keras-imdb.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 20180110 - My Neural Network Lab notes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "import keras\n",
    "from keras.datasets import imdb\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense, Dropout, Activation\n",
    "from keras.preprocessing.text import Tokenizer\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "\n",
    "np.random.seed(666)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Data\n",
    "Keras include sample datasets, e.g. movie revies from IMDB\n",
    "\n",
    "Dataset of 25,000 movies reviews from IMDB, labeled by sentiment (positive/negative). \n",
    "\n",
    "Reviews have been preprocessed, and each review is encoded as a sequence of word indexes (integers). For convenience, words are indexed by overall frequency in the dataset, so that for instance the integer \"3\" encodes the 3rd most frequent word in the data. This allows for quick filtering operations such as: \"only consider the top 10,000 most common words, but eliminate the top 20 most common words\".\n",
    "\n",
    "As a convention, \"0\" does not stand for a specific word, but instead is used to encode any unknown word.\n",
    "\n",
    "This dataset comes preloaded with Keras, so one simple command will get us training and testing data. There is a parameter for how many words we want to look at. We've set it at 1000, but feel free to experiment.\n",
    "\n",
    "Usage:\n",
    "\n",
    "```python\n",
    "from keras.datasets import imdb\n",
    "\n",
    "(x_train, y_train), (x_test, y_test) = imdb.load_data(path=\"imdb.npz\",\n",
    "                                                      num_words=None,\n",
    "                                                      skip_top=0,\n",
    "                                                      maxlen=None,\n",
    "                                                      seed=113,\n",
    "                                                      start_char=1,\n",
    "                                                      oov_char=2,\n",
    "                                                      index_from=3)\n",
    "```\n",
    "\n",
    "Returns:\n",
    "\n",
    "2 tuples:\n",
    "\n",
    "x_train, x_test: list of sequences, which are lists of indexes (integers). If the num_words argument was specific, the maximum possible index value is num_words-1. If the maxlen argument was specified, the largest possible sequence length is maxlen.\n",
    "\n",
    "y_train, y_test: list of integer labels (1 or 0).\n",
    "\n",
    "\n",
    "Arguments:\n",
    "path: if you do not have the data locally (at '~/.keras/datasets/' + path), it will be downloaded to this location.\n",
    "\n",
    "num_words: integer or None. Top most frequent words to consider. Any less frequent word will appear as oov_char value in the sequence data.\n",
    "\n",
    "skip_top: integer. Top most frequent words to ignore (they will appear as oov_char value in the sequence data).\n",
    "\n",
    "maxlen: int. Maximum sequence length. Any longer sequence will be truncated.\n",
    "\n",
    "seed: int. Seed for reproducible data shuffling.\n",
    "\n",
    "start_char: int. The start of a sequence will be marked with this character. Set to 1 because 0 is usually the padding character.\n",
    "\n",
    "oov_char: int. words that were cut out because of the num_words or skip_top limit will be replaced with this character.\n",
    "\n",
    "index_from: int. Index actual words with this index and higher."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Length of data vector\n",
      "25000\n",
      "Shape of data vectors:\n",
      "(25000,)\n",
      "(25000,)\n",
      "\n",
      "First element of data vectors\n",
