keras_pg_68.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# IMDB reviews dataset"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Importing dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.datasets import imdb\n",
    "\n",
    "(train_data, train_labels), (test_data, test_labels) = imdb.load_data(num_words=10000)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Get a feel of the data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Word indexes --  [1, 14, 22, 16, 43, 530, 973, 1622, 1385, 65, 458, 4468, 66, 3941, 4, 173, 36, 256, 5, 25, 100, 43, 838, 112, 50, 670, 2, 9, 35, 480, 284, 5, 150, 4, 172, 112, 167, 2, 336, 385, 39, 4, 172, 4536, 1111, 17, 546, 38, 13, 447, 4, 192, 50, 16, 6, 147, 2025, 19, 14, 22, 4, 1920, 4613, 469, 4, 22, 71, 87, 12, 16, 43, 530, 38, 76, 15, 13, 1247, 4, 22, 17, 515, 17, 12, 16, 626, 18, 2, 5, 62, 386, 12, 8, 316, 8, 106, 5, 4, 2223, 5244, 16, 480, 66, 3785, 33, 4, 130, 12, 16, 38, 619, 5, 25, 124, 51, 36, 135, 48, 25, 1415, 33, 6, 22, 12, 215, 28, 77, 52, 5, 14, 407, 16, 82, 2, 8, 4, 107, 117, 5952, 15, 256, 4, 2, 7, 3766, 5, 723, 36, 71, 43, 530, 476, 26, 400, 317, 46, 7, 4, 2, 1029, 13, 104, 88, 4, 381, 15, 297, 98, 32, 2071, 56, 26, 141, 6, 194, 7486, 18, 4, 226, 22, 21, 134, 476, 26, 480, 5, 144, 30, 5535, 18, 51, 36, 28, 224, 92, 25, 104, 4, 226, 65, 16, 38, 1334, 88, 12, 16, 283, 5, 16, 4472, 113, 103, 32, 15, 16, 5345, 19, 178, 32]\n",
      "Positive -- 1 and Negative -- 0 1\n"
     ]
    }
   ],
   "source": [
    "print('Word indexes -- ', train_data[0])\n",
    "print('Positive -- 1 and Negative -- 0', train_labels[0])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### None of the word indexes go beyond 9999 (because num_words = 10000)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "9999"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "max([max(sequence) for sequence in train_data])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Try get a feel of the data by seeing what was the review in training and what's the label "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Decoded review --  ? this film was just brilliant casting location scenery story direction everyone's really suited the part they played and you could just imagine being there robert ? is an amazing actor and now the same being director ? father came from the same scottish island as myself so i loved the fact there was a real connection with this film the witty remarks throughout the film were great it was just brilliant so much that i bought the film as soon as it was released for ? and would recommend it to everyone to watch and the fly fishing was amazing really cried at the end it was so sad and you know what they say if you cry at a film it must have been good and this definitely was also ? to the two little boy's that played the ? of norman and paul they were just brilliant children are often left out of the ? list i think because the stars that play them all grown up are such a big profile for the whole film but these children are amazing and should be praised for what they have done don't you think the whole story was so lovely because it was true and was someone's life after all that was shared with us all\n",
      "Label alloted --  1\n"
     ]
    }
   ],
   "source": [
    "word_index = imdb.get_word_index() # word_index contains -- (word, index)\n",
    "reverse_word_index = dict( \n",
    "        [(value, key) for (key, value) in word_index.items()]) # Just reversing the mapping to obtain word from index\n",
    "decoded_review = ' '.join(\n",
    "        [reverse_word_index.get(i - 3, '?') for i in train_data[0]]) # Get the statement of X_train[0]\n",
    "print('Decoded review -- ', decoded_review)\n",
    "print('Label alloted -- ' , train_labels[0])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Preparing data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Cannot use lists in neural nets, so One-hot vectorize your lists -- eg. [3, 5] would appear as numpy 1_D vector with 3 and 5 index as 1 and rest zeros"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "\n",
    "def vectorize_sequences(sequences, dimension=10000):\n",
    "    results = np.zeros((len(sequences), dimension))\n",
    "    for i, sequence in enumerate(sequences):\n",
    "        results[i, sequence] = 1\n",
    "    return results\n",
    "\n",
    "X_train = vectorize_sequences(train_data)\n",
    "X_test = vectorize_sequences(test_data)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Vectorize labels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_train = np.asarray(train_labels).astype('float32')\n",
    "y_test = np.asarray(test_labels).astype('float32')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Defining the network\n",
    "\n",
    "> ### Two intermediate layers with 16 hidden units each\n",
    "> ### A third layer that will output the scalar prediction regarding the sentiment of the current review"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.models import Sequential\n",
    "from keras.layers import Dense\n",
    "\n",
    "model = Sequential()\n",
    "model.add(Dense(units=16, activation='relu', input_shape=(10000, )))\n",
    "model.add(Dense(16, activation='relu'))\n",
    "model.add(Dense(1, activation='sigmoid'))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Compiling the model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras import optimizers\n",
    "from keras import metrics"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Create validation set to monitor accuracy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "x_val = X_train[:10000]\n",
