HARWISDM.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "from scipy import stats\n",
    "from sklearn import metrics\n",
    "from sklearn.model_selection import train_test_split\n",
    "from keras.models import Sequential, Model\n",
    "from keras.layers import Dense\n",
    "from keras.layers import Flatten\n",
    "from keras.layers import Dropout\n",
    "from keras.callbacks import ModelCheckpoint\n",
    "from keras.layers.convolutional import Conv1D\n",
    "from keras.layers.convolutional import MaxPooling1D\n",
    "RANDOM_SEED = 42"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Load Dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "columns = ['user','activity','timestamp', 'x-axis', 'y-axis', 'z-axis']\n",
    "df = pd.read_csv('data/WISDM_ar_v1.1_raw.txt', header = None, names = columns)\n",
    "df = df.dropna()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Total rows :  1098203\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>user</th>\n",
       "      <th>activity</th>\n",
       "      <th>timestamp</th>\n",
       "      <th>x-axis</th>\n",
       "      <th>y-axis</th>\n",
       "      <th>z-axis</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>33</td>\n",
       "      <td>Jogging</td>\n",
       "      <td>49105962326000</td>\n",
       "      <td>-0.694638</td>\n",
       "      <td>12.680544</td>\n",
       "      <td>0.503953</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>33</td>\n",
       "      <td>Jogging</td>\n",
       "      <td>49106062271000</td>\n",
       "      <td>5.012288</td>\n",
       "      <td>11.264028</td>\n",
       "      <td>0.953424</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>33</td>\n",
       "      <td>Jogging</td>\n",
       "      <td>49106112167000</td>\n",
       "      <td>4.903325</td>\n",
       "      <td>10.882658</td>\n",
       "      <td>-0.081722</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>33</td>\n",
       "      <td>Jogging</td>\n",
       "      <td>49106222305000</td>\n",
       "      <td>-0.612916</td>\n",
       "      <td>18.496431</td>\n",
       "      <td>3.023717</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>33</td>\n",
       "      <td>Jogging</td>\n",
       "      <td>49106332290000</td>\n",
       "      <td>-1.184970</td>\n",
       "      <td>12.108489</td>\n",
       "      <td>7.205164</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   user activity       timestamp    x-axis     y-axis    z-axis\n",
       "0    33  Jogging  49105962326000 -0.694638  12.680544  0.503953\n",
       "1    33  Jogging  49106062271000  5.012288  11.264028  0.953424\n",
       "2    33  Jogging  49106112167000  4.903325  10.882658 -0.081722\n",
       "3    33  Jogging  49106222305000 -0.612916  18.496431  3.023717\n",
       "4    33  Jogging  49106332290000 -1.184970  12.108489  7.205164"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print('Total rows : ', len(df))\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Plot Activity Types"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "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": [
    "df['activity'].value_counts().plot(kind='bar', title='Training examples by activity type');"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Normalize Training Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Normalize features for training data set (values between 0 and 1)\n",
    "# Surpress warning for next 3 operation\n",
    "pd.options.mode.chained_assignment = None  # default='warn'\n",
    "\n",
    "df['x-axis'] = df['x-axis'] / df['x-axis'].max()\n",
    "df['y-axis'] = df['y-axis'] / df['y-axis'].max()\n",
    "df['z-axis'] = df['z-axis'] / df['z-axis'].max()\n",
    "# Round numbers\n",
