notebooks/1_rnn_vs_rforest.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Recurrent Neural Networks vs Random Forest"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import sys\n",
    "nb_dir = os.path.split(os.getcwd())[0]\n",
    "if nb_dir not in sys.path:\n",
    "    sys.path.append(nb_dir)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from keras.models import load_model\n",
    "\n",
    "from core.random_forest import load_data as load_data_rf, train_random_forest, eval_random_forest\n",
    "from core.rnn import create_model, load_data as load_data_rnn, train as train_rnn\n",
    "\n",
    "LABELS_4 = [\"walking\", \"stairs_down\", \"stairs_up\", \"other\"]\n",
    "LABELS_6 = [\"walking\", \"stairs_down\", \"stairs_up\", \"sitting\", \"standing\", \"laying\"]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "## Random Forest"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "def train_rf(train_path, test_path):\n",
    "    x_train, y_train = load_data_rf(train_path)\n",
    "    x_test, y_test = load_data_rf(test_path)\n",
    "    \n",
    "    model = train_random_forest(x_train, y_train)\n",
    "    score = eval_random_forest(model, x_test, y_test)\n",
    "    \n",
    "    return score"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 6 Labels (walking, stairs_down, stairs_up, sitting, standing, laying)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test Accuracy: 0.7567017305734646\n"
     ]
    }
   ],
   "source": [
    "rf_6_accuracy = train_rf(\"../data/uci_train_6.csv\", \"../data/uci_test_6.csv\")\n",
    "\n",
    "print(f\"Test Accuracy: {rf_6_accuracy}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 4 Labels (walking, stairs_down, stairs_up, other)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test Accuracy: 0.831353919239905\n"
     ]
    }
   ],
   "source": [
    "rf_4_accuracy = train_rf(\"../data/uci_train_4.csv\", \"../data/uci_test_4.csv\")\n",
    "\n",
    "print(f\"Test Accuracy: {rf_4_accuracy}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Recurrent Neural Network"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 6 Labels (walking, stairs_down, stairs_up, sitting, standing, laying)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2947/2947 [==============================] - 1s 239us/step\n",
      "Test Accuracy: 0.9056667796142391\n"
     ]
    }
   ],
   "source": [
    "rnn_model_6 = load_model(\"../models/uci6-128.h5\")\n",
    "\n",
    "x_test, y_test = load_data_rnn(\"../data/uci_test_6.csv\", 128, LABELS_6)\n",
    "\n",
    "[_, rnn_6_accuracy] = rnn_model_6.evaluate(x_test, y_test, 512)\n",
    "\n",
    "print(f\"Test Accuracy: {rnn_6_accuracy}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 4 Labels (walking, stairs_down, stairs_up, other)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2947/2947 [==============================] - 1s 187us/step\n",
      "Test Accuracy: 0.9545300297305104\n"
     ]
    }
   ],
   "source": [
    "rnn_model_4 = load_model(\"../models/uci4-128.h5\")\n",
    "\n",
    "x_test, y_test = load_data_rnn(\"../data/uci_test_4.csv\", 128, LABELS_4)\n",
    "\n",
    "[_, rnn_4_accuracy] = rnn_model_4.evaluate(x_test, y_test, 512)\n",
    "\n",
    "print(f\"Test Accuracy: {rnn_4_accuracy}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Conclusions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots()\n",
    "\n",
    "models = ('Random Forest, 6 Labels', 'Random Forest, 4 Labels', 'RNN, 6 Labels', 'RNN, 4 Labels')\n",
    "y_pos = np.arange(len(models))\n",
    "performance = np.array([rf_6_accuracy, rf_4_accuracy, rnn_6_accuracy, rnn_4_accuracy])\n",
    "\n",
    "ax.barh(y_pos, performance, align='center')\n",
    "ax.set_yticks(y_pos)\n",
    "ax.set_yticklabels(models)\n",
    "ax.invert_yaxis()\n",
    "ax.set_xlabel('Accuracy')\n",
    "ax.set_title('RNN vs Random Forest')\n",
    "\n",
    "plt.show()"
   ]
  }
 ],
 "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": "3.6.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}