14_lstm_concat.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 54
    },
    "colab_type": "code",
    "id": "9nn-YIyzJ09Y",
    "outputId": "6b6aa014-3e5f-4079-d7c5-cccef9c9ef56"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Drive already mounted at /content/drive; to attempt to forcibly remount, call drive.mount(\"/content/drive\", force_remount=True).\n"
     ]
    }
   ],
   "source": [
    "from google.colab import drive\n",
    "drive.mount('/content/drive')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "rhUo7OzGQl4Z"
   },
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import tensorflow as tf\n",
    "from math import sqrt\n",
    "\n",
    "\n",
    "from tensorflow.keras.layers import Input, Dense, Embedding, LSTM, Activation, GlobalMaxPooling1D\n",
    "from tensorflow.keras.preprocessing.text import Tokenizer\n",
    "from tensorflow.keras.preprocessing.sequence import pad_sequences\n",
    "\n",
    "from tensorflow.keras.models import Sequential\n",
    "from tensorflow.keras.callbacks import ReduceLROnPlateau, EarlyStopping\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "\n",
    "from sklearn.model_selection import StratifiedShuffleSplit"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "2pNg4ay8JZmg"
   },
   "outputs": [],
   "source": [
    "#Run Local\n",
    "#df = pd.read_csv(\"csv/book_info_complete.csv\")\n",
    "#Run on COLAB\n",
    "df = pd.read_csv(\"/content/drive/My Drive/final_project/book_info_complete.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 301
    },
    "colab_type": "code",
    "id": "5hNMh6xrJZmm",
    "outputId": "e3929fd8-a4d3-48db-b8d8-aa50726ae1fd"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 5816 entries, 0 to 5815\n",
      "Data columns (total 9 columns):\n",
      " #   Column            Non-Null Count  Dtype \n",
      "---  ------            --------------  ----- \n",
      " 0   title             5816 non-null   object\n",
      " 1   description       5816 non-null   object\n",
      " 2   plot              5816 non-null   object\n",
      " 3   csm_review        5816 non-null   object\n",
      " 4   need_to_know      5816 non-null   object\n",
      " 5   csm_rating        5816 non-null   int64 \n",
      " 6   Genre             5816 non-null   object\n",
      " 7   Book type         5816 non-null   object\n",
      " 8   Publication date  5816 non-null   object\n",
      "dtypes: int64(1), object(8)\n",
      "memory usage: 409.1+ KB\n"
     ]
    }
   ],
   "source": [
    "df = df.dropna(axis=1)\n",
    "df[\"Publication date\"] = df[\"Publication date\"].str[-4:]\n",
    "df = df.drop(\"Last updated\", axis=1)\n",
    "df.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "V7b04dMoJZms"
   },
   "outputs": [],
   "source": [
    "df[\"text\"] = df['title'] + \" \" + df['plot'] + \" \" + df[\"description\"] + \" \" + df[\"csm_review\"] + \" \" + df[\"need_to_know\"] + \" \" + df[\"Genre\"] + \" \" + df[\"Book type\"]+ \" \" + df[\"Publication date\"]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "a1RnxUyXJZmx"
   },
   "source": [
    "## Create the splits"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "M0_-59xOJZmy"
   },
   "outputs": [],
   "source": [
    "def splitter(df):\n",
    "    split = StratifiedShuffleSplit(n_splits=1, test_size=0.2, random_state=999)\n",
    "    for train_index, test_index in split.split(df, df['csm_rating']):\n",
    "        train_data= df.loc[train_index]\n",
    "        test_data = df.loc[test_index]\n",
    "    \n",
    "    \n",
    "    return train_data, test_data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "3wkFIijOJZm4"
   },
   "outputs": [],
   "source": [
    "train_data, test_data = splitter(df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "YjI0Lfc8JZm9"
   },
   "outputs": [],
   "source": [
    "x_tr, y_tr = train_data['text'].values, train_data[\"csm_rating\"].values\n",
    "x_val, y_val = test_data[\"text\"].values, test_data[\"csm_rating\"].values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "id": "64omNypEJZnC",
    "outputId": "d2344a23-48bb-44eb-bb4d-501fcf5b1a05"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(4652,) (4652,)\n"
     ]
    }
   ],
   "source": [
    "print(x_tr.shape, y_tr.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "id": "6SCDNHa2JZnJ",
