notebooks/05_train_model.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "import itertools\n",
    "\n",
    "import matplotlib as mpl\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import seaborn as sns\n",
    "from LearningAlgorithms import ClassificationAlgorithms\n",
    "from sklearn.metrics import accuracy_score, confusion_matrix\n",
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "mpl.style.use(\"fivethirtyeight\")\n",
    "mpl.rcParams[\"figure.figsize\"] = (15, 8)\n",
    "mpl.rcParams[\"figure.dpi\"] = 50\n",
    "mpl.rcParams[\"lines.linewidth\"] = 2\n",
    "mpl.rcParams[\"font.size\"] = 12\n",
    "mpl.rcParams[\"figure.titlesize\"] = 25"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "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>acc_x</th>\n",
       "      <th>acc_y</th>\n",
       "      <th>acc_z</th>\n",
       "      <th>gyr_x</th>\n",
       "      <th>gyr_y</th>\n",
       "      <th>gyr_z</th>\n",
       "      <th>participant</th>\n",
       "      <th>label</th>\n",
       "      <th>category</th>\n",
       "      <th>set</th>\n",
       "      <th>...</th>\n",
       "      <th>gyr_r_pse</th>\n",
       "      <th>gyr_r_freq_0.0_Hz_ws_14</th>\n",
       "      <th>gyr_r_freq_0.357_Hz_ws_14</th>\n",
       "      <th>gyr_r_freq_0.714_Hz_ws_14</th>\n",
       "      <th>gyr_r_freq_1.071_Hz_ws_14</th>\n",
       "      <th>gyr_r_freq_1.429_Hz_ws_14</th>\n",
       "      <th>gyr_r_freq_1.786_Hz_ws_14</th>\n",
       "      <th>gyr_r_freq_2.143_Hz_ws_14</th>\n",
       "      <th>gyr_r_freq_2.5_Hz_ws_14</th>\n",
       "      <th>cluster</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>epoch (ms)</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2019-01-11 15:08:08.000</th>\n",
       "      <td>-0.170494</td>\n",
       "      <td>0.888970</td>\n",
       "      <td>-0.172353</td>\n",
       "      <td>-25.990313</td>\n",
       "      <td>-10.162423</td>\n",
       "      <td>25.426906</td>\n",
       "      <td>B</td>\n",
       "      <td>bench</td>\n",
       "      <td>heavy</td>\n",
       "      <td>28</td>\n",
       "      <td>...</td>\n",
       "      <td>0.401838</td>\n",
       "      <td>227.845540</td>\n",
       "      <td>-29.795388</td>\n",
       "      <td>18.195603</td>\n",
       "      <td>0.294594</td>\n",
       "      <td>-24.424951</td>\n",
       "      <td>11.515400</td>\n",
       "      <td>-51.920406</td>\n",
       "      <td>-13.567962</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2019-01-11 15:08:08.400</th>\n",
       "      <td>-0.027634</td>\n",
       "      <td>0.878027</td>\n",
       "      <td>-0.056321</td>\n",
       "      <td>-4.311919</td>\n",
       "      <td>-1.963210</td>\n",
       "      <td>3.088785</td>\n",
       "      <td>B</td>\n",
       "      <td>bench</td>\n",
       "      <td>heavy</td>\n",
       "      <td>28</td>\n",
       "      <td>...</td>\n",
       "      <td>0.366606</td>\n",
       "      <td>255.346459</td>\n",
       "      <td>-32.701710</td>\n",
       "      <td>-65.942225</td>\n",
       "      <td>-18.156214</td>\n",
       "      <td>5.435095</td>\n",
       "      <td>16.330497</td>\n",
       "      <td>4.416252</td>\n",
       "      <td>-1.494355</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2019-01-11 15:08:08.800</th>\n",
       "      <td>-0.060139</td>\n",
       "      <td>0.982341</td>\n",
       "      <td>-0.036019</td>\n",
       "      <td>3.691356</td>\n",
       "      <td>-1.273743</td>\n",
       "      <td>-2.492890</td>\n",
       "      <td>B</td>\n",
       "      <td>bench</td>\n",
       "      <td>heavy</td>\n",
       "      <td>28</td>\n",
       "      <td>...</td>\n",
       "      <td>0.432324</td>\n",
       "      <td>256.323480</td>\n",
       "      <td>-40.455694</td>\n",
       "      <td>-30.695845</td>\n",
       "      <td>3.361290</td>\n",
       "      <td>-10.888301</td>\n",
       "      <td>-51.138403</td>\n",
       "      <td>32.709146</td>\n",
       "      <td>-10.263861</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>3 rows × 121 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                            acc_x     acc_y     acc_z      gyr_x      gyr_y  \\\n",
       "epoch (ms)                                                                    \n",
       "2019-01-11 15:08:08.000 -0.170494  0.888970 -0.172353 -25.990313 -10.162423   \n",
       "2019-01-11 15:08:08.400 -0.027634  0.878027 -0.056321  -4.311919  -1.963210   \n",
       "2019-01-11 15:08:08.800 -0.060139  0.982341 -0.036019   3.691356  -1.273743   \n",
       "\n",
       "                             gyr_z participant  label category  set  ...  \\\n",
       "epoch (ms)                                                           ...   \n",
       "2019-01-11 15:08:08.000  25.426906           B  bench    heavy   28  ...   \n",
       "2019-01-11 15:08:08.400   3.088785           B  bench    heavy   28  ...   \n",
       "2019-01-11 15:08:08.800  -2.492890           B  bench    heavy   28  ...   \n",
       "\n",
       "                         gyr_r_pse  gyr_r_freq_0.0_Hz_ws_14  \\\n",
       "epoch (ms)                                                    \n",
       "2019-01-11 15:08:08.000   0.401838               227.845540   \n",
       "2019-01-11 15:08:08.400   0.366606               255.346459   \n",
       "2019-01-11 15:08:08.800   0.432324               256.323480   \n",
       "\n",
       "                         gyr_r_freq_0.357_Hz_ws_14  gyr_r_freq_0.714_Hz_ws_14  \\\n",
       "epoch (ms)                                                                      \n",
