Embedding.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import io\n",
    "import numpy as np\n",
    "from sklearn.feature_extraction.text import CountVectorizer\n",
    "from sklearn.feature_extraction.text import TfidfTransformer\n",
    "from sklearn.linear_model import LogisticRegression \n",
    "\n",
    "###########################################################################\n",
    "#generate training/test data\n",
    "########################################################################### \n",
    "bragging_file = io.open(\"Happy_clean.txt\",'r',encoding=\"utf-8\") \n",
    "bitching_file = io.open(\"FML_clean.txt\",'r',encoding=\"utf-8\") \n",
    "\n",
    "training_samples = 30000\n",
    "bragging_tweets = []\n",
    "bitching_tweets = []\n",
    "\n",
    "for _ in range(249000):\n",
    "    bragging_tweets.append([bragging_file.readline().strip(\"\\n\"),0])\n",
    "    bitching_tweets.append([bitching_file.readline().strip(\"\\n\"),1])\n",
    "    \n",
    "tweets = bragging_tweets+bitching_tweets\n",
    "np.random.seed(42)\n",
    "np.random.shuffle(tweets)\n",
    "\n",
    "\n",
    "tweets_train = np.array(tweets[:348600])\n",
    "tweets_test = np.array(tweets[348600:498000])\n",
    "\n",
    "X_train = tweets_train[:,0]\n",
    "Y_train = tweets_train[:,1]\n",
    "# X_val = tweets_train[training_samples:maximum_data,0]\n",
    "# Y_val = tweets_train[training_samples:maximum_data,1]\n",
    "X_test = tweets_test[:,0]\n",
    "Y_test = tweets_test[:,1]\n",
    "\n",
    "# #########################################################################\n",
    "# #bag of words\n",
    "# #########################################################################\n",
    "\n",
    "# vectorizer = CountVectorizer()\n",
    "# X_train_counts = vectorizer.fit_transform(X_train)\n",
    "# X_val_counts = vectorizer.transform(X_val)\n",
    "# X_test_counts = vectorizer.transform(X_test)\n",
    "\n",
    "\n",
    "# # tfidf_transformer = TfidfTransformer()\n",
    "# # X_train_tfidf = tfidf_transformer.fit_transform(X_train_counts)\n",
    "# # X_test_tfidf = tfidf_transformer.transform(X_test_counts)\n",
    "\n",
    "# #########################################################################\n",
    "# #logistic regression\n",
    "# #########################################################################"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 271907 unique tokens.\n"
     ]
    }
   ],
   "source": [
    "from keras.preprocessing.text import Tokenizer\n",
    "from keras.preprocessing.sequence import pad_sequences\n",
    "import numpy as np\n",
    "\n",
    "maxlen = 50\n",
    "training_samples = 200000\n",
    "max_words = 10000\n",
    "tokenizer = Tokenizer(num_words = max_words)\n",
    "tokenizer.fit_on_texts(X_train)\n",
    "sequences = tokenizer.texts_to_sequences(X_train)\n",
    "word_index = tokenizer.word_index\n",
    "print ('Found %s unique tokens.' % len(word_index))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "shape of the data tensor: (348600, 50)\n",
      "shape of the label tensor:  (348600,)\n"
     ]
    }
   ],
   "source": [
    "data = pad_sequences(sequences, maxlen = maxlen)\n",
    "labels = np.asarray(Y_train)\n",
    "print (\"shape of the data tensor:\", data.shape)\n",
    "print (\"shape of the label tensor: \", labels.shape)\n",
    "x_train = data[:]\n",
    "y_train = labels[:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 1193514 word vectors.\n"
     ]
    }
   ],
   "source": [
    "# splits data into a training set and a validation set\n",
    "# indices = np.arange(data.shape[0])\n",
    "# np.random.shuffle(indices)\n",
    "# data = data[indices]\n",
    "# labels = labels[indices]\n",
    "import os\n",
    "# x_val = data[training_samples:]\n",
    "# y_val = labels[training_samples:]\n",
    "\n",
    "glove_dir = '/Users/hamzakhan/Desktop/Winter 2018/Hashtag_Recommendations_NN/glove.twitter.27B'\n",
    "embeddings_index = {}\n",
    "#your maxlen and you .txt files must correspond\n",
    "f = open(os.path.join(glove_dir, 'glove.twitter.27B.50d.txt'))\n",
    "for line in f:\n",
    "    values = line.split()\n",
    "    word = values[0]\n",
    "    coefs=np.asarray(values[1:], dtype='float32')\n",
    "    embeddings_index[word] = coefs\n",
    "f.close()\n",
    "print('Found %s word vectors.'%len(embeddings_index))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "embedding_dim = maxlen\n",
