bidirectional_lstm_final.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "SUCCESS\n",
      "(1100000,)\n"
     ]
    }
   ],
   "source": [
    "import io\n",
    "import numpy as np\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense, Input, TimeDistributed, Flatten, Dropout, Conv1D, MaxPooling1D\n",
    "from keras.layers import LSTM, SimpleRNN, Bidirectional\n",
    "from keras.layers.embeddings import Embedding\n",
    "from keras.preprocessing import sequence\n",
    "from keras.utils import to_categorical\n",
    "from sklearn.feature_extraction.text import CountVectorizer\n",
    "from nltk.corpus import stopwords\n",
    "from keras.preprocessing.text import Tokenizer\n",
    "from nltk.stem.porter import *\n",
    "\n",
    "stemmer = PorterStemmer()\n",
    "\n",
    "np.random.seed(42)\n",
    "\n",
    "###########################################################################\n",
    "#generate training/test data\n",
    "########################################################################### \n",
    "\n",
    "positive_file1 = io.open(\"Happy_clean.txt\",'r',encoding=\"latin-1\")\n",
    "positive_file2 = io.open(\"Blessed_clean.txt\",'r',encoding=\"latin-1\")\n",
    "positive_file3 = io.open(\"Proud_clean.txt\",'r',encoding=\"latin-1\")\n",
    "\n",
    "negative_file1 = io.open(\"FML_clean.txt\",'r',encoding=\"latin-1\")\n",
    "negative_file2 = io.open(\"Sad_clean.txt\",'r',encoding=\"latin-1\")\n",
    "negative_file3 = io.open(\"Ugh_clean.txt\",'r',encoding=\"latin-1\")\n",
    "\n",
    "\n",
    "positive_tweets = []\n",
    "\n",
    "negative_tweets = []\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "for _ in range(230000):\n",
    "    positive_tweets.append([positive_file1.readline().strip(\"\\n\"),0])\n",
    "    positive_tweets.append([positive_file2.readline().strip(\"\\n\"),0])\n",
    "    positive_tweets.append([positive_file3.readline().strip(\"\\n\"),0])\n",
    "    \n",
    "    negative_tweets.append([negative_file1.readline().strip(\"\\n\"),1])\n",
    "    negative_tweets.append([negative_file2.readline().strip(\"\\n\"),1])\n",
    "    negative_tweets.append([negative_file3.readline().strip(\"\\n\"),1])\n",
    "    \n",
    "tweets = positive_tweets+negative_tweets\n",
    "np.random.shuffle(tweets)\n",
    "\n",
    "\n",
    "tweets_train = np.array(tweets[:1100000])\n",
    "tweets_test = np.array(tweets[1100000:])\n",
    "\n",
    "X_train = tweets_train[:,0]\n",
    "Y_train = to_categorical(np.array([[int(i)] for i in tweets_train[:,1]],dtype=np.int))\n",
    "X_test = tweets_test[:,0]\n",
    "Y_test = to_categorical(np.array([[int(i)] for i in tweets_test[:,1]],dtype=np.int))\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "print(\"SUCCESS\")\n",
    "print(X_train.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 789660 unique tokens.\n"
     ]
    }
   ],
   "source": [
    "top_words = 750000\n",
    "max_tweet_length = 50\n",
    "\n",
    "tokenizer = Tokenizer(num_words = top_words)\n",
    "tokenizer.fit_on_texts(X_train)\n",
    "sequences = tokenizer.texts_to_sequences(X_train)\n",
    "word_index = tokenizer.word_index\n",
    "\n",
    "\n",
    "X_train = sequence.pad_sequences(sequences, maxlen=max_tweet_length)\n",
    "Y_train = np.asarray(Y_train)\n",
    "\n",
    "tokenizer.fit_on_texts(X_test)\n",
    "sequences = tokenizer.texts_to_sequences(X_test)\n",
    "\n",
    "\n",
    "X_test = sequence.pad_sequences(sequences, maxlen=max_tweet_length)\n",
    "Y_test = np.asarray(Y_test)\n",
    "\n",
    "\n",
    "print ('Found %s unique tokens.' % len(word_index))\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "embedding_1 (Embedding)      (None, 50, 50)            37500000  \n",
      "_________________________________________________________________\n",
      "dropout_1 (Dropout)          (None, 50, 50)            0         \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 50, 300)           15300     \n",