       "2019-01-11 15:08:08.000                 -29.795388                  18.195603   \n",
       "2019-01-11 15:08:08.400                 -32.701710                 -65.942225   \n",
       "2019-01-11 15:08:08.800                 -40.455694                 -30.695845   \n",
       "\n",
       "                         gyr_r_freq_1.071_Hz_ws_14  gyr_r_freq_1.429_Hz_ws_14  \\\n",
       "epoch (ms)                                                                      \n",
       "2019-01-11 15:08:08.000                   0.294594                 -24.424951   \n",
       "2019-01-11 15:08:08.400                 -18.156214                   5.435095   \n",
       "2019-01-11 15:08:08.800                   3.361290                 -10.888301   \n",
       "\n",
       "                         gyr_r_freq_1.786_Hz_ws_14  gyr_r_freq_2.143_Hz_ws_14  \\\n",
       "epoch (ms)                                                                      \n",
       "2019-01-11 15:08:08.000                  11.515400                 -51.920406   \n",
       "2019-01-11 15:08:08.400                  16.330497                   4.416252   \n",
       "2019-01-11 15:08:08.800                 -51.138403                  32.709146   \n",
       "\n",
       "                         gyr_r_freq_2.5_Hz_ws_14  cluster  \n",
       "epoch (ms)                                                 \n",
       "2019-01-11 15:08:08.000               -13.567962        1  \n",
       "2019-01-11 15:08:08.400                -1.494355        1  \n",
       "2019-01-11 15:08:08.800               -10.263861        1  \n",
       "\n",
       "[3 rows x 121 columns]"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# --------------------------------------------------------------\n",
    "# Load data\n",
    "# --------------------------------------------------------------\n",
    "\n",
    "df = pd.read_pickle(\"../data/interim/03_data_features.pkl\")\n",
    "\n",
    "df.head(3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# --------------------------------------------------------------\n",
    "# Create a training and test set\n",
    "# --------------------------------------------------------------\n",
    "\n",
    "df_train = df.drop(columns=[\"participant\", \"category\", \"set\"], axis=1)\n",
    "\n",
    "X = df_train.drop(columns=[\"label\"], axis=1)\n",
    "y = df_train[\"label\"]\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(\n",
    "    X, y, test_size=0.25, random_state=42, stratify=y\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 750x400 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot the distribution of the labels in the training set\n",
    "fig, ax = plt.subplots(figsize=(15, 8))\n",
    "df_train[\"label\"].value_counts().plot(\n",
    "    kind=\"bar\", ax=ax, color=\"lightblue\", label=\"Total\"\n",
    ")\n",
    "y_train.value_counts().plot(kind=\"bar\", ax=ax, color=\"dodgerblue\", label=\"Train\")\n",
    "y_test.value_counts().plot(kind=\"bar\", ax=ax, color=\"royalblue\", label=\"Test\")\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Basic features: 6\n",
      "Square features: 2\n",
      "PCA features: 3\n",
      "Time features: 16\n",
      "Frequency features: 88\n",
      "Cluster features: 1\n"
     ]
    }
   ],
   "source": [
    "# --------------------------------------------------------------\n",
    "# Split feature subsets\n",
    "# --------------------------------------------------------------\n",
    "\n",
    "basic_features = [\"acc_x\", \"acc_y\", \"acc_z\", \"gyr_x\", \"gyr_y\", \"gyr_z\"]\n",
    "square_features = [\"acc_r\", \"gyr_r\"]\n",
    "pca_features = [\"pca_1\", \"pca_2\", \"pca_3\"]\n",
    "time_features = [i for i in df_train.columns if \"temp\" in i]\n",
    "freq_features = [i for i in df_train.columns if (\"_freq\" in i) or (\"_pse\" in i)]\n",
    "cluster_features = [\"cluster\"]\n",
    "\n",
    "print(\"Basic features:\", len(basic_features))\n",
    "print(\"Square features:\", len(square_features))\n",
    "print(\"PCA features:\", len(pca_features))\n",
    "print(\"Time features:\", len(time_features))\n",
    "print(\"Frequency features:\", len(freq_features))\n",
    "print(\"Cluster features:\", len(cluster_features))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Creating feature sets\n",
    "feature_set_1 = list(set(basic_features))\n",
    "feature_set_2 = list(set(basic_features + square_features + pca_features))\n",
    "feature_set_3 = list(set(feature_set_2 + time_features))\n",
    "feature_set_4 = list(set(feature_set_3 + freq_features + cluster_features))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0\n",
      "1\n",
      "2\n",
      "3\n",
      "4\n",
      "5\n",
      "6\n",
      "7\n",
      "8\n",
      "9\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAd4AAADsCAYAAADaW8fRAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8pXeV/AAAACXBIWXMAAAexAAAHsQEGxWGGAAAodElEQVR4nO3de1hT9+E/8HdCuMolKEUuVpS26LQCKoqAFVEcBhSqztVnfmvpBVfbrrN1ne1039rN/Vy/25762OKltd+y7lZXdTJB6PpFCFqpYK1i61OcQ7w0QdBiuObkdn5/UFMRgihJTjDv1/P00ZyTfHgn9smbnM/J58iuXbsmgoiIiJxCLnUAIiIid8LiJSIiciIWLxERkROxeImIiJzIbYpXq9VKHeG2MK9jMa9jMa9jMa9jOTqv2xSvKA6tk7eZ17GY17GY17GY17EcnddtipeIiMgVsHiJiIiciMVLRETkRHYt3ra2NsydOxeRkZE4ffp0j30mkwmrVq2CSqXC2rVrrdvz8/ORkZGBRx55BDqdzp5xiIiIXI5di9fX1xe7du1CdnZ2r30lJSWIjIxESUkJurq6cPToUVy5cgWlpaUoLS3FkiVLsHPnTnvGISIicjkKuw6mUCAkJKTPfTU1NcjIyAAAzJ07F9XV1dDpdJg5cyZkMhnS09OxatUqe8YhckmiKMIkAiYLYBJFmL/902QBTJbufT22iSKadCKavAxSRx+wZuZ1KOZ1HB+FDIEO/hl2Ld7+6HQ6BAQEAACCgoLQ0tLSY1tgYCBaWlp6PU6r1drl1G69Xg+NRjPocZzlbs4riiLMImAW0V0yff3dYmP7jX9abjFGP/cRjCbIznw98HGsZWgjyw25