embedding_and_simpleRNN_layer_on_imdb.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\Rajiv\\Anaconda3\\lib\\site-packages\\h5py\\__init__.py:34: FutureWarning: Conversion of the second argument of issubdtype from `float` to `np.floating` is deprecated. In future, it will be treated as `np.float64 == np.dtype(float).type`.\n",
      "  from ._conv import register_converters as _register_converters\n",
      "Using TensorFlow backend.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "length of train sequence 25000\n",
      "length of test sequence 25000\n",
      "after padding\n",
      "shape of train sequences (25000, 500)\n",
      "shape of test sequences (25000, 500)\n"
     ]
    }
   ],
   "source": [
    "# importing dataset\n",
    "from keras.datasets import imdb\n",
    "from keras.preprocessing import sequence\n",
    "max_features =1000                                          # no of word considerd as a feature/vocabulary\n",
    "maxlen =500                                                 # maximum no of words in a sequence\n",
    "batch_size =32\n",
    "(x_train ,y_train) ,(x_test ,y_test) = imdb.load_data(num_words =max_features)\n",
    "print('length of train sequence' ,len(x_train))\n",
    "print('length of test sequence' ,len(x_test))\n",
    "print('after padding')\n",
    "x_train = sequence.pad_sequences(x_train ,maxlen =maxlen)\n",
    "x_test =  sequence.pad_sequences(x_test ,maxlen =maxlen)\n",
    "print('shape of train sequences' ,x_train.shape)\n",
    "print('shape of test sequences' ,x_test.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "embedding_1 (Embedding)      (None, None, 32)          32000     \n",
      "_________________________________________________________________\n",
      "simple_rnn_1 (SimpleRNN)     (None, 32)                2080      \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 1)                 33        \n",
      "=================================================================\n",
      "Total params: 34,113\n",
      "Trainable params: 34,113\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "# training model with Embedding and simple RNN\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense ,SimpleRNN ,Embedding\n",
    "model = Sequential()\n",
    "model.add(Embedding(max_features ,32))\n",
    "model.add(SimpleRNN(32))\n",
    "model.add(Dense(1 ,activation ='sigmoid'))\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 20000 samples, validate on 5000 samples\n",
      "Epoch 1/15\n",
      "20000/20000 [==============================] - 18s 903us/step - loss: 0.1468 - acc: 0.9494 - val_loss: 0.6498 - val_acc: 0.7962\n",
      "Epoch 2/15\n",
      "20000/20000 [==============================] - 20s 981us/step - loss: 0.1193 - acc: 0.9591 - val_loss: 0.7218 - val_acc: 0.7782\n",
      "Epoch 3/15\n",
      "20000/20000 [==============================] - 19s 965us/step - loss: 0.1137 - acc: 0.9624 - val_loss: 0.8781 - val_acc: 0.7412\n",
      "Epoch 4/15\n",
      "20000/20000 [==============================] - 19s 971us/step - loss: 0.1019 - acc: 0.9670 - val_loss: 0.7478 - val_acc: 0.7956\n",
      "Epoch 5/15\n",
      "20000/20000 [==============================] - 19s 967us/step - loss: 0.1007 - acc: 0.9666 - val_loss: 0.8755 - val_acc: 0.7448\n",
      "Epoch 6/15\n",
      "20000/20000 [==============================] - 20s 976us/step - loss: 0.0931 - acc: 0.9714 - val_loss: 0.8087 - val_acc: 0.7948\n",
      "Epoch 7/15\n",
      "20000/20000 [==============================] - 19s 954us/step - loss: 0.0906 - acc: 0.9710 - val_loss: 0.8862 - val_acc: 0.7706\n",
      "Epoch 8/15\n",
      "20000/20000 [==============================] - 19s 968us/step - loss: 0.0845 - acc: 0.9759 - val_loss: 0.9418 - val_acc: 0.7544\n",
      "Epoch 9/15\n",
      "20000/20000 [==============================] - 19s 964us/step - loss: 0.0850 - acc: 0.9737 - val_loss: 1.0188 - val_acc: 0.7528\n",
      "Epoch 10/15\n",
      "20000/20000 [==============================] - 20s 981us/step - loss: 0.0903 - acc: 0.9706 - val_loss: 0.9075 - val_acc: 0.7832\n",
      "Epoch 11/15\n",
      "20000/20000 [==============================] - 19s 970us/step - loss: 0.0830 - acc: 0.9731 - val_loss: 0.9362 - val_acc: 0.7710\n",
      "Epoch 12/15\n",
      "20000/20000 [==============================] - 20s 976us/step - loss: 0.0605 - acc: 0.9815 - val_loss: 1.0583 - val_acc: 0.7478\n",
      "Epoch 13/15\n",