      "_________________________________________________________________\n",
      "dropout_2 (Dropout)          (None, 50, 300)           0         \n",
      "_________________________________________________________________\n",
      "dense_2 (Dense)              (None, 50, 300)           90300     \n",
      "_________________________________________________________________\n",
      "bidirectional_1 (Bidirection (None, 50, 600)           1442400   \n",
      "_________________________________________________________________\n",
      "time_distributed_1 (TimeDist (None, 50, 1)             601       \n",
      "_________________________________________________________________\n",
      "flatten_1 (Flatten)          (None, 50)                0         \n",
      "_________________________________________________________________\n",
      "dense_4 (Dense)              (None, 2)                 102       \n",
      "=================================================================\n",
      "Total params: 39,048,703\n",
      "Trainable params: 39,048,703\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n",
      "None\n"
     ]
    }
   ],
   "source": [
    "import keras.optimizers\n",
    "\n",
    "\n",
    "embedding_vector_length = 50\n",
    "model = Sequential()\n",
    "model.add(Embedding(top_words, embedding_vector_length, input_length=max_tweet_length))\n",
    "model.add(Dropout(0.5))\n",
    "model.add(Dense(300,activation='softmax'))\n",
    "model.add(Dropout(0.5))\n",
    "model.add(Dense(300,activation='softmax'))\n",
    "model.add(Bidirectional(LSTM(300,return_sequences=True,recurrent_dropout=0.5,dropout=0.5)))\n",
    "model.add(TimeDistributed(Dense(1)))\n",
    "model.add(Flatten())\n",
    "model.add(Dense(2, activation='softmax'))\n",
    "adam = keras.optimizers.Adam(lr=0.001)\n",
    "model.compile(optimizer=adam,loss='categorical_crossentropy', metrics=['accuracy'])\n",
    "print(model.summary())\n",
    "# history = model.fit(X_train, Y_train, nb_epoch=4, batch_size=32, validation_split=0.1)\n",
    "# model.save_weights('bidirectional_LSTM_final.h5')\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYQAAAEICAYAAABfz4NwAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nDHBAMDOzxAHBzMwABwQzM0v+P9hq1RJqiYziAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x27bae295c18>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAX0AAAEICAYAAACzliQjAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x27c30fe8278>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "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",
    "\n",
    "epochs = range(1, len(acc) +1)\n",
    "\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",
    "\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",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "model.load_weights('bidirectional_LSTM_final.h5')\n",
    "from keras.utils import plot_model\n",
    "plot_model(model, to_file='model.png')\n",
    "# print(\"Classification Error on Test Set: \",model.evaluate(X_test,Y_test,verbose=0))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fuck you \n",
      " [0.17569716 0.82430285]\n",
      "I love you \n",
      " [0.8687122 0.1312878]\n",
      "I hate homework \n",
      " [0.80361545 0.19638455]\n",
      "I like homework \n",
      " [0.47045943 0.5295406 ]\n",
      "Life sucks \n",
      " [0.49002728 0.5099727 ]\n",
      "Life is good \n",
      " [0.7854692  0.21453084]\n",
      "Jesus loves me \n",
      " [0.9355118  0.06448819]\n",
      "Put a ring on it \n",
      " [0.39489925 0.60510075]\n",
      "Mercedes Benz: $20,000 \n",
      " [0.06225213 0.9377479 ]\n",
      "You suck \n",
      " [0.6228995  0.37710053]\n",
      "The decision on Sergeant Bergdahl is a complete and total disgrace to our Country and to our Military. \n",
      " [0.29189077 0.7081092 ]\n",