    "df = df.round({'x-axis': 4, 'y-axis': 4, 'z-axis': 4})"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Data Preprocessing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\admin\\Anaconda3\\lib\\site-packages\\scipy\\stats\\stats.py:245: RuntimeWarning: The input array could not be properly checked for nan values. nan values will be ignored.\n",
      "  \"values. nan values will be ignored.\", RuntimeWarning)\n"
     ]
    }
   ],
   "source": [
    "N_TIME_STEPS = 200\n",
    "N_FEATURES = 3\n",
    "step = 20\n",
    "segments = []\n",
    "labels = []\n",
    "for i in range(0, len(df) - N_TIME_STEPS, step):\n",
    "    xs = df['x-axis'].values[i: i + N_TIME_STEPS]\n",
    "    ys = df['y-axis'].values[i: i + N_TIME_STEPS]\n",
    "    zs = df['z-axis'].values[i: i + N_TIME_STEPS]\n",
    "    label = stats.mode(df['activity'][i: i + N_TIME_STEPS])[0][0]\n",
    "    segments.append([xs, ys, zs])\n",
    "    labels.append(label)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "reshaped_segments = np.asarray(segments, dtype= np.float32).reshape(-1, N_TIME_STEPS, N_FEATURES)\n",
    "labels = np.asarray(pd.get_dummies(labels), dtype = np.float32)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(54901, 200, 3)\n",
      "[0. 1. 0. 0. 0. 0.]\n"
     ]
    }
   ],
   "source": [
    "print(reshaped_segments.shape)\n",
    "print(labels[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(\n",
    "        reshaped_segments, labels, test_size=0.5, random_state=RANDOM_SEED)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "X_train :  27450\n",
      "X_test :  27451\n"
     ]
    }
   ],
   "source": [
    "print('X_train : ', len(X_train))\n",
    "print('X_test : ', len(X_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "n_timesteps :  200\n",
      "n_features :  3\n",
      "n_outputs :  6\n"
     ]
    }
   ],
   "source": [
    "verbose = 1\n",
    "epochs = 10\n",
    "batch_size = 400\n",
    "\n",
    "n_timesteps = X_train.shape[1]\n",
    "n_features = X_train.shape[2]\n",
    "n_outputs = y_train.shape[1]\n",
    "\n",
    "print('n_timesteps : ', n_timesteps)\n",
    "print('n_features : ', n_features)\n",
    "print('n_outputs : ', n_outputs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "checkpoint = ModelCheckpoint(\"wisdm_weights.h5\", monitor='val_acc', verbose=1, \n",
    "                             save_best_only=True, save_weights_only=False, mode='auto', period=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = Sequential()\n",
    "model.add(Conv1D(filters=64, kernel_size=3, activation='relu', input_shape=(n_timesteps,n_features)))\n",
    "model.add(Conv1D(filters=64, kernel_size=3, activation='relu', padding = 'same'))\n",
    "model.add(Dropout(0.4))\n",
    "model.add(MaxPooling1D(pool_size=2))\n",
    "model.add(Flatten())\n",
    "model.add(Dense(100, activation='relu'))\n",
    "model.add(Dense(n_outputs, activation='softmax'))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 27450 samples, validate on 27451 samples\n",
      "Epoch 1/10\n",
      "27450/27450 [==============================] - 49s 2ms/step - loss: 0.7255 - acc: 0.7325 - val_loss: 0.5023 - val_acc: 0.8227\n",
      "\n",
      "Epoch 00001: val_acc improved from -inf to 0.82274, saving model to wisdm_weights.h5\n",
      "Epoch 2/10\n",
      "27450/27450 [==============================] - 48s 2ms/step - loss: 0.3985 - acc: 0.8485 - val_loss: 0.3787 - val_acc: 0.8521\n",
      "\n",
      "Epoch 00002: val_acc improved from 0.82274 to 0.85210, saving model to wisdm_weights.h5\n",
      "Epoch 3/10\n",
      "27450/27450 [==============================] - 48s 2ms/step - loss: 0.3067 - acc: 0.8877 - val_loss: 0.3393 - val_acc: 0.8739\n",
      "\n",
      "Epoch 00003: val_acc improved from 0.85210 to 0.87388, saving model to wisdm_weights.h5\n",
      "Epoch 4/10\n",
      "27450/27450 [==============================] - 48s 2ms/step - loss: 0.2531 - acc: 0.9101 - val_loss: 0.2676 - val_acc: 0.8951\n",
      "\n",