      "[1, 14, 22, 16, 43, 530, 973, 2, 2, 65, 458, 2, 66, 2, 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, 2, 2, 17, 546, 38, 13, 447, 4, 192, 50, 16, 6, 147, 2, 19, 14, 22, 4, 2, 2, 469, 4, 22, 71, 87, 12, 16, 43, 530, 38, 76, 15, 13, 2, 4, 22, 17, 515, 17, 12, 16, 626, 18, 2, 5, 62, 386, 12, 8, 316, 8, 106, 5, 4, 2, 2, 16, 480, 66, 2, 33, 4, 130, 12, 16, 38, 619, 5, 25, 124, 51, 36, 135, 48, 25, 2, 33, 6, 22, 12, 215, 28, 77, 52, 5, 14, 407, 16, 82, 2, 8, 4, 107, 117, 2, 15, 256, 4, 2, 7, 2, 5, 723, 36, 71, 43, 530, 476, 26, 400, 317, 46, 7, 4, 2, 2, 13, 104, 88, 4, 381, 15, 297, 98, 32, 2, 56, 26, 141, 6, 194, 2, 18, 4, 226, 22, 21, 134, 476, 26, 480, 5, 144, 30, 2, 18, 51, 36, 28, 224, 92, 25, 104, 4, 226, 65, 16, 38, 2, 88, 12, 16, 283, 5, 16, 2, 113, 103, 32, 15, 16, 2, 19, 178, 32]\n",
      "1\n"
     ]
    }
   ],
   "source": [
    "(x_train, y_train), (x_test, y_test) = imdb.load_data(num_words=1000)\n",
    "\n",
    "print('\\nLength of data vector')\n",
    "print(len(x_train))\n",
    "\n",
    "print('Shape of data vectors:')\n",
    "print(x_train.shape)\n",
    "print(x_test.shape)\n",
    "\n",
    "print('\\nFirst element of data vectors')\n",
    "print(x_train[0])\n",
    "print(y_train[0])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Data encoding\n",
    "Converting the input vectors into one-hot encoded vectors. \n",
    "\n",
    "0\t00000001\n",
    "\n",
    "1\t00000010\n",
    "\n",
    "2\t00000100\n",
    "\n",
    "https://en.wikipedia.org/wiki/One-hot\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(25000, 1000)\n",
      "(25000, 1000)\n",
      "(25000, 2)\n",
      "(25000, 2)\n"
     ]
    }
   ],
   "source": [
    "tokenizer = Tokenizer(num_words=1000)\n",
    "\n",
    "x_train = tokenizer.sequences_to_matrix(x_train, mode='binary')\n",
    "x_test = tokenizer.sequences_to_matrix(x_test, mode='binary')\n",
    "\n",
    "y_train = keras.utils.to_categorical(y_train, 2)\n",
    "y_test = keras.utils.to_categorical(y_test, 2)\n",
    "\n",
    "print(x_train.shape)\n",
    "print(x_test.shape)\n",
    "\n",
    "print(y_train.shape)\n",
    "print(y_test.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Neural net model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "dense_1 (Dense)              (None, 512)               512512    \n",
      "_________________________________________________________________\n",
      "dropout_1 (Dropout)          (None, 512)               0         \n",
      "_________________________________________________________________\n",
      "dense_2 (Dense)              (None, 2)                 1026      \n",
      "=================================================================\n",
      "Total params: 513,538.0\n",
      "Trainable params: 513,538.0\n",
      "Non-trainable params: 0.0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "# sequential model architecture with one layer of length 100\n",
    "model = Sequential()\n",
    "model.add(Dense(512, activation='relu', input_dim=1000))\n",
    "model.add(Dropout(0.5))\n",
    "model.add(Dense(num_classes, activation='softmax'))\n",
    "model.summary()\n",
    "\n",
    "# Compiling the model using categorical_crossentropy loss, and rmsprop optimizer.\n",
    "model.compile(loss='categorical_crossentropy',\n",
    "              optimizer='rmsprop',\n",
    "              metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Model training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 25000 samples, validate on 25000 samples\n",
      "Epoch 1/10\n",
      "25000/25000 [==============================] - 3s - loss: 0.3967 - acc: 0.8255 - val_loss: 0.3427 - val_acc: 0.8580\n",
      "Epoch 2/10\n",
      "25000/25000 [==============================] - 2s - loss: 0.3344 - acc: 0.8664 - val_loss: 0.3422 - val_acc: 0.8637\n",
      "Epoch 3/10\n",
      "25000/25000 [==============================] - 2s - loss: 0.3223 - acc: 0.8784 - val_loss: 0.3548 - val_acc: 0.8612\n",
      "Epoch 4/10\n",
      "25000/25000 [==============================] - 2s - loss: 0.3105 - acc: 0.8856 - val_loss: 0.3726 - val_acc: 0.8602\n",
      "Epoch 5/10\n",
      "25000/25000 [==============================] - 2s - loss: 0.3051 - acc: 0.8904 - val_loss: 0.4121 - val_acc: 0.8554\n",