b/WcLHf8Kjff8SOlwbyOxbyOMC4A+NsUwS7vvzKZDOHh4b22O614lUol2traAHSXcHBwMJRKJerr63tsu1lfoe+ERqNBRESEXcZyhqGQ12wRcahRwJ76LnyqFQEPjxs+sdn+VGd20lf6FDJAIQcUMhk8vv3z+m3RAvh4ekAhl0EhAzzk1/cBCrkMHt/+qZABnnIZfPsZy+OGx1nH6nH7Nu7/7b6bx7hypRn33HOPc144O2huZl5HYl7H8fWQwaez2aHvv04r3oSEBJSXlyMlJQVlZWX4r//6L0RHR2PLli1Yu3YtysrKkJiY6Kw4dIcsooiaJgN2n+tCYUMXmrosGBvggcRgICTQZ8DlcmOxfVd03xWUxw1FdfP9PeU3l1/vopLLun/btKX7F5swJ75yg6MxyBAxwkvqGAOmEZjXkZjXsTSdjh3f7sW7dOlSnDp1CmfPnkVubi5qamqwefNmqFQqFBcXQ6VSYdKkSZg+fToAYP78+cjIyIBSqcTbb79t7zhkB6IoovYbI/bUd2HvuS5c6jAjwk+OpdF++EG0L+JHeEKr1SIiIkjqqERELs/uxfvhhx/2uL18+fLuH6RQYPv27b3u/+yzz+LZZ5+1dwyygzPXjNh9rgt767twttWEEd5yPDzWF4vH+iJppBfk/XyiJCKivjntUDMNDefbTNh7rgt7znXhi2+MCPSUYUGUL16fEYTUcG8o5CxbIqLBYPESGjvN2NfQhT31nahpNsLXQwbVaB+8HB+A9Egf+ChYtkRE9sLidVPf6M3Yf16P3fWdONxogEIOpEf6YGeqP+bf6wN/T64mSkTkCCxeN9JmtODABT321nei7GsBFgCzwr2xJUWJhVG+UHqzbImIHI3Fe5frMon41yU99p7rxEcX9dCbgRmhXtiUGIScMb4I9fWQOiIRkVth8d6FjBYR5V8L2HOuEwcu6NFmFBE3whPrJgfi4bG+uNef/+xERFLhO/BdwmwR8cllA/bWd6LwfBdaBBExQQo896A/loz1xf1BnlJHJCIisHiHNFEUcazZiD3nOrHvXBcauywY7e+B3JhhWBLth4nBin5XbyIiIudj8Q4xoijiixYT9tZ3Ys+5LlxoNyPMt3thix9E+2FqiCfLlojIhbF4h4izOiP2fLuKVJ3OhGBvGXKifLE42g8pI73gwYUtiIiGBBavC7vYbsI/vl1F6uRVI/wVMmRF+eDX04KQFukNT5YtEdGQw+J1QQ1tJjxxVMTxa5fh4wFk3OuDF2MD8P1RPvDlKlJEREMai9cFvftVB+o7gB2zgqG61weBXlzYgojobsHidUEVGgGp9wCP3OcndRQiIrIzfpRyMVf1Zpz6xogZw6VOQkREjsDidTGHtAYAwIwREgchIiKHYPG6GLVWjweCFBjpw5OoiIjuRixeF6PWCJgd7i11DCIichAWrwu52G5CfZsZsyJYvEREdysWrwtRawXIZcBDYSxeIqK7FYvXhag1AuJGePKC9EREdzG+w7sIURRRqRWQyvldIqK7GovXRXx1zYTLXRYWLxHRXY7F6yLUWgFeciBxpJfUUYiIyIFYvC5CrREwPdQLfgr+kxAR3c34Lu8CTBYRnzQKmB3hI3UUIiJyMBavCzhx1YhWo8j5XSIiN8DidQEVGgGBnjJMDvGUOgoRETmYXYt3/fr1UKlUyMvLg8FgsG5vbW3FsmXLsGDBAqxbt866/bXXXsO8efMwb948HDt2zJ5RhhS1Ro/kMG8o5FyfmYjobme34j158iSamppQUlKCcePGobCw0LqvoKAAmZmZKCoqgl6vx7Fjx9DS0oLDhw/j448/xvbt2/HGG2/YK8qQ0mUSUd1s4GFmIiI3YbfirampQVpaGgAgPT0d1dXV1n0NDQ2IjY0FAMTFxaGqqgr+/v4YPnw4jEYjdDodQkJC7BVlSDnaJEAwA6lcn5mIyC0o7DWQTqdDWFgYACAwMBAtLS3WfTExMaisrER8fDwqKioQHR0NT09PjB8/HgkJCTAYDPjwww/7HFer1UIUxUHn0+v10Gg0gx7H3orOiBjhBQR1NkHT9d2hZlfNawvzOhbzOhbzOpa75pXJZAgPD++13W7Fq1Qq0dbWBqC7hIODg637VqxYgTVr1iAnJwejR49GaGgo6urqUFtbi+PHj+Py5cvIy8tDcXFxr3H7Cn0nNBoNIiIi7DKWPR3/rAlzRikQGTm8x3ZXzWsL8zoW8zoW8zoW8/Zkt0PNCQkJKC8vBwCUlZUhMTHRus/Pzw/btm2zzvtmZGQAAAICAuDh4QF/f390dHTYK8qQcU2w4MRVI2ZxfpeIyG3YrXjj4uIQGhoKlUqFuro6ZGdnY/Xq1QCA2tpaZGVlITs7G0lJSYiKisK4ceMQGRmJ+fPnIycnBz/72c/sFWXIONQowCJyfpeIyJ3Y7VAzAGzcuLHH7c2bNwMAYmNj+zyMvGnTJnv++CGnUiNgbIAHRvvb9Z+BiIhcGBfQkJCalwEkInI7LF6JaDrMOKMz8TAzEZGbYfFKpFIrAAAe4ideIiK3wuKViFor4MHhngjx8ZA6ChERORGLVwKiKKJSI2A2P+0SEbkdFq8Ezraa8HWnmfO7RERuiMUrAbVGgEIGJI30kjoKERE5GYtXAmqtgGmhXvD35MtPRORu+M7vZGaLiENagctEEhG5KRavk536xohrBpELZxARuSkWr5OptQL8FDIk3MP5XSIid8TidTK1RkDKSC94echufWciIrrrsHidSDCLqLpswCx+jYiIyG2xeJ2ousmALjPnd4mI3BmL14nUWgEjvOV4cLin1FGIiEgiLF4nqtQIeCjcG3IZ53eJiNwVi9dJWg0WfHbFwMPMRERujsXrJEcuCzCL4PrMRERujsXrJGqNgHv9PTA2gJcBJCJyZyxeJ1FrBaSGe0PG+V0iIrfG4nWCpi4zTreYOL9LREQsXmeo1AoAwAsjEBERi9cZ1BoB31MqMNKP87tERO6OxesEal4GkIiIvsXidbCGNhMutJv5NSIiIgLA4nU4tUaAXAakhLF4iYiIxetwaq2AqSGeCPLiS01ERCxeh7KIItQaAanhPlJHISIiF8HidaAvW0y4Klh4/V0iIrKya/GuX78eKpUKeXl5MBgM1u2tra1YtmwZFixYgHXr1lm3nz59GosXL0ZmZibee+89e0ZxCWqNHj4ewPR7vKSOQkRELsJuxXvy5Ek0NTWhpKQE48aNQ2FhoXVfQUEBMjMzUVRUBL1ej2PHjgEAXnvtNRQUFODAgQN4/PHH7RXFZVRqBcwY6Q0fBZeJJCKibjaLt7S0FBaLZcAD1dTUIC0tDQCQnp6O6upq676GhgbExsYCAOLi4lBVVYWGhgaYTCasXLkSixcvxpkzZ+70Obgko0XEkUZeBpCIiHpS2Nqxd+9evPzyy8jOzsby5csxbty4fgfS6XQICwsDAAQGBqKlpcW6LyYmBpWVlYiPj0dFRQWio6PR1NSEr776ClVVVbh06RLWrVuHDz/8sNe4Wq0Woije6fOz0uv10Gg0gx5noI63iGg3ARM8W6HRtN32452dd7CY17GY17GY17HcNa9MJkN4eHiv7TaL9+2330Zrayv27NmDZ599FjKZDMuXL8eSJUsQEBDQ6/5KpRJtbd0Fo9PpEBwcbN23YsUKrFmzBjk5ORg9ejRCQ0MRFBSE+Ph4BAYGYsKECbh69WqfOfoKfSc0Gg0iIiLsMtZA/KmpFUFe7UgfFw4P+e0fanZ23sFiXsdiXsdiXsdi3p76neMNDAxEdnY2Fi9ejMbGRhQVFSE1NRU7duzodd+EhASUl5cDAMrKypCYmGjd5+fnh23btlnnfTMyMnDffffh6tWrMBqN+Prrr/ss86GsQiPgoTDvOypdIiK6e9ks3pKSEixfvhwLFy6EyWTCwYMHsXv3bhw+fBhvvvlmr/vHxcUhNDQUKpUKdXV1yM7OxurVqwEAtbW1yMrKQnZ2NpKSkhAVFQWFQoFnnnkGCxcuxOOPP44NGzY46jk6XYfRgmPNBi4TSUREvdg81Lxv3z4888wzSElJ6bHdz88P+fn5fT5m48aNPW5v3rwZABAbG4vi4uJe98/OzkZ2dvbtZnZ5VZcNMFrAE6uIiKgXm8X7yiuvWE+WAoCuri40NTUhKioKqampTgk3VKm1AsL95HggyObLS0REbsrmoebc3FzI5d/tlsvlyM3NdUamIU+t6b4MoEzG+V0iIurJZvGaTCZ4eX234pK3t3eP1aiob9/ozTj1jZGHmYmIqE82izckJAQHDhyw3i4uLsaIESOcEmooO9RogAggNYIXRiAiot5sTkK+8cYbyMvLw89//nOIoojIyEhs377dmdmGpAqNHg8EKRA5zEPqKERE5IJsFu/YsWPxf//3f2hvb4coinfd92wdRa0RMCeSn3aJiKhv/Z52+9FHH+Grr76CXq+3blu7dq3DQw1VF9tNqG8zYwPnd4mIyAabxfvCCy+gs7MThw8fxqOPPop//vOfmDJlijOzDTlqrQAZgIdYvEREZIPNk6uOHj2KHTt2ICgoCC+//DL+9a9/4euvv3ZmtiGnUiMgboQngr3tepljIiK6i9hsCB+f7nlKPz8/aLVaeHp64vz5804LNtSIogi1VuDXiIiIqF82DzXPnz8f165dw09+8hOkpqZCJpNhxYoVzsw2pNTpTLjcZeH6zERE1K8+i9disSAtLQ1KpRI5OTmYP38+9Ho9goKCnJ1vyKjQCPCSAzNGet36zkRE5Lb6PNQsl8vxi1/8wnrb29ubpXsLao2A6aFe8FNwfpeIiGyz2RJpaWkoLCyEKIrOzDMkmSwiPmnk/C4REd2azTnerVu3oqOjAwqFAj4+PhBFETKZDBcvXnRmviHhxFUjWo0i53eJiOiWbBbvpUuXnJljSFNrBAR4yjAlhPO7RETUP5vF+8knn/S5PSUlxWFhhiq1VkBymDcUcl4GkIiI+mezeLds2WL9uyAIOH78OOLi4rB//36nBBsqukwijjYJeHUqTz4jIqJbs1m8u3bt6nH70qVL2LBhg6PzDDlHmwQIZmA253eJiGgABvzdl8jISJw+fdqRWYYktUZAqK8c31P2e70JIiIiAP184n3ppZcgk3XPWYqiiFOnTuHBBx90WrChQq0VMCvc2/paERER9cdm8U6ePPm7OykUWLJkCWbMmOGUUEPFNcGCE1eNyB03TOooREQ0RNgs3pycHPj4+MDDwwMAYDab0dnZCT8/P6eFc3WHGwVYRHDhDCIiGjCbc7w5OTno6uqy3u7s7EROTo5TQg0Vaq2AMQEeiArg/C4REQ2MzeLV6/Xw9/e33g4ICEBnZ6dTQg0VlRouE0lERLfHZvEOGzYMJ06csN4+ceIEDzPfQNNhRp3OxOIlIqLbYvMY6aZNm5Cbm4vw8HAAQGNjI9577z2nBXN1lVoBADCL398lIqLbYLN4p0yZgpqaGvz73/+GKIqIiYmBp6enM7O5NLVWwIPDPRHi4yF1FCIiGkJsHmp+55130NHRgQkTJmDixIno6OjAzp07+x1s/fr1UKlUyMvLg8FgsG5vbW3FsmXLsGDBAqxbt67HY6qrq6FUKtHe3j7Ip+I8oihyfpeIiO6IzeL94x//CKVSab2tVCrxxz/+0eZAJ0+eRFNTE0pKSjBu3DgUFhZa9xUUFCAzMxNFRUXQ6/U4duyYdd+OHTsQHx8/uGfhZP9pNeHrTjOLl4iIbpvN4rVYLBBF0XrbbDbDaDTaHKimpgZpaWkAgPT0dFRXV1v3NTQ0IDY2FgAQFxeHqqoqAEBVVRUmTpyIYcOG1gIUaq0AhQxIDuNlAImI6PbYnOOdO3cucnNz8cQTTwAA/vd//xfp6ek2B9LpdAgLCwMABAYGoqWlxbovJiYGlZWViI+PR0VFBaKjowEA27dvx1tvvYWysjKb42q12h6/ANwpvV4PjUYz6HEA4KN6EbFBQGtzI1rtMmJv9szrDMzrWMzrWMzrWO6aVyaTWU9QvpHN4n3ttddQUFCAd999F6IoYvLkyf0GUSqVaGtrA9BdwsHBwdZ9K1aswJo1a5CTk4PRo0cjNDQUhw8fxoMPPoiAgIB+g/cV+k5oNBpEREQMehyzRUR1ixYrJ/gjIiLQDsn6Zq+8zsK8jsW8jsW8jsW8Pdk81CyXyzFt2jRER0fj5MmTOHjwIGJiYmwOlJCQgPLycgBAWVkZEhMTrfv8/Pywbds267xvRkYGvvjiC6jVaixZsgRffvklnnvuOXs9J4c69Y0R1wwiZnN+l4iI7kCvT7xnz57Fnj17sHfvXgQHB2PRokWwWCwoKirqd6C4uDiEhoZCpVJh1KhReP7557F69Wps3rwZtbW1eOWVV+Dh4YFly5YhKioKTz/9NJ5++mkAQFZWFt566y3HPEM7U2sF+ClkSLiH87tERHT7ehXvtGnTkJSUhL/97W/WuditW7cOaLCNGzf2uL1582YAQGxsLIqLi20+rr99rkatEZA80gteHrwMIBER3b5eh5rff/99jBw5EgsXLsTzzz8PtVptl5Ob7gaCWUTVZQO/RkRERHesV/EuXLgQ7733HqqrqzFz5kxs3boVzc3NePHFF3Hw4EEpMrqMmmYDuswil4kkIqI71u9FEn74wx9i165dOH36NCZNmoQ33njDmdlcjlojYLi3HJOGc+lMIiK6MzaL90bBwcF4/PHHsX//fkfncWlqjYCHwr0gl3F+l4iI7syAipeAVoMFn10xYHa4j9RRiIhoCGPxDtCRywLMIpDK+V0iIhoEFu8AqTUCRg3zwNgAXgaQiIjuHIt3gNRaAakR3pBxfpeIiAaBxTsAzV1mnG4x8fu7REQ0aCzeAajUCgCAWSxeIiIaJBbvAFRoBIxXKhDmx/ldIiIaHBbvAKi1Aj/tEhGRXbB4b6GhzYQL7WbM5teIiIjIDli8t6DWCJDLgJQwFi8REQ0ei/cW1FoBU0I8EeTFl4qIiAaPbdIPiyiiUivwa0RERGQ3LN5+nG4x4Yregllcn5mIiOyExdsPtVaAjweQGOoldRQiIrpLsHj7odbokRjqDR8Fl4kkIiL7YPHaYLSIONJo4NeIiIjIrli8NnzWbEC7SeSJVUREZFcsXhvUWgFBXjLEjfCUOgoREd1FWLw2qDUCZoZ5w0PO+V0iIrIfFm8fOowW1DQbeJiZiIjsjsXbh0+bDDBagFSeWEVERHbG4u1DhUZAmK8cMUEKqaMQEdFdhsXbB7VGQGqEN2Qyzu8SEZF9sXhv8o3ejFPfGDm/S0REDmHXY6nr16/HZ599hlGjRiE/Px9eXt1LLba2tmLlypVob29HXFwcfvOb3+DixYt4+umnIYoi/Pz8sHPnTiiVSnvGuSOHGg0QAaRGcH1mIiKyP7t94j158iSamppQUlKCcePGobCw0LqvoKAAmZmZKCoqgl6vx7FjxxAQEIA///nPOHDgABYsWID333/fXlEGRa0RcH+gApHDPKSOQkREdyG7FW9NTQ3S0tIAAOnp6aiurrbua2hoQGxsLAAgLi4OVVVVUCqVCA4OBgAoFAp4eLhG0am1ep7NTEREDmO3Q806nQ5hYWEAgMDAQLS0tFj3xcTEoLKyEvHx8aioqEB0dHSPxxUUFGD37t19jqvVaiGK4qDz6fV6aDSafu+j6RLxn1bgJ2M7oNF0DvpnDsZA8roS5nUs5nUs5nUsd80rk8kQHh7ea7vdilepVKKtrQ1Ad5le/zQLACtWrMCaNWuQk5OD0aNHIzQ0FABgNBqRl5eHX//61zbnd/sKfSc0Gg0iIiL6vc/Bf3dAhmvImRiOYG9pzzsbSF5XwryOxbyOxbyOxbw92a1dEhISUF5eDgAoKytDYmKidZ+fnx+2bdtmnffNyMgAAKxZswYPP/wwkpKS7BVjUCo1AmJHeEpeukREdPeyW8PExcUhNDQUKpUKdXV1yM7OxurVqwEAtbW1yMrKQnZ2NpKSkhAVFYXq6mrs3r0bf/nLX5CVlYVt27bZK8odEUURaq2A2fwaEREROZBdv060cePGHrc3b94MAIiNjUVxcXGPfdOnT3epY/51OhMud1l4YhURETkUj6l+S60R4CUHZoz0kjoKERHdxVi831JrBUwL9YKfgi8JERE5DlsGgMki4nCjwGUiiYjI4Vi8AE5cNaLVILJ4iYjI4Vi86J7f9VfIMOUezu8SEZFjsXjRPb+bEu4NTzkvA0hERI7l9sXbZRJxtInzu0RE5BxuX7zVTQIEM1i8RETkFG5fvGqtgHt85JgQbNe1RIiIiPrE4tUImBXuDZmM87tEROR4bl281wQLPr9q5DKRRETkNG5dvIcbBVhEYBbnd4mIyEncunjVWgFjAjwwJoDzu0RE5BxuXbyVGn6NiIiInMtti1fbaUadzsTiJSIip3Lb4q3UCgCAWTyxioiInMhti1etETAxWIEQHw+poxARkRtxy+IVRRFqjcCvERERkdO5ZfH+p9WErzvNSA33kToKERG5GbcsXrVWgEIGJIfxMoBERORc7lm8GgEJ93ghwNMtnz4REUnI7ZrHIoo41CjwbGYiIpKE2xVv7VUjWgSR398lIiJJuF3xVmoF+ClkmHYP53eJiMj53K54KzQCkkZ6wcuDlwEkIiLnc6viFcwiqi4beJiZiIgk41bFW9NsQJdZ5GUAiYhIMm5VvGqNgGBvGWJHeEodhYiI3JRbFW+lVsCscG/IZZzfJSIiadi1eNevXw+VSoW8vDwYDAbr9tbWVixbtgwLFizAunXrrNvz8/ORkZGBRx55BDqdzp5Reukwifis2cBlIomISFJ2K96TJ0+iqakJJSUlGDduHAoLC637CgoKkJmZiaKiIuj1ehw7dgxXrlxBaWkpSktLsWTJEuzcudNeUfp0rAUwieCJVUREJCm7FW9NTQ3S0tIAAOnp6aiurrbua2hoQGxsLAAgLi4OVVVVOH78OGbOnAmZTNbr/o5QdRUYNcwD0YG8DCAREUlHYa+BdDodwsLCAACBgYFoaWmx7ouJiUFlZSXi4+NRUVGB6Oho6HQ6BAQE9Hn/G2m1WoiiOOh8R65YkBDUPd5QoNfrodFopI4xYMzrWMzrWMzrWO6aVyaTITw8vNd2uxWvUqlEW1sbgO4SDg4Otu5bsWIF1qxZg5ycHIwePRqhoaFQKpWor6/v8/436iv07WruMuNsRyNemhqMiAi/QY/nDBqNBhEREVLHGDDmdSzmdSzmdSzm7cluh5oTEhJQXl4OACgrK0NiYqJ1n5+fH7Zt22ad983IyMDkyZNx+PDhPu9vb50mEQ9HcH6XiIikZ7fijYuLQ2hoKFQqFerq6pCdnY3Vq1cDAGpra5GVlYXs7GwkJSUhKioKISEhmD9/PjIyMrBnzx48+eST9orSS1SAAv9vkgxhfpzfJSIiacmuXbs2+AnUIYCHOhyLeR2LeR2LeR2LeXtyqwU0iIiIpMbiJSIiciIWLxERkROxeImIiJyIxUtEROREblO8siF2RSLmdSzmdSzmdSzmdSxH53WbrxMRERG5Arf5xEtEROQKWLxEREROxOIlIiJyIrco3ra2NsydOxeRkZE4ffq01HH6deLECahUKmRmZiI3NxdGo1HqSP06ffo0MjIykJmZiR/+8Idob2+XOtKA7N69G/fdd5/UMW7p/PnzuO+++5CVlYWsrCxcuXJF6ki3dOjQIWRnZyMzMxPFxcVSx+nXZ599Zn1tExIS8Morr0gdqV8WiwVPP/00VCoVVCoVzp07J3WkfpnNZuTl5SErKwurVq1yyfezvvph7969mDdvHhYuXIhLly7Z/We6RfH6+vpi165dyM7OljrKLYWHh2PPnj04cOAAoqOjXf6N64EHHsBHH32EAwcOYMqUKSgqKpI60i1ZLBYUFhYiMjJS6igDkpKSguLiYhQXFyMkJETqOP3S6/V46623sHv3bhw4cABZWVlSR+rX1KlTra9tUlKSy+etra2FwWBASUkJfv7zn+Ptt9+WOlK/9u/fj6ioKBQXF2P8+PHYv3+/1JF6ubkfjEYj8vPzUVxcjHXr1uF3v/ud3X+mWxSvQqFw+Tes60aOHAk/v+5rBisUCnh4uPYVlTw9Pa1/7+rqQkxMjIRpBubDDz9ETk4O5PKh8b//0aNHoVKp8Ktf/Qqi6NpfQjh69Ch8fX2xbNkyLF++HJcvX5Y60oCYTCYcO3YMycnJUkfp1/VfFkVRhE6nc/n3tYaGBkyaNAlA9xXsqqqqJE7U28398J///Afjx4+Hl5cXZsyY4ZCjpEPjnccNXbx4EWq1GvPnz5c6yi2Vl5fjoYcewqFDhzB27Fip4/TLbDbjH//4BxYvXix1lAEJCwvD8ePHceDAATQ3N7vkJ4YbNTc34/z58/jggw+Qm5uL3/72t1JHGpDKykqkpKS4/C9jI0aMgFwux/Tp0/Hf//3f+NGPfiR1pH7FxMTg0KFDAICKigrodDqJE92aTqdDQECA9bbZbLb7z3Dt/8vcVGtrK3784x8jPz+/xydKV5WWloZDhw4hJycHBQUFUsfp165du7Bo0SKXf4O9ztvbG8OGDYNMJkN2djZOnToldaR+BQUFYcaMGfDy8sKsWbPw1VdfSR1pQPbt24ecnBypY9xSWVkZfHx8UFNTg/fffx/r1q2TOlK/5s+fD09PTyxYsAAdHR0IDQ2VOtItKZVKtLW1WW874qjj0Hj3cSPXT0ZYu3Yt7r//fqnj3JIgCNa/BwYGWg+Tu6q6ujp88MEHWLJkCerr613+ZJob3wCOHDmC6OhoCdPc2tSpU1FXVwegez5yzJgx0gYaAJPJhJqaGqSkpEgdZUCUSqX1T1f/BCmXy7Fp0yYUFRVh+PDhLj+HDgDR0dGoq6uDwWDAp59+iokTJ9r9Z7jNylVLly7FqVOncO+99yI3NxfLly+XOlKfdu/ejZdeegkTJkwAADz55JMufVi0tLQUW7ZsgVwuR0hICLZu3ery5Xvd7NmzUVFRIXWMfn388cfYuHEj/Pz8MHr0aOTn50OhUEgdq1/vvPMO9u7dC7lcjvz8fJcv3/LychQVFeEPf/iD1FFuyWw248c//jE0Gg0MBgN+85vfIDExUepYNl2+fBlPPPEEFAoFZs+ejRdeeEHqSH26uR98fHywfft2eHt7Y/v27Rg1apRdf57bFC8REZEr4KFmIiIiJ2LxEhERORGLl4iIyIlYvERERE7E4iUaBKVS2eO7lG+++SY2bdpkl7FXrVqFwsJCu4zVn3379mH69OlYsGBBj+3nz59HWFgYZs6caf3PYDDc9vh/+ctfoNVq7RWXaMhj8RINgre3N/bv34+rV69KHaWH21lt509/+hN+//vf97nO9tixY3H48GHrf15eXred5a9//SsaGxtv6zGOWC2IyFWweIkGQaFQIDc3F1u3bu217+ZPrNfX2T106JD16lNTp07Fhg0b8Pe//x1z5sxBcnJyjyvOVFRUQKVSYerUqSgtLQXQXUq//OUvkZaWhuTkZLz33nvWcRcsWICnnnoKM2fO7JVn9+7dSE5ORlJSEl599VUAwOuvv45PP/0UL774In75y18O6DkfPHgQ8+bNw6xZs/DYY49Zr0j1+uuvIy0tDUlJSfjpT38KURRRWFiIEydOIC8vDzNnzkRXVxcmTZpk/UXl888/ty6qsGnTJvz0pz/FokWLsGrVKly5cgWPPvoo0tLSkJaWhk8//RQAcPjwYesn8IceeqjHIiNEQ4FrfxOfaAh46qmnkJKSgueff37Aj/niiy9QXV2N4OBgxMfH49FHH8XBgwexbds27Nixw7rG8YULF1BcXIwLFy4gMzMTs2fPxgcffIDAwECUl5dDEARkZGQgLS0NAHD8+HEcOXKk16IVWq0WGzZsQEVFBZRKJRYtWoSioiKsXbsWlZWV2LhxIyZPntwr57lz56wlPmPGDLzyyiv43e9+h3379mHYsGHYvHkz8vPzsXbtWqxcuRJr164FAKxcuRKlpaXIycnB22+/bXP8m504cQKlpaXw9fXFU089hWeeeQZJSUm4ePEilixZgurqarz55pv4/e9/jxkzZqC9vR0+Pj4Dft2JXAGLl2iQAgMDsWzZMuzYsQO+vr4DesyUKVMQFhYGABgzZgzmzJkDAJgwYYJ1UXkAePjhhyGXyzFmzBhERUXhzJkzOHjwIL788kvrp+nW1lbU19fD09MTU6ZM6XOlqOPHjyMlJcV6FZalS5fiyJEjveZ1b3b9UPN1paWlqKurQ0ZGBoDuS6hNmzYNQPeFBrZs2YKuri60tLTge9/7HlQq1YBej+tUKpX1NayoqOix1nNbWxva2towY8YM/OIXv8DSpUuRnZ0Nf3//2/oZRFJj8RLZwTPPPINZs2b1WIpUoVDAYrEA6L6M240nJnl7e1v/LpfLrbflcnmP+U2ZTNbj58hkMoiiiP/5n//B3Llze+w7dOgQhg0b1mc+e11OUBRFpKWl4d133+2xXa/X42c/+xnKy8sxatQobNq0CXq9vs8xbnxdbr7PjfktFgs+/vjjXr/MvPDCC/j+97+Pjz/+GPPmzcO+ffuGxOUoia7jHC+RHQQHB2PRokX485//bN02evRonDhxAgBQXFwMo9F42+MWFhbCYrGgoaEBFy5cwAMPPIC5c+fi3XfftY539uxZdHR09DtOQkICPvnkE1y9ehVmsxl79uy5o4sCTJs2DUePHkV9fT0AoLOzE2fPnrUW6IgRI9De3o5//vOf1sf4+/v3mIe98XW58X43mzNnDt555x3r7draWgDdh78nTpyI1atXIz4+HmfOnLnt50EkJRYvkZ0899xzPc5ufuyxx/DJJ59gzpw5+Pzzz21+Gu3P/fffj6ysLPzgBz/AH/7wB/j4+GDFihUYP348UlNTkZSUhNWrV8NkMvU7TlhYGF599VUsXLgQM2fORFxc3B1dKSYkJAT5+fl48sknkZycjPT0dJw5cwZKpRKPPfYYkpOTsXz58h7zuT/60Y/w4osvWk+uWrt2LV5++WWoVKp+L3v5+uuv4/PPP0dycjISExOtJ5Ft3boVSUlJSElJga+vL+bNm3fbz4NISrxIAhERkRPxEy8REZETsXiJiIiciMVLRETkRCxeIiIiJ2LxEhEROdH/BxqDRxb7gsPZAAAAAElFTkSuQmCC",
      "text/plain": [
       "<Figure size 500x250 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# --------------------------------------------------------------\n",
    "# Perform forward feature selection using simple decision tree\n",
    "# --------------------------------------------------------------\n",
    "\n",
    "learner = ClassificationAlgorithms()\n",
    "max_features = 10\n",
    "\n",
    "selected_features, ordered_features, ordered_scores = learner.forward_selection(\n",
    "    max_features, X_train, y_train\n",
    ")\n",
    "\n",
    "# Plot the feature importance\n",
    "plt.figure(figsize=(10, 5))\n",
    "plt.plot(np.arange(1, max_features + 1, 1), ordered_scores)\n",
    "plt.xlabel(\"Number of Features\")\n",
    "plt.ylabel(\"Accuracy\")\n",
    "plt.xticks(np.arange(1, max_features + 1, 1))\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Feature set of selected feature in forward selection\n",
    "selected_features = [\n",
    "    \"acc_z_freq_0.0_Hz_ws_14\",\n",
    "    \"duration\",\n",
    "    \"acc_x_freq_0.0_Hz_ws_14\",\n",
    "    \"acc_y\",\n",
    "    \"gyr_r_freq_0.0_Hz_ws_14\",\n",
    "    \"gyr_x_temp_std_ws_5\",\n",
    "    \"gyr_z_freq_2.5_Hz_ws_14\",\n",
    "    \"acc_y_freq_2.143_Hz_ws_14\",\n",
    "    \"acc_r\",\n",
    "    \"gyr_r_freq_weighted\",\n",
    "]\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "# --------------------------------------------------------------\n",
    "# Grid search for best hyperparameters and model selection\n",
    "# --------------------------------------------------------------\n",
    "\n",
    "possible_feature_sets = [\n",
    "    feature_set_1,\n",
    "    feature_set_2,\n",
    "    feature_set_3,\n",
    "    feature_set_4,\n",
    "    selected_features,\n",
    "]\n",
    "\n",
    "feature_names = [\n",
    "    \"Feature set 1\",\n",
    "    \"Feature set 2\",\n",
    "    \"Feature set 3\",\n",
    "    \"Feature set 4\",\n",
    "    \"Selected features\",\n",
    "]\n",
    "\n",
    "iterations = 1  # Just one iteration\n",
    "score_df = pd.DataFrame()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Feature set: 0\n",
      "\tTraining neural network, 0\n",
      "\tTraining random forest, 0\n",
      "\tTraining KNN\n",
      "\tTraining decision tree\n",
      "\tTraining naive bayes\n",
      "Feature set: 1\n",
      "\tTraining neural network, 0\n",
      "\tTraining random forest, 0\n",
      "\tTraining KNN\n",
      "\tTraining decision tree\n",
      "\tTraining naive bayes\n",
      "Feature set: 2\n",
      "\tTraining neural network, 0\n",
      "\tTraining random forest, 0\n",
      "\tTraining KNN\n",
      "\tTraining decision tree\n",
      "\tTraining naive bayes\n",
      "Feature set: 3\n",
      "\tTraining neural network, 0\n",
      "\tTraining random forest, 0\n",
      "\tTraining KNN\n",
      "\tTraining decision tree\n",
      "\tTraining naive bayes\n",
      "Feature set: 4\n",
      "\tTraining neural network, 0\n",
      "\tTraining random forest, 0\n",
      "\tTraining KNN\n",
      "\tTraining decision tree\n",
      "\tTraining naive bayes\n"
     ]
    }
   ],
   "source": [
    "# This code performs a systematic evaluation of different feature sets using\n",
    "# multiple classifiers. The goal is to compare the performance of these classifiers for\n",
    "# various combinations of features and identify which feature set works best with each\n",
    "# classifier. The results are saved in the DataFrame score_df, which can be further\n",
    "# analyzed or visualized to make decisions about feature selection and classifier choices.\n",
    "for i, f in zip(range(len(possible_feature_sets)), feature_names):\n",
    "    print(\"Feature set:\", i)\n",
    "    selected_train_X = X_train[possible_feature_sets[i]].values\n",