    "partial_x_train = X_train[10000:]\n",
    "\n",
    "y_val = y_train[:10000]\n",
    "partial_y_train = y_train[10000:]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Training your model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 15000 samples, validate on 10000 samples\n",
      "Epoch 1/20\n",
      "15000/15000 [==============================] - 4s 263us/step - loss: 0.5222 - acc: 0.7840 - val_loss: 0.3928 - val_acc: 0.8677\n",
      "Epoch 2/20\n",
      "15000/15000 [==============================] - 2s 131us/step - loss: 0.3151 - acc: 0.8974 - val_loss: 0.3156 - val_acc: 0.8808\n",
      "Epoch 3/20\n",
      "15000/15000 [==============================] - 2s 103us/step - loss: 0.2296 - acc: 0.9255 - val_loss: 0.2791 - val_acc: 0.8919\n",
      "Epoch 4/20\n",
      "15000/15000 [==============================] - 2s 101us/step - loss: 0.1806 - acc: 0.9393 - val_loss: 0.2899 - val_acc: 0.8823\n",
      "Epoch 5/20\n",
      "15000/15000 [==============================] - 2s 105us/step - loss: 0.1428 - acc: 0.9535 - val_loss: 0.3023 - val_acc: 0.8818\n",
      "Epoch 6/20\n",
      "15000/15000 [==============================] - 2s 103us/step - loss: 0.1178 - acc: 0.9629 - val_loss: 0.2924 - val_acc: 0.8864\n",
      "Epoch 7/20\n",
      "15000/15000 [==============================] - 2s 106us/step - loss: 0.0972 - acc: 0.9713 - val_loss: 0.3093 - val_acc: 0.8829\n",
      "Epoch 8/20\n",
      "15000/15000 [==============================] - 2s 104us/step - loss: 0.0767 - acc: 0.9781 - val_loss: 0.3459 - val_acc: 0.8797\n",
      "Epoch 9/20\n",
      "15000/15000 [==============================] - 2s 101us/step - loss: 0.0653 - acc: 0.9825 - val_loss: 0.3670 - val_acc: 0.8749\n",
      "Epoch 10/20\n",
      "15000/15000 [==============================] - 1s 98us/step - loss: 0.0476 - acc: 0.9889 - val_loss: 0.3788 - val_acc: 0.8785\n",
      "Epoch 11/20\n",
      "15000/15000 [==============================] - 1s 97us/step - loss: 0.0405 - acc: 0.9909 - val_loss: 0.4293 - val_acc: 0.8744\n",
      "Epoch 12/20\n",
      "15000/15000 [==============================] - 1s 97us/step - loss: 0.0312 - acc: 0.9941 - val_loss: 0.4344 - val_acc: 0.8770\n",
      "Epoch 13/20\n",
      "15000/15000 [==============================] - 1s 97us/step - loss: 0.0240 - acc: 0.9962 - val_loss: 0.4748 - val_acc: 0.8749\n",
      "Epoch 14/20\n",
      "15000/15000 [==============================] - 1s 97us/step - loss: 0.0192 - acc: 0.9967 - val_loss: 0.5151 - val_acc: 0.8712\n",
      "Epoch 15/20\n",
      "15000/15000 [==============================] - 2s 112us/step - loss: 0.0161 - acc: 0.9969 - val_loss: 0.5294 - val_acc: 0.8717\n",
      "Epoch 16/20\n",
      "15000/15000 [==============================] - 2s 126us/step - loss: 0.0099 - acc: 0.9989 - val_loss: 0.5972 - val_acc: 0.8666\n",
      "Epoch 17/20\n",
      "15000/15000 [==============================] - 2s 108us/step - loss: 0.0072 - acc: 0.9996 - val_loss: 0.6559 - val_acc: 0.8559\n",
      "Epoch 18/20\n",
      "15000/15000 [==============================] - 1s 98us/step - loss: 0.0050 - acc: 0.9998 - val_loss: 0.6390 - val_acc: 0.8652\n",
      "Epoch 19/20\n",
      "15000/15000 [==============================] - 2s 103us/step - loss: 0.0044 - acc: 0.9999 - val_loss: 0.6676 - val_acc: 0.8674\n",
      "Epoch 20/20\n",
      "15000/15000 [==============================] - 1s 97us/step - loss: 0.0045 - acc: 0.9993 - val_loss: 0.7048 - val_acc: 0.8635\n"
     ]
    }
   ],
   "source": [
    "model.compile(optimizer='rmsprop', \n",
    "             loss='binary_crossentropy',\n",
    "             metrics=['acc'])\n",
    "\n",
    "history = model.fit(partial_x_train,\n",
    "                   partial_y_train,\n",
    "                   epochs=20,\n",
    "                   batch_size=512,\n",
    "                   validation_data = (x_val, y_val))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## History"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "dict_keys(['val_loss', 'val_acc', 'loss', 'acc'])"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "history_dict = history.history\n",
    "history_dict.keys()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Plotting the training and validation loss"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "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": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "history_dict = history.history\n",
    "loss_values = history_dict['loss']\n",
    "val_loss_values = history_dict['val_loss']\n",
    "acc = history_dict['acc']\n",
    "val_acc_values = history_dict['val_acc']\n",
    "\n",
    "epochs = range(1, len(acc)+1)\n",
    "\n",
    "plt.plot(epochs, loss_values, 'bo', label='Training Loss')\n",
    "plt.plot(epochs, val_loss_values, 'b', label='Validation loss')\n",
    "plt.title('Training and validation loss')\n",
    "plt.xlabel('Training and validation loss')\n",
    "plt.ylabel('Epochs')\n",
    "plt.ylabel('Loss')\n",
    "plt.legend()\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Plotting the training and validation accuracy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "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": [
    "plt.clf() ## Clears the figure\n",
    "\n",
    "plt.plot(epochs, acc, 'bo', label='Training accuracy')\n",
    "plt.plot(epochs, val_acc_values, 'b', label='Validation accuracy')\n",
    "plt.title('Training and validation accuracy')\n",
    "plt.xlabel('Epochs')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.legend()\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "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
}