      "20000/20000 [==============================] - 20s 995us/step - loss: 0.0600 - acc: 0.9811 - val_loss: 1.0763 - val_acc: 0.7536\n",
      "Epoch 14/15\n",
      "20000/20000 [==============================] - 19s 972us/step - loss: 0.0606 - acc: 0.9823 - val_loss: 1.0161 - val_acc: 0.7764\n",
      "Epoch 15/15\n",
      "20000/20000 [==============================] - 19s 975us/step - loss: 0.0613 - acc: 0.9820 - val_loss: 1.0338 - val_acc: 0.7800\n"
     ]
    }
   ],
   "source": [
    "model.compile(optimizer ='rmsprop' ,loss= 'binary_crossentropy' ,metrics =['acc'])\n",
    "history = model.fit(x_train ,y_train ,epochs =15 ,batch_size =128 ,validation_split =0.2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAX0AAAEICAYAAACzliQjAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n3xhjAsSSvjHGBIglfWOMCZD/D/LsVm2bWwOUAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1fa01cd9b70>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXcAAAEICAYAAACktLTqAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nqCSn/U4AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1fa023b4240>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# plotting the results\n",
    "import matplotlib.pyplot as plt\n",
    "acc = history.history['acc']\n",
    "val_acc = history.history['val_acc']\n",
    "loss = history.history['loss']\n",
    "val_loss = history.history['val_loss']\n",
    "epochs = range(1, len(acc) + 1)\n",
    "plt.plot(epochs, acc, 'bo', label='Training acc')\n",
    "plt.plot(epochs, val_acc, 'b', label='Validation acc')\n",
    "plt.title('Training and validation accuracy')\n",
    "plt.legend()\n",
    "plt.figure()\n",
    "plt.plot(epochs, loss, 'bo', label='Training loss')\n",
    "plt.plot(epochs, val_loss, 'b', label='Validation loss')\n",
    "plt.title('Training and validation loss')\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "maximum Validation accuracy we got is approx 80% ,reason behind this is following;"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "inputs only consider the first 500 words, rather than full sequences—hence, the RNN has access to less information.\n",
    "The remainder of the problem is that SimpleRNN isn’t good at processing long sequences."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 20000 samples, validate on 5000 samples\n",
      "Epoch 1/10\n",
      "20000/20000 [==============================] - 54s 3ms/step - loss: 0.5485 - acc: 0.7169 - val_loss: 0.6671 - val_acc: 0.6848\n",
      "Epoch 2/10\n",
      "20000/20000 [==============================] - 53s 3ms/step - loss: 0.3975 - acc: 0.8267 - val_loss: 0.3489 - val_acc: 0.8546\n",
      "Epoch 3/10\n",
      "20000/20000 [==============================] - 53s 3ms/step - loss: 0.3646 - acc: 0.8460 - val_loss: 0.4813 - val_acc: 0.8074\n",
      "Epoch 4/10\n",
      "20000/20000 [==============================] - 53s 3ms/step - loss: 0.3448 - acc: 0.8555 - val_loss: 0.3306 - val_acc: 0.8636\n",
      "Epoch 5/10\n",
      "20000/20000 [==============================] - 53s 3ms/step - loss: 0.3370 - acc: 0.8620 - val_loss: 0.3782 - val_acc: 0.8506\n",
      "Epoch 6/10\n",
      "20000/20000 [==============================] - 52s 3ms/step - loss: 0.3240 - acc: 0.8649 - val_loss: 0.3501 - val_acc: 0.8562\n",
      "Epoch 7/10\n",
      "20000/20000 [==============================] - 53s 3ms/step - loss: 0.3182 - acc: 0.8705 - val_loss: 0.4434 - val_acc: 0.7984\n",
      "Epoch 8/10\n",
      "20000/20000 [==============================] - 53s 3ms/step - loss: 0.3138 - acc: 0.8695 - val_loss: 0.3536 - val_acc: 0.8502\n",
      "Epoch 9/10\n",
      "20000/20000 [==============================] - 53s 3ms/step - loss: 0.3083 - acc: 0.8710 - val_loss: 0.3530 - val_acc: 0.8482\n",
      "Epoch 10/10\n",
      "20000/20000 [==============================] - 53s 3ms/step - loss: 0.3017 - acc: 0.8772 - val_loss: 0.3554 - val_acc: 0.8514\n"
     ]
    }
   ],
   "source": [
    "from keras.layers import LSTM ,Embedding ,Dense\n",
    "from keras.models import Sequential\n",
    "model1 = Sequential()\n",
    "model1.add(Embedding(max_features ,32))\n",
    "model1.add(LSTM(32))\n",
    "model1.add(Dense(1 ,activation ='sigmoid'))\n",
    "model1.compile(optimizer ='rmsprop' ,loss ='binary_crossentropy' ,metrics =['acc'])\n",
    "history1 =model1.fit(x_train ,y_train ,epochs =10 ,batch_size =128 ,validation_split =0.2)"
   ]
  },
  {
   "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": "3.6.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}