    "    selected_test_X = X_test[possible_feature_sets[i]].values\n",
    "\n",
    "    # We run the classifiers for a number of iterations to get an average score.\n",
    "\n",
    "    # First run non deterministic classifiers to average their score.\n",
    "    performance_test_nn = 0\n",
    "    performance_test_rf = 0\n",
    "\n",
    "    for it in range(0, iterations):\n",
    "        print(\"\\tTraining neural network,\", it)\n",
    "        (\n",
    "            class_train_y,\n",
    "            class_test_y,\n",
    "            class_train_prob_y,\n",
    "            class_test_prob_y,\n",
    "        ) = learner.feedforward_neural_network(\n",
    "            selected_train_X,\n",
    "            y_train,\n",
    "            selected_test_X,\n",
    "            gridsearch=False,\n",
    "        )\n",
    "        performance_test_nn += accuracy_score(y_test, class_test_y)\n",
    "\n",
    "        print(\"\\tTraining random forest,\", it)\n",
    "        (\n",
    "            class_train_y,\n",
    "            class_test_y,\n",
    "            class_train_prob_y,\n",
    "            class_test_prob_y,\n",
    "        ) = learner.random_forest(\n",
    "            selected_train_X, y_train, selected_test_X, gridsearch=True\n",
    "        )\n",
    "        performance_test_rf += accuracy_score(y_test, class_test_y)\n",
    "\n",
    "    performance_test_nn = performance_test_nn / iterations\n",
    "    performance_test_rf = performance_test_rf / iterations\n",
    "\n",
    "    # And we run our deterministic classifiers:\n",
    "    print(\"\\tTraining KNN\")\n",
    "    (\n",
    "        class_train_y,\n",
    "        class_test_y,\n",
    "        class_train_prob_y,\n",
    "        class_test_prob_y,\n",
    "    ) = learner.k_nearest_neighbor(\n",
    "        selected_train_X, y_train, selected_test_X, gridsearch=True\n",
    "    )\n",
    "    performance_test_knn = accuracy_score(y_test, class_test_y)\n",
    "\n",
    "    print(\"\\tTraining decision tree\")\n",
    "    (\n",
    "        class_train_y,\n",
    "        class_test_y,\n",
    "        class_train_prob_y,\n",
    "        class_test_prob_y,\n",
    "    ) = learner.decision_tree(\n",
    "        selected_train_X, y_train, selected_test_X, gridsearch=True\n",
    "    )\n",
    "    performance_test_dt = accuracy_score(y_test, class_test_y)\n",
    "\n",
    "    print(\"\\tTraining naive bayes\")\n",
    "    (\n",
    "        class_train_y,\n",
    "        class_test_y,\n",
    "        class_train_prob_y,\n",
    "        class_test_prob_y,\n",
    "    ) = learner.naive_bayes(selected_train_X, y_train, selected_test_X)\n",
    "\n",
    "    performance_test_nb = accuracy_score(y_test, class_test_y)\n",
    "\n",
    "    # Save results to dataframe\n",
    "    models = [\"NN\", \"RF\", \"KNN\", \"DT\", \"NB\"]\n",
    "    new_scores = pd.DataFrame(\n",
    "        {\n",
    "            \"model\": models,\n",
    "            \"feature_set\": f,\n",
    "            \"accuracy\": [\n",
    "                performance_test_nn,\n",
    "                performance_test_rf,\n",
    "                performance_test_knn,\n",
    "                performance_test_dt,\n",
    "                performance_test_nb,\n",
    "            ],\n",
    "        }\n",
    "    )\n",
    "    score_df = pd.concat([score_df, new_scores])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "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>model</th>\n",
       "      <th>feature_set</th>\n",
       "      <th>accuracy</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>DT</td>\n",
       "      <td>Selected features</td>\n",
       "      <td>0.996587</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>RF</td>\n",
       "      <td>Selected features</td>\n",
       "      <td>0.996587</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>RF</td>\n",
       "      <td>Feature set 4</td>\n",
       "      <td>0.994312</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>NN</td>\n",
       "      <td>Feature set 4</td>\n",
       "      <td>0.989761</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>DT</td>\n",
       "      <td>Feature set 4</td>\n",
       "      <td>0.984073</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>NN</td>\n",
       "      <td>Feature set 3</td>\n",
       "      <td>0.984073</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>RF</td>\n",
       "      <td>Feature set 3</td>\n",
       "      <td>0.981797</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>NN</td>\n",
       "      <td>Selected features</td>\n",
       "      <td>0.981797</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>NB</td>\n",
       "      <td>Selected features</td>\n",
       "      <td>0.972696</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>KNN</td>\n",
       "      <td>Feature set 4</td>\n",
       "      <td>0.972696</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>NB</td>\n",
       "      <td>Feature set 4</td>\n",
       "      <td>0.968146</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>DT</td>\n",
       "      <td>Feature set 3</td>\n",
       "      <td>0.957907</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>RF</td>\n",
       "      <td>Feature set 2</td>\n",
       "      <td>0.953356</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>RF</td>\n",
       "      <td>Feature set 1</td>\n",
       "      <td>0.953356</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>NB</td>\n",
       "      <td>Feature set 3</td>\n",
       "      <td>0.938567</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>KNN</td>\n",
