...                   ...                   ...   \n",
       "28329             -0.318423              0.168400              0.192235   \n",
       "28330             -0.366974             -0.015138              0.109395   \n",
       "28331             -0.347520              0.128928              0.241542   \n",
       "28332             -0.359385             -0.032202              0.021750   \n",
       "28333             -0.374733             -0.065324              0.015547   \n",
       "\n",
       "       0_Wavelet variance_4  0_Wavelet variance_5  0_Wavelet variance_6  \\\n",
       "0                  0.331125              0.315621              0.098100   \n",
       "1                  0.088765              0.165894              0.080628   \n",
       "2                  0.106690              0.184444              0.094334   \n",
       "3                  0.319330              0.356808              0.163995   \n",
       "4                  0.149181              0.180033              0.048406   \n",
       "...                     ...                   ...                   ...   \n",
       "28329              0.379790              0.526924              0.523551   \n",
       "28330              0.403192              0.671188              0.715732   \n",
       "28331              0.517434              0.863002              1.043649   \n",
       "28332              0.247934              0.569930              0.783378   \n",
       "28333              0.207774              0.423876              0.529738   \n",
       "\n",
       "       0_Wavelet variance_7  0_Wavelet variance_8  0_Zero crossing rate  label  \n",
       "0                 -0.074447             -0.154870             -0.797524      0  \n",
       "1                 -0.012129             -0.059523             -0.797524      0  \n",
       "2                 -0.005094             -0.057600             -0.797524      0  \n",
       "3                 -0.007081             -0.092325             -0.797524      0  \n",
       "4                 -0.062516             -0.113203             -0.797524      0  \n",
       "...                     ...                   ...                   ...    ...  \n",
       "28329              0.513470              0.526327              3.423047      1  \n",
       "28330              0.682506              0.642059              3.423047      1  \n",
       "28331              1.095488              1.050245              3.423047      1  \n",
       "28332              0.929218              1.027704              3.423047      1  \n",
       "28333              0.598626              0.677700              3.423047      1  \n",
       "\n",
       "[28334 rows x 236 columns]"
      ]
     },
     "execution_count": 91,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_x_features_scaled"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "id": "8d70f219",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "138a51ee",
   "metadata": {},
   "source": [
    "### Divide into training and validation sets 70/30 split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "id": "f1ce0e7d",
   "metadata": {},
   "outputs": [],
   "source": [
    "training_set_x, validation_set_x = train_test_split(df_x_features_scaled, test_size = 0.3, random_state = 21)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "id": "84842cd7",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train_x = training_set_x.iloc[:,0:-1].values\n",
    "Y_train_x = training_set_x.iloc[:,-1].values\n",
    "X_val_x = validation_set_x.iloc[:,0:-1].values\n",
    "y_val_x = validation_set_x.iloc[:,-1].values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "id": "6fb66354",
   "metadata": {},
   "outputs": [],
   "source": [
    "def accuracy(confusion_matrix):\n",
    "   diagonal_sum = confusion_matrix.trace()\n",
    "   sum_of_all_elements = confusion_matrix.sum()\n",
    "   return diagonal_sum / sum_of_all_elements"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "id": "8db68771",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.neural_network import MLPClassifier"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2a82b305",
   "metadata": {},
   "source": [
    "### Train classifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "id": "be67320f",
   "metadata": {},
   "outputs": [],
   "source": [
    "classifier = MLPClassifier(hidden_layer_sizes=(150,100,50), max_iter=300,activation = 'relu',solver='adam',random_state=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "id": "25274ece",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Wall time: 8.07 s\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "MLPClassifier(hidden_layer_sizes=(150, 100, 50), max_iter=300, random_state=1)"
      ]
     },
     "execution_count": 98,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%%time\n",
    "classifier.fit(X_train_x, Y_train_x)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3d9a52f3",
   "metadata": {},
   "source": [
    "### Evaluate prediction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "id": "01ba4248",
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred_x = classifier.predict(X_val_x)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "id": "7a794184",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import confusion_matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "id": "e7dc2453",
   "metadata": {},
   "outputs": [],
   "source": [
    "cm = confusion_matrix(y_pred_x, y_val_x)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "id": "fccb0774",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy of MLPClassifier :  0.9869427126220445\n"
     ]
    }
   ],
   "source": [
    "print(\"Accuracy of MLPClassifier : \", accuracy(cm))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "id": "784b88a4",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[5624,   86],\n",
       "       [  25, 2766]], dtype=int64)"
      ]
     },
     "execution_count": 103,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cm"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}