lstm_cnn.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "from sklearn.model_selection import train_test_split\n",
    "import pandas as pd\n",
    "import matplotlib as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_csv(\"extracted_features.csv\")\n",
    "X = df.drop(columns=['pid', 'window10', 'timestamp', 'intoxicated', \"Unnamed: 0\", \"Unnamed: 0.1\", 'datetime'], axis=1)\n",
    "y = df[['intoxicated']]\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "import tensorflow as tf\n",
    "from tensorflow import keras"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "x = X.to_numpy()\n",
    "x = x.reshape((len(df['window10']), X.shape[1], 1))\n",
    "y = y.to_numpy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Xshape = (72521, 136, 1)\n",
      "Yshape = (72521, 1)\n"
     ]
    }
   ],
   "source": [
    "print(f\"Xshape = {x.shape}\")\n",
    "print(f\"Yshape = {y.shape}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_X, test_X, train_y, test_y = train_test_split(X, y, test_size=0.3, random_state=1)\n",
    "valid_X, test_X, valid_y, test_y = train_test_split(test_X, test_y, test_size=0.5, random_state=1)\n",
    "train_X = np.asarray(train_X).astype('float32')\n",
    "test_X = np.asarray(test_X).astype('float32')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = keras.Sequential([\n",
    "\tkeras.layers.Input((136,1)), \n",
    "\tkeras.layers.Conv1D(filters=64, kernel_size=3, activation='relu', input_shape = (136,1)),\n",
    "\tkeras.layers.Conv1D(filters=64, kernel_size=3, activation='relu'),\n",
    "\tkeras.layers.Dropout(0.5),\n",
    "\tkeras.layers.MaxPool1D(pool_size=2), \n",
    "\tkeras.layers.LSTM(64),\n",
    "\tkeras.layers.Dropout(0.5),\n",
    "\tkeras.layers.Flatten(),\n",
    "\tkeras.layers.Dense(136, activation = 'relu'),\n",
    "\tkeras.layers.Dense(1, activation = 'sigmoid',),\n",
    "])\n",
    "\n",
    "model.compile(optimizer='adam', \n",
    "              loss='binary_crossentropy',\n",
    "              metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/10\n",
      "1270/1270 [==============================] - 82s 62ms/step - loss: 0.2768 - accuracy: 0.8615 - val_loss: 0.0782 - val_accuracy: 0.9655\n",
      "Epoch 2/10\n",
      "1270/1270 [==============================] - 81s 64ms/step - loss: 0.0869 - accuracy: 0.9634 - val_loss: 0.0506 - val_accuracy: 0.9834\n",
      "Epoch 3/10\n",
      "1270/1270 [==============================] - 82s 65ms/step - loss: 0.0740 - accuracy: 0.9710 - val_loss: 0.0496 - val_accuracy: 0.9816\n",
      "Epoch 4/10\n",
      "1270/1270 [==============================] - 91s 71ms/step - loss: 0.0666 - accuracy: 0.9734 - val_loss: 0.0573 - val_accuracy: 0.9777\n",
      "Epoch 5/10\n",
      "1270/1270 [==============================] - 88s 69ms/step - loss: 0.0623 - accuracy: 0.9759 - val_loss: 0.0629 - val_accuracy: 0.9758\n",
      "Epoch 6/10\n",
      "1270/1270 [==============================] - 85s 67ms/step - loss: 0.0632 - accuracy: 0.9754 - val_loss: 0.0684 - val_accuracy: 0.9701\n",
      "Epoch 7/10\n",
      "1270/1270 [==============================] - 91s 72ms/step - loss: 0.0625 - accuracy: 0.9752 - val_loss: 0.0433 - val_accuracy: 0.9840\n",
      "Epoch 8/10\n",
      "1270/1270 [==============================] - 87s 69ms/step - loss: 0.0562 - accuracy: 0.9779 - val_loss: 0.0382 - val_accuracy: 0.9854\n",
      "Epoch 9/10\n",
      "1270/1270 [==============================] - 83s 65ms/step - loss: 0.0579 - accuracy: 0.9775 - val_loss: 0.0505 - val_accuracy: 0.9795\n",
      "Epoch 10/10\n",
      "1270/1270 [==============================] - 85s 67ms/step - loss: 0.0549 - accuracy: 0.9795 - val_loss: 0.0340 - val_accuracy: 0.9863\n"
     ]
    }
   ],
   "source": [
    "history =model.fit(x = train_X, y = train_y, batch_size=None, epochs=10, validation_split=0.2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "prediction = model.predict(test_X).round()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9808119044511161 0.9864877286515304 0.9926023778071335 0.9692982456140351\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import accuracy_score, f1_score, recall_score, precision_score, confusion_matrix\n",
    "f1 = f1_score(prediction, test_y)\n",
    "accuracy = accuracy_score(prediction, test_y)\n",
    "rec = recall_score(prediction, test_y)\n",
    "prec = precision_score(prediction, test_y)\n",
    "print(f1, accuracy, rec, prec)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "def plot_confusion_matrix(test_y, y_pred):\n",
    "    \"\"\"\n",
    "    Given test_y and y_predictions, plot confusion matrix.\n",
    "    \"\"\"\n",
    "    cm = confusion_matrix(test_y, y_pred)\n",
    "    fig, ax = plt.subplots()\n",
    "    ax.matshow(cm)\n",
    "    plt.title('Confusion Matrix')\n",
    "    plt.xlabel('Predicted')\n",
    "    plt.ylabel('True')\n",
    "\n",
    "    for (i, j), z in np.ndenumerate(cm):\n",
    "        ax.text(j, i, '{:d}'.format(z), ha='center', va='center')\n",
    "\n",
    "    plt.show()\n",
    "plot_confusion_matrix(test_y, prediction)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "plt.figure()\n",
    "plt.plot(history.history['loss'])\n",
    "plt.plot(history.history['val_loss'])\n",
    "plt.title('Model Loss')\n",
    "plt.ylabel('Loss')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['train','val'], loc = 'upper right')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "plt.figure()\n",
    "plt.plot(history.history['accuracy'])\n",
    "plt.plot(history.history['val_accuracy'])\n",
    "plt.title('Model Accuracy')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['train','val'], loc = 'upper left')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " conv1d (Conv1D)             (None, 134, 64)           256       \n",
      "                                                                 \n",
      " conv1d_1 (Conv1D)           (None, 132, 64)           12352     \n",
      "                                                                 \n",
      " dropout (Dropout)           (None, 132, 64)           0         \n",
      "                                                                 \n",
      " max_pooling1d (MaxPooling1D  (None, 66, 64)           0         \n",
      " )                                                               \n",
      "                                                                 \n",
      " lstm (LSTM)                 (None, 64)                33024     \n",
      "                                                                 \n",
      " dropout_1 (Dropout)         (None, 64)                0         \n",
      "                                                                 \n",
      " flatten (Flatten)           (None, 64)                0         \n",
      "                                                                 \n",
      " dense (Dense)               (None, 136)               8840      \n",
      "                                                                 \n",
      " dense_1 (Dense)             (None, 1)                 137       \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 54,609\n",
      "Trainable params: 54,609\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model.summary()"
   ]
  }
 ],
 "metadata": {
  "interpreter": {
   "hash": "81794d4967e6c3204c66dcd87b604927b115b27c00565d3d43f05ba2f3a2cb0d"
  },
  "kernelspec": {
   "display_name": "Python 3.9.4 64-bit",
   "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": "3.9.4"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}