       "      <td>Selected features</td>\n",
       "      <td>0.932878</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>NN</td>\n",
       "      <td>Feature set 2</td>\n",
       "      <td>0.928328</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>NN</td>\n",
       "      <td>Feature set 1</td>\n",
       "      <td>0.926052</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>DT</td>\n",
       "      <td>Feature set 2</td>\n",
       "      <td>0.916951</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>KNN</td>\n",
       "      <td>Feature set 3</td>\n",
       "      <td>0.916951</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>DT</td>\n",
       "      <td>Feature set 1</td>\n",
       "      <td>0.908987</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>NB</td>\n",
       "      <td>Feature set 1</td>\n",
       "      <td>0.862344</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>NB</td>\n",
       "      <td>Feature set 2</td>\n",
       "      <td>0.854380</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>KNN</td>\n",
       "      <td>Feature set 1</td>\n",
       "      <td>0.775882</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>KNN</td>\n",
       "      <td>Feature set 2</td>\n",
       "      <td>0.773606</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  model        feature_set  accuracy\n",
       "3    DT  Selected features  0.996587\n",
       "1    RF  Selected features  0.996587\n",
       "1    RF      Feature set 4  0.994312\n",
       "0    NN      Feature set 4  0.989761\n",
       "3    DT      Feature set 4  0.984073\n",
       "0    NN      Feature set 3  0.984073\n",
       "1    RF      Feature set 3  0.981797\n",
       "0    NN  Selected features  0.981797\n",
       "4    NB  Selected features  0.972696\n",
       "2   KNN      Feature set 4  0.972696\n",
       "4    NB      Feature set 4  0.968146\n",
       "3    DT      Feature set 3  0.957907\n",
       "1    RF      Feature set 2  0.953356\n",
       "1    RF      Feature set 1  0.953356\n",
       "4    NB      Feature set 3  0.938567\n",
       "2   KNN  Selected features  0.932878\n",
       "0    NN      Feature set 2  0.928328\n",
       "0    NN      Feature set 1  0.926052\n",
       "3    DT      Feature set 2  0.916951\n",
       "2   KNN      Feature set 3  0.916951\n",
       "3    DT      Feature set 1  0.908987\n",
       "4    NB      Feature set 1  0.862344\n",
       "4    NB      Feature set 2  0.854380\n",
       "2   KNN      Feature set 1  0.775882\n",
       "2   KNN      Feature set 2  0.773606"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "score_df.sort_values(by=[\"accuracy\"], ascending=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 600x400 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# --------------------------------------------------------------\n",
    "# Create a grouped bar plot to compare the results\n",
    "# --------------------------------------------------------------\n",
    "\n",
    "# Plotting to see how much each feature contributed to the performance of the model\n",
    "fig, ax = plt.subplots(figsize=(12, 8))\n",
    "sns.barplot(\n",
    "    data=score_df,\n",
    "    x=\"model\",\n",
    "    y=\"accuracy\",\n",
    "    hue=\"feature_set\",\n",
    "    ax=ax,\n",
    ")\n",
    "plt.ylim(0.7, 1)\n",
    "plt.legend(loc=\"lower left\")\n",
    "plt.show()\n",
    "\n",
    "# Selected features and feature set 4 perform better "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training random forest (Selected features)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "0.9977246871444824"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# --------------------------------------------------------------\n",
    "# Select best model and evaluate results (Randon Forest)\n",
    "# --------------------------------------------------------------\n",
    "\n",
    "print(\"Training random forest (Selected features)\")\n",
    "(\n",
    "    class_train_y,\n",
    "    class_test_y,\n",
    "    class_train_prob_y,\n",
    "    class_test_prob_y,\n",
    ") = learner.random_forest(\n",
    "    X_train[selected_features], y_train, X_test[selected_features], gridsearch=True\n",
    ")\n",
    "\n",
    "accuracy = accuracy_score(y_test, class_test_y)\n",
    "\n",
    "accuracy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 600x600 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "classes = class_test_prob_y.columns  # Get the classes that were predicted by the model\n",
    "cm = confusion_matrix(y_test, class_test_y, labels=classes)  # Get the confusion matrix\n",
    "\n",
    "# Create confusion matrix for cm\n",
    "plt.figure(figsize=(12, 12))\n",
    "plt.imshow(cm, interpolation=\"nearest\", cmap=plt.cm.Blues)\n",
    "plt.title(\"Confusion matrix\")\n",
    "plt.colorbar()\n",
    "tick_marks = np.arange(len(classes))\n",
    "plt.xticks(tick_marks, classes, rotation=45)\n",
    "plt.yticks(tick_marks, classes)\n",
    "\n",
    "thresh = cm.max() / 2.0\n",
    "for i, j in itertools.product(range(cm.shape[0]), range(cm.shape[1])):\n",
    "    plt.text(\n",
    "        j,\n",
    "        i,\n",
    "        format(cm[i, j]),\n",
    "        horizontalalignment=\"center\",\n",
    "        color=\"white\" if cm[i, j] > thresh else \"black\",\n",
    "    )\n",
    "plt.ylabel(\"True label\")\n",
    "plt.xlabel(\"Predicted label\")\n",
    "plt.grid(False)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "# --------------------------------------------------------------\n",