      "Does anybody really believe that a reporter, who nobody ever heard of, went to his mailbox and found my tax returns? @NBCNews FAKE NEWS! \n",
      " [0.86135995 0.13864009]\n",
      "After Turkey call I will be heading over to Trump National Golf Club, Jupiter, to play golf (quickly) with Tiger Woods and Dustin Johnson. Then back to Mar-a-Lago for talks on bringing even more jobs and companies back to the USA! \n",
      " [0.30206624 0.69793373]\n",
      "“The President has accomplished some absolutely historic things during this past year.” Thank you Charlie Kirk of Turning Points USA. Sadly, the Fake Mainstream Media will NEVER talk about our accomplishments in their end of year reviews. We are compiling a long & beautiful list. \n",
      " [0.97554487 0.02445516]\n",
      "Crazy Joe Scarborough and dumb as a rock Mika are not bad people, but their low rated show is dominated by their NBC bosses. Too bad! \n",
      " [0.66128385 0.33871612]\n",
      "Robert Pattinson should not take back Kristen Stewart. She cheated on him like a dog & will do it again--just watch. He can do much better! \n",
      " [0.22104256 0.7789574 ]\n"
     ]
    }
   ],
   "source": [
    "test_tweets = [\"Fuck you\",\"I love you\",\"I hate homework\",\"I like homework\", \"Life sucks\",\"Life is good\",\n",
    "               \"Jesus loves me\",\"Put a ring on it\",\"Mercedes Benz: $20,000\", \"You suck\",\n",
    "                \"The decision on Sergeant Bergdahl is a complete and total disgrace to our Country and to our Military.\",\n",
    "                \"Does anybody really believe that a reporter, who nobody ever heard of, went to his mailbox and found my tax returns? @NBCNews FAKE NEWS!\",\n",
    "                \"After Turkey call I will be heading over to Trump National Golf Club, Jupiter, to play golf (quickly) with Tiger Woods and Dustin Johnson. Then back to Mar-a-Lago for talks on bringing even more jobs and companies back to the USA!\",\n",
    "                \"“The President has accomplished some absolutely historic things during this past year.” Thank you Charlie Kirk of Turning Points USA. Sadly, the Fake Mainstream Media will NEVER talk about our accomplishments in their end of year reviews. We are compiling a long & beautiful list.\",\n",
    "                \"Crazy Joe Scarborough and dumb as a rock Mika are not bad people, but their low rated show is dominated by their NBC bosses. Too bad!\",\n",
    "                \"Robert Pattinson should not take back Kristen Stewart. She cheated on him like a dog & will do it again--just watch. He can do much better!\"]\n",
    "\n",
    "tokenizer.fit_on_texts(test_tweets)\n",
    "sequences = tokenizer.texts_to_sequences(test_tweets)\n",
    "\n",
    "\n",
    "test_tweets_fit = sequence.pad_sequences(sequences, maxlen=max_tweet_length)\n",
    "\n",
    "predictions = model.predict(test_tweets_fit)\n",
    "for i in range(len(predictions)):\n",
    "    print(test_tweets[i],\"\\n\",predictions[i])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "accuracy =  54.520357142857144\n"
     ]
    }
   ],
   "source": [
    "Y_pred=model.predict(X_test,verbose=0)\n",
    "acc=0\n",
    "for item in range(len(Y_pred)):\n",
    "    if Y_test[item][0]==1 and Y_test[item][1]==0:\n",
    "        if tweets_test[item][1]=='0' and Y_pred[item][1]>=0.5:\n",
    "            acc+=1\n",
    "        elif tweets_test[item][1]=='1' and Y_pred[item][0]>=0.5:\n",
    "            acc+=1   \n",
    "    if Y_test[item][0]==0 and Y_test[item][1]==1:\n",
    "        if tweets_test[item][1]=='0' and Y_pred[item][0]>=0.5:\n",
    "            acc+=1\n",
    "        elif tweets_test[item][1]=='1' and Y_pred[item][1]>=0.5:\n",
    "            acc+=1   \n",
    "            \n",
    "print(\"accuracy = \",acc/len(Y_pred)*100)"
   ]
  },
  {
   "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
}