      "Epoch 00004: val_acc improved from 0.87388 to 0.89509, saving model to wisdm_weights.h5\n",
      "Epoch 5/10\n",
      "27450/27450 [==============================] - 47s 2ms/step - loss: 0.2087 - acc: 0.9291 - val_loss: 0.2141 - val_acc: 0.9352\n",
      "\n",
      "Epoch 00005: val_acc improved from 0.89509 to 0.93523, saving model to wisdm_weights.h5\n",
      "Epoch 6/10\n",
      "27450/27450 [==============================] - 48s 2ms/step - loss: 0.1712 - acc: 0.9445 - val_loss: 0.2687 - val_acc: 0.9138\n",
      "\n",
      "Epoch 00006: val_acc did not improve from 0.93523\n",
      "Epoch 7/10\n",
      "27450/27450 [==============================] - 50s 2ms/step - loss: 0.1540 - acc: 0.9501 - val_loss: 0.1719 - val_acc: 0.9414\n",
      "\n",
      "Epoch 00007: val_acc improved from 0.93523 to 0.94139, saving model to wisdm_weights.h5\n",
      "Epoch 8/10\n",
      "27450/27450 [==============================] - 50s 2ms/step - loss: 0.1366 - acc: 0.9556 - val_loss: 0.1506 - val_acc: 0.9533\n",
      "\n",
      "Epoch 00008: val_acc improved from 0.94139 to 0.95326, saving model to wisdm_weights.h5\n",
      "Epoch 9/10\n",
      "27450/27450 [==============================] - 49s 2ms/step - loss: 0.1261 - acc: 0.9577 - val_loss: 0.1528 - val_acc: 0.9532\n",
      "\n",
      "Epoch 00009: val_acc did not improve from 0.95326\n",
      "Epoch 10/10\n",
      "27450/27450 [==============================] - 51s 2ms/step - loss: 0.1114 - acc: 0.9646 - val_loss: 0.1433 - val_acc: 0.9554\n",
      "\n",
      "Epoch 00010: val_acc improved from 0.95326 to 0.95545, saving model to wisdm_weights.h5\n",
      "27451/27451 [==============================] - 11s 394us/step\n",
      "[INFO] loss=0.1433, accuracy: 95.5448%\n"
     ]
    }
   ],
   "source": [
    "# Load Weights\n",
    "final_model = Model(inputs = model.input, outputs = model.output)\n",
    "#final_model.load_weights('./har_weights.h5')\n",
    "final_model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])\n",
    "\n",
    "# fit network\n",
    "history = final_model.fit(X_train, y_train, validation_data=(X_test, y_test), \n",
    "                    epochs=epochs, batch_size=batch_size, callbacks = [checkpoint], verbose=verbose)\n",
    "\n",
    "\n",
    "# evaluate model\n",
    "(loss, accuracy) = final_model.evaluate(X_test, y_test, batch_size=batch_size, verbose=verbose)\n",
    "print(\"[INFO] loss={:.4f}, accuracy: {:.4f}%\".format(loss, accuracy * 100))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "\n",
    "# Print confusion matrix for training data\n",
    "y_pred_train = final_model.predict(X_train)\n",
    "# Take the class with the highest probability from the train predictions\n",
    "max_y_pred_train = np.argmax(y_pred_train, axis=1)\n",
    "#print(classification_report(trainy, max_y_pred_train))\n",
    "\n",
    "# confusion matrix\n",
    "LABELS = ['WALKING',\n",
    "          'WALKING_UPSTAIRS',\n",
    "          'WALKING_DOWNSTAIRS',\n",
    "          'SITTING',\n",
    "          'STANDING',\n",
    "          'LAYING']\n",
    "y_pred_test = final_model.predict(X_test)\n",
    "# Take the class with the highest probability from the test predictions\n",
    "max_y_pred_test = np.argmax(y_pred_test, axis=1)\n",
    "max_y_test = np.argmax(y_test, axis=1)\n",
    "\n",
    "matrix = metrics.confusion_matrix(max_y_test, max_y_pred_test)\n",
    "plt.figure(figsize=(6, 4))\n",
    "sns.heatmap(matrix,\n",
    "            cmap='PuOr',\n",
    "            linecolor='white',\n",
    "            linewidths=1,\n",
    "            xticklabels=LABELS,\n",
    "            yticklabels=LABELS,\n",
    "            annot=True,\n",
    "            fmt='d')\n",
    "plt.title('Confusion Matrix')\n",
    "plt.ylabel('True Label')\n",
    "plt.xlabel('Predicted Label')\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.6.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}