    "embedding_matrix = np.zeros((max_words, embedding_dim))\n",
    "for word, i in word_index.items():\n",
    "    if (i < max_words):\n",
    "        embedding_vector= embeddings_index.get(word)\n",
    "        if (embedding_vector is not None):\n",
    "            embedding_matrix[i] = embedding_vector            "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "embedding_1 (Embedding)      (None, 50, 50)            500000    \n",
      "_________________________________________________________________\n",
      "flatten_1 (Flatten)          (None, 2500)              0         \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 32)                80032     \n",
      "_________________________________________________________________\n",
      "dense_2 (Dense)              (None, 1)                 33        \n",
      "=================================================================\n",
      "Total params: 580,065\n",
      "Trainable params: 580,065\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "from keras.models import Sequential\n",
    "from keras.layers import Embedding, Flatten, Dense\n",
    "\n",
    "model = Sequential()\n",
    "model.add(Embedding(max_words, embedding_dim, input_length = maxlen))\n",
    "model.add(Flatten())\n",
    "model.add(Dense(32, activation = 'relu'))\n",
    "model.add(Dense(1, activation = 'sigmoid'))\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#this is where we add the embeddings to the model\n",
    "model.layers[0].set_weights([embedding_matrix])\n",
    "model.layers[0].trainable = False"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 331170 samples, validate on 17430 samples\n",
      "Epoch 1/10\n",
      "331170/331170 [==============================] - 32s 96us/step - loss: 0.2929 - acc: 0.8774 - val_loss: 0.3924 - val_acc: 0.8301\n",
      "Epoch 2/10\n",
      "331170/331170 [==============================] - 33s 99us/step - loss: 0.2634 - acc: 0.8908 - val_loss: 0.2687 - val_acc: 0.8901\n",
      "Epoch 3/10\n",
      "331170/331170 [==============================] - 35s 106us/step - loss: 0.2537 - acc: 0.8954 - val_loss: 0.2704 - val_acc: 0.8893\n",
      "Epoch 4/10\n",
      "331170/331170 [==============================] - 31s 92us/step - loss: 0.2468 - acc: 0.8985 - val_loss: 0.2525 - val_acc: 0.8974\n",
      "Epoch 5/10\n",
      "331170/331170 [==============================] - 34s 103us/step - loss: 0.2414 - acc: 0.9005 - val_loss: 0.2532 - val_acc: 0.8985\n",
      "Epoch 6/10\n",
      "331170/331170 [==============================] - 35s 106us/step - loss: 0.2368 - acc: 0.9023 - val_loss: 0.2481 - val_acc: 0.8980\n",
      "Epoch 7/10\n",
      "331170/331170 [==============================] - 34s 103us/step - loss: 0.2328 - acc: 0.9040 - val_loss: 0.2517 - val_acc: 0.8991\n",
      "Epoch 8/10\n",
      "331170/331170 [==============================] - 33s 99us/step - loss: 0.2291 - acc: 0.9053 - val_loss: 0.3613 - val_acc: 0.8679\n",
      "Epoch 9/10\n",
      "331170/331170 [==============================] - 33s 101us/step - loss: 0.2262 - acc: 0.9065 - val_loss: 0.2463 - val_acc: 0.8992\n",
      "Epoch 10/10\n",
      "331170/331170 [==============================] - 34s 102us/step - loss: 0.2230 - acc: 0.9080 - val_loss: 0.2459 - val_acc: 0.9007\n"
     ]
    }
   ],
   "source": [
    "import keras.optimizers\n",
    "sgd = keras.optimizers.SGD(lr=0.0001*400)\n",
    "model.compile(optimizer = sgd,\n",
    "             loss = 'binary_crossentropy',\n",
    "             metrics = ['acc'])\n",
    "\n",
    "history = model.fit(x_train, y_train, epochs = 10, batch_size= 32, validation_split = 0.05 )\n",
    "model.save_weights('pre-trained_glove_twitter_model.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAX0AAAEICAYAAACzliQjAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1ab8c4def0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYQAAAEICAYAAABfz4NwAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nMcYAFhCMMcZ4/j/TAWfZNO4AEwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1a9edd2e48>"
      ]
     },
     "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 (with embeddings)')\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 (with embeddings)')\n",
    "plt.legend()\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "149400/149400 [==============================] - 8s 56us/step\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[0.24519523541691154, 0.90046854083317829]"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "## These will be results on the test\n",
