Part 2(d) Light-GBM(clasifying participants).ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Part 2d. Participants Classification based on Light-BGM model\n",
    "by GONG ZERUI 2022.04.21"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 1. Load Libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "# import libraries\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "from numpy import pi \n",
    "from lightgbm import LGBMClassifier\n",
    "\n",
    "from sklearn.preprocessing import StandardScaler, LabelEncoder\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import accuracy_score\n",
    "from sklearn.decomposition import PCA\n",
    "from sklearn.manifold import TSNE\n",
    "\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2. Load Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Shape Train:\t(7352, 564)\n",
      "Shape Test:\t(2947, 564)\n",
      "\n"
     ]
    }
   ],
   "source": [
    "df_train = pd.read_csv(\"source/train.csv\")\n",
    "df_test = pd.read_csv(\"source/test.csv\")\n",
    "\n",
    "# combine both dataframes\n",
    "df_train['Data'] = 'Train'\n",
    "df_test['Data'] = 'Test'\n",
    "df_total = pd.concat([df_train,df_test], axis = 0).reset_index(drop = True)\n",
    "df_total['subject'] = '#' + df_total['subject'].astype(str)\n",
    "\n",
    "# create label\n",
    "activity_label = df_total.pop('Activity')\n",
    "\n",
    "print('Shape Train:\\t{}'.format(df_train.shape))\n",
    "print('Shape Test:\\t{}\\n'.format(df_test.shape))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3. Classifying Participants data using Light-BGM model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Activity:LAYING\n",
      "Accuracy:\t0.64403292\n",
      "\n",
      "Activity:STANDING\n",
      "Accuracy:\t0.54088050\n",
      "\n",
      "Activity:SITTING\n",
      "Accuracy:\t0.53258427\n",
      "\n",
      "Activity:WALKING\n",
      "Accuracy:\t0.96055684\n",
      "\n",
      "Activity:WALKING_UPSTAIRS\n",
      "Accuracy:\t0.92487047\n",
      "\n",
      "Activity:WALKING_DOWNSTAIRS\n",
      "Accuracy:\t0.91477273\n",
      "\n"
     ]
    }
   ],
   "source": [
    "activity_label_counts = activity_label.value_counts()\n",
    "\n",
    "# iterate over each activity and print out the accuracy\n",
    "for activity in activity_label_counts.index:\n",
    "\n",
    "    activity_dataset = df_total[activity_label == activity].copy()\n",
    "    activity_dataset.pop(\"Data\")\n",
    "    activity_subject = activity_dataset.pop(\"subject\")\n",
    "\n",
    "    # scale data and dimension reducing\n",
    "    scl = StandardScaler()\n",
    "    activity_dataset = scl.fit_transform(activity_dataset)\n",
    "\n",
    "    pca = PCA(n_components=0.9, random_state=3)\n",
    "    activity_dataset = pca.fit_transform(activity_dataset)\n",
    "\n",
    "    # split train-test\n",
    "    enc = LabelEncoder()\n",
    "    label_encoded = enc.fit_transform(activity_subject)\n",
    "    X_train, X_test, y_train, y_test = train_test_split(\n",
    "        activity_dataset, label_encoded, random_state=3\n",
    "    )\n",
    "\n",
    "    # fit LBGM model\n",
    "    print('Activity:'+ str(activity))\n",
    "    lgbm = LGBMClassifier(n_estimators=1000, random_state = 3)\n",
    "    lgbm = lgbm.fit(X_train, y_train)\n",
    "    \n",
    "    score = accuracy_score(y_true=y_test, y_pred=lgbm.predict(X_test))\n",
    "    print('Accuracy:\\t{:.8f}\\n'.format(score))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can see that it's very accurate to detect the specific person based on their activity. And the accuracy is much higher for movement. (96.1% for walking) This suggests we can find a unique movement pattern for each individual"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 4. Important features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1440x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import warnings\n",
    "warnings.simplefilter(action='ignore', category=FutureWarning)\n",
    "\n",
    "# sorted(zip(clf.feature_importances_, X.columns), reverse=True)\n",
    "feature_imp = pd.DataFrame(sorted(zip(lgbm.feature_importances_, df_train.columns)), columns=['Value','Feature'])\n",
    "\n",
    "plt.figure(figsize=(20, 10))\n",
    "# feature_importances = (feature_imp / sum(feature_imp)) * 100\n",
    "sns.barplot(x=\"Value\", y=\"Feature\", data=feature_imp.sort_values(by=\"Value\", ascending=False).head(20))\n",
    "plt.title('LightGBM Features (avg over folds)')\n",
    "plt.tight_layout()\n",
    "plt.show()"
   ]
  }
 ],
 "metadata": {
  "interpreter": {
   "hash": "fbc8aa4214fedb48b0b7e9159cb57cac96f213c2036602ef91c90215122c989b"
  },
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}