    "# Select train and test data based on participant\n",
    "# --------------------------------------------------------------\n",
    "\n",
    "# Create a participant dataframe\n",
    "participant_df = df.drop([\"set\", \"category\"], axis=1)\n",
    "\n",
    "# Removing the participant A, as it is not used for training\n",
    "X_train = participant_df[participant_df[\"participant\"] != \"A\"].drop(\n",
    "    [\"label\", \"participant\"], axis=1\n",
    ")\n",
    "y_train = participant_df[participant_df[\"participant\"] != \"A\"][\"label\"]\n",
    "\n",
    "X_test = participant_df[participant_df[\"participant\"] == \"A\"].drop(\n",
    "    [\"label\", \"participant\"], axis=1\n",
    ")\n",
    "y_test = participant_df[participant_df[\"participant\"] == \"A\"][\"label\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 750x400 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot the distribution of the labels in the training set\n",
    "fig, ax = plt.subplots(figsize=(15, 8))\n",
    "df_train[\"label\"].value_counts().plot(\n",
    "    kind=\"bar\", ax=ax, color=\"lightblue\", label=\"Total\"\n",
    ")\n",
    "y_train.value_counts().plot(kind=\"bar\", ax=ax, color=\"dodgerblue\", label=\"Train\")\n",
    "y_test.value_counts().plot(kind=\"bar\", ax=ax, color=\"royalblue\", label=\"Test\")\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training random forest (Selected features) - Participant 'A' dropped to use it as test\n",
      "0.9982158786797503\n"
     ]
    }
   ],
   "source": [
    "# --------------------------------------------------------------\n",
    "# Use best model again and evaluate results\n",
    "# --------------------------------------------------------------\n",
    "\n",
    "print(\n",
    "    \"Training random forest (Selected features) - Participant 'A' dropped to use it as test\"\n",
    ")\n",
    "(\n",
    "    class_train_y,\n",
    "    class_test_y,\n",
    "    class_train_prob_y,\n",
    "    class_test_prob_y,\n",
    ") = learner.random_forest(\n",
    "    X_train[selected_features], y_train, X_test[selected_features], gridsearch=True\n",
    ")\n",
    "\n",
    "accuracy = accuracy_score(y_test, class_test_y)\n",
    "print(accuracy)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 600x600 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Create confusion matrix for cm\n",
    "plt.figure(figsize=(12, 12))\n",
    "plt.imshow(cm, interpolation=\"nearest\", cmap=plt.cm.Blues)\n",
    "plt.title(\"Confusion matrix\")\n",
    "plt.colorbar()\n",
    "tick_marks = np.arange(len(classes))\n",
    "plt.xticks(tick_marks, classes, rotation=45)\n",
    "plt.yticks(tick_marks, classes)\n",
    "\n",
    "thresh = cm.max() / 2.0\n",
    "for i, j in itertools.product(range(cm.shape[0]), range(cm.shape[1])):\n",
    "    plt.text(\n",
    "        j,\n",
    "        i,\n",
    "        format(cm[i, j]),\n",
    "        horizontalalignment=\"center\",\n",
    "        color=\"white\" if cm[i, j] > thresh else \"black\",\n",
    "    )\n",
    "plt.ylabel(\"True label\")\n",
    "plt.xlabel(\"Predicted label\")\n",
    "plt.grid(False)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Decision Tree (Selected features) - Participant 'A' dropped to use it as test\n",
      "0.9589652096342551\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 600x600 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# --------------------------------------------------------------\n",
    "# Try a simpler model with the selected features\n",
    "# --------------------------------------------------------------\n",
    "\n",
    "print(\n",
    "    \"Training Decision Tree (Selected features) - Participant 'A' dropped to use it as test\"\n",
    ")\n",
    "(\n",
    "    class_train_y,\n",
    "    class_test_y,\n",
    "    class_train_prob_y,\n",
    "    class_test_prob_y,\n",
    ") = learner.decision_tree(\n",
    "    X_train[selected_features], y_train, X_test[selected_features], gridsearch=True\n",
    ")\n",
    "\n",
    "accuracy = accuracy_score(y_test, class_test_y)\n",
    "print(accuracy)\n",
    "\n",
    "classes = class_test_prob_y.columns  # Get the classes that were predicted by the model\n",
    "cm = confusion_matrix(y_test, class_test_y, labels=classes)  # Get the confusion matrix\n",
    "\n",
    "# Create confusion matrix for cm\n",
    "plt.figure(figsize=(12, 12))\n",
    "plt.imshow(cm, interpolation=\"nearest\", cmap=plt.cm.Blues)\n",
    "plt.title(\"Confusion matrix\")\n",
    "plt.colorbar()\n",
    "tick_marks = np.arange(len(classes))\n",
    "plt.xticks(tick_marks, classes, rotation=45)\n",
    "plt.yticks(tick_marks, classes)\n",
    "\n",
    "thresh = cm.max() / 2.0\n",
    "for i, j in itertools.product(range(cm.shape[0]), range(cm.shape[1])):\n",
    "    plt.text(\n",
    "        j,\n",
    "        i,\n",
    "        format(cm[i, j]),\n",
    "        horizontalalignment=\"center\",\n",
    "        color=\"white\" if cm[i, j] > thresh else \"black\",\n",
    "    )\n",
    "plt.ylabel(\"True label\")\n",
    "plt.xlabel(\"Predicted label\")\n",
    "plt.grid(False)\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Códigos orignais:** `../src/data/train_model.py`\n",
    "\n",
    "**Próxima etapa:** `06_count_repetitions.ipynb`"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3.10.12 ('fitness-tracker')",
   "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": "3.10.12"
  },
  "orig_nbformat": 4,
  "vscode": {
   "interpreter": {
    "hash": "4646813149621b46ca6d4f33c3fc66958fdef2358074d2ca1d1c681bb450ff99"
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}