    "sequences = tokenizer.texts_to_sequences(X_test)\n",
    "x_test = pad_sequences(sequences, maxlen = maxlen)\n",
    "y_test = np.asarray(Y_test)\n",
    "model.load_weights('pre-trained_glove_twitter_model.h5')\n",
    "model.evaluate(x_test, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "embedding_2 (Embedding)      (None, 50, 50)            500000    \n",
      "_________________________________________________________________\n",
      "flatten_2 (Flatten)          (None, 2500)              0         \n",
      "_________________________________________________________________\n",
      "dense_3 (Dense)              (None, 32)                80032     \n",
      "_________________________________________________________________\n",
      "dense_4 (Dense)              (None, 1)                 33        \n",
      "=================================================================\n",
      "Total params: 580,065\n",
      "Trainable params: 580,065\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n",
      "Train on 331170 samples, validate on 17430 samples\n",
      "Epoch 1/10\n",
      "331170/331170 [==============================] - 65s 198us/step - loss: 0.3138 - acc: 0.8609 - val_loss: 0.2528 - val_acc: 0.8947\n",
      "Epoch 2/10\n",
      "331170/331170 [==============================] - 62s 188us/step - loss: 0.2250 - acc: 0.9087 - val_loss: 0.2559 - val_acc: 0.8939\n",
      "Epoch 3/10\n",
      "331170/331170 [==============================] - 66s 200us/step - loss: 0.2071 - acc: 0.9163 - val_loss: 0.2223 - val_acc: 0.9101\n",
      "Epoch 4/10\n",
      "331170/331170 [==============================] - 59s 178us/step - loss: 0.1939 - acc: 0.9219 - val_loss: 0.2719 - val_acc: 0.8871\n",
      "Epoch 5/10\n",
      "331170/331170 [==============================] - 68s 205us/step - loss: 0.1841 - acc: 0.9269 - val_loss: 0.2149 - val_acc: 0.9155\n",
      "Epoch 6/10\n",
      "331170/331170 [==============================] - 68s 204us/step - loss: 0.1769 - acc: 0.9292 - val_loss: 0.2119 - val_acc: 0.9163\n",
      "Epoch 7/10\n",
      "331170/331170 [==============================] - 67s 204us/step - loss: 0.1698 - acc: 0.9325 - val_loss: 0.2057 - val_acc: 0.9192\n",
      "Epoch 8/10\n",
      "331170/331170 [==============================] - 67s 203us/step - loss: 0.1637 - acc: 0.9351 - val_loss: 0.2076 - val_acc: 0.9189\n",
      "Epoch 9/10\n",
      "331170/331170 [==============================] - 70s 210us/step - loss: 0.1575 - acc: 0.9378 - val_loss: 0.2135 - val_acc: 0.9160\n",
      "Epoch 10/10\n",
      "331170/331170 [==============================] - 70s 210us/step - loss: 0.1510 - acc: 0.9404 - val_loss: 0.2404 - val_acc: 0.9115\n"
     ]
    }
   ],
   "source": [
    "##Training without embeddings \n",
    "from keras.models import Sequential\n",
    "from keras.layers import Embedding, Flatten, Dense\n",
    "\n",
    "model = Sequential()\n",
    "model.add(Embedding(max_words, embedding_dim, input_length = maxlen))\n",
    "model.add(Flatten())\n",
    "model.add(Dense(32, activation = 'relu'))\n",
    "model.add(Dense(1, activation = 'sigmoid'))\n",
    "model.summary()\n",
    "import keras.optimizers\n",
    "sgd = keras.optimizers.SGD(lr=0.0001*400)\n",
    "model.compile(optimizer = sgd,\n",
    "             loss = 'binary_crossentropy',\n",
    "             metrics = ['acc'])\n",
    "\n",
    "history = model.fit(x_train, y_train, epochs = 10, batch_size= 32, validation_split = 0.05 )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAX0AAAEICAYAAACzliQjAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nvnPOZZH/BwZGECLl01ZCAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1a9f395a90>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAX0AAAEICAYAAACzliQjAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1a9f3a6390>"
      ]
     },
     "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 (without embeddings)')\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 (without embeddings)')\n",
    "plt.legend()\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "149400/149400 [==============================] - 7s 46us/step\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[0.22842103302129702, 0.91319277108752905]"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.evaluate(x_test, y_test)"
   ]
  }
 ],
 "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
}