      "Epoch 6/10\n",
      "25000/25000 [==============================] - 2s - loss: 0.2978 - acc: 0.8969 - val_loss: 0.3929 - val_acc: 0.8602\n",
      "Epoch 7/10\n",
      "25000/25000 [==============================] - 2s - loss: 0.2879 - acc: 0.9040 - val_loss: 0.4132 - val_acc: 0.8590\n",
      "Epoch 8/10\n",
      "25000/25000 [==============================] - 2s - loss: 0.2742 - acc: 0.9096 - val_loss: 0.4721 - val_acc: 0.8517\n",
      "Epoch 9/10\n",
      "25000/25000 [==============================] - 2s - loss: 0.2614 - acc: 0.9159 - val_loss: 0.4710 - val_acc: 0.8578\n",
      "Epoch 10/10\n",
      "25000/25000 [==============================] - 2s - loss: 0.2522 - acc: 0.9218 - val_loss: 0.4867 - val_acc: 0.8568\n"
     ]
    }
   ],
   "source": [
    "hist = model.fit(x_train, y_train,\n",
    "          batch_size=16,\n",
    "          epochs=10,\n",
    "          validation_data=(x_test, y_test), \n",
    "          verbose=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Model evaluation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Quick visualization of model training history\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYwAAAEWCAYAAAB1xKBvAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7ff3745f8828>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYwAAAEWCAYAAAB1xKBvAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nUsURAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7ff378b9fcc0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "accuracy:  0.8568\n"
     ]
    }
   ],
   "source": [
    "print('Quick visualization of model training history')\n",
    "plt.plot(hist.history['acc'])\n",
    "plt.plot(hist.history['val_acc'])\n",
    "plt.title('model accuracy')\n",
    "plt.ylabel('accuracy')\n",
    "plt.xlabel('epoch')\n",
    "plt.legend(['train', 'val'], loc='upper left')\n",
    "plt.show()\n",
    "\n",
    "# summarize history for loss\n",
    "plt.plot(hist.history['loss'])\n",
    "plt.plot(hist.history['val_loss'])\n",
    "plt.title('model loss')\n",
    "plt.ylabel('loss')\n",
    "plt.xlabel('epoch')\n",
    "plt.legend(['train', 'val'], loc='upper left')\n",
    "plt.show()\n",
    "\n",
    "score = model.evaluate(x_test, y_test, verbose=0)\n",
    "print(\"accuracy: \", score[1])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Prediction\n",
    "\n",
    "Take a look at model.predict function\n",
    "\n",
    "predict(self, x, batch_size=None, verbose=0, steps=None)\n",
    "\n",
    "Generates output predictions for the input samples.\n",
    "\n",
    "The input samples are processed batch by batch.\n",
    "\n",
    "Arguments\n",
    "\n",
    "x: the input data, as a Numpy array.\n",
    "\n",
    "batch_size: Integer. If unspecified, it will default to 32.\n",
    "\n",
    "verbose: verbosity mode, 0 or 1.\n",
    "\n",
    "steps: Total number of steps (batches of samples) before declaring the prediction round finished. Ignored with the default value of None.\n",
    "\n",
    "Returns\n",
    "\n",
    "A Numpy array of predictions."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[  3.89038888e-03   9.96109664e-01]\n",
      " [  9.94147539e-01   5.85240684e-03]\n",
      " [  9.99389052e-01   6.10984396e-04]\n",
      " ..., \n",
      " [  9.99670982e-01   3.29033152e-04]\n",
      " [  3.45729035e-03   9.96542752e-01]\n",
      " [  8.86964262e-01   1.13035776e-01]]\n",
      "25000\n"
     ]
    }
   ],
   "source": [
    "# TODO - correct the output of predict function\n",
    "\n",
    "prediction = model.predict(x_train)\n",
    "print(prediction)\n",
    "print(len(prediction))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# More advance exmaple\n",
    "\n",
    "https://www.kaggle.com/c/sentiment-analysis-on-movie-reviews"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "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": "3.6.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}