    "outputId": "474de6d1-8145-4cb7-df17-e4a27b2c6f26"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1164,) (1164,)\n"
     ]
    }
   ],
   "source": [
    "print(x_val.shape, y_val.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "Z1CH5m22JZnO"
   },
   "source": [
    "## Prepare the Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 52
    },
    "colab_type": "code",
    "id": "liitdrp9JZnP",
    "outputId": "37254626-f72a-4c70-b6f7-36393e5f1d15"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "979\n",
      "117\n"
     ]
    }
   ],
   "source": [
    "#Tokenize the sentences\n",
    "tokenizer = Tokenizer()\n",
    "\n",
    "#preparing vocabulary\n",
    "tokenizer.fit_on_texts(list(x_tr))\n",
    "\n",
    "#converting text into integer sequences\n",
    "x_tr_seq  = tokenizer.texts_to_sequences(x_tr) \n",
    "x_val_seq = tokenizer.texts_to_sequences(x_val)\n",
    "\n",
    "print(len(max(x_tr_seq, key=len)))\n",
    "max_length = len(max(x_tr_seq, key=len))\n",
    "print(len(min(x_tr_seq, key=len)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 52
    },
    "colab_type": "code",
    "id": "nGP4OL9XJZnT",
    "outputId": "a7530b27-f5e2-4c27-d5fe-6c33de0eab2e"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "979\n",
      "979\n"
     ]
    }
   ],
   "source": [
    "#padding to prepare sequences of same length\n",
    "x_tr_seq  = pad_sequences(x_tr_seq, maxlen=max_length)\n",
    "x_val_seq = pad_sequences(x_val_seq, maxlen=max_length)\n",
    "print(len(max(x_tr_seq, key=len)))\n",
    "print(len(min(x_tr_seq, key=len)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "id": "NuoM9u8HJZnZ",
    "outputId": "8dee58d9-2bcb-4ff8-fb60-3481820c0fe7"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "45235\n"
     ]
    }
   ],
   "source": [
    "size_of_vocabulary=len(tokenizer.word_index) + 1 #+1 for padding\n",
    "print(size_of_vocabulary)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "ThBDgrP8JZnf"
   },
   "source": [
    "## Create embeddings"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "id": "VdR43fJQJZng",
    "outputId": "6f6c02a0-b051-46c4-e14a-9ece3cb87862"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 45234 unique tokens.\n"
     ]
    }
   ],
   "source": [
    "word_index = tokenizer.word_index\n",
    "print(\"Found %s unique tokens.\" % len(word_index))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "uF9AT0LdJZnl"
   },
   "source": [
    "## Create the Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 354
    },
    "colab_type": "code",
    "id": "aUdRk0UDJZnl",
    "outputId": "fb682e4e-edd6-4b38-fd74-0aab724ae960"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "embedding (Embedding)        (None, 979, 300)          13570500  \n",
      "_________________________________________________________________\n",
      "lstm (LSTM)                  (None, 979, 128)          219648    \n",
      "_________________________________________________________________\n",
      "global_max_pooling1d (Global (None, 128)               0         \n",
      "_________________________________________________________________\n",
      "dense (Dense)                (None, 64)                8256      \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 1)                 65        \n",
      "=================================================================\n",
      "Total params: 13,798,469\n",
      "Trainable params: 13,798,469\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n",
      "None\n"
     ]
    }
   ],
   "source": [
    "model=Sequential()\n",
    "\n",
    "#embedding layer\n",
    "model.add(Embedding(size_of_vocabulary,300,input_length=max_length,trainable=True)) \n",
    "\n",
    "#lstm layer\n",
    "model.add(LSTM(128,return_sequences=True,dropout=0.2))\n",
    "\n",
    "#Global Maxpooling\n",
    "model.add(GlobalMaxPooling1D())\n",
    "\n",
    "#Dense Layer\n",
    "model.add(Dense(64,activation='relu')) \n",
    "model.add(Dense(1,activation='relu')) \n",
    "\n",
    "model.compile(optimizer=\"RMSprop\", loss='mean_squared_error', metrics=[\"mae\"]) \n",
    "\n",
    "#Print summary of model\n",
    "print(model.summary())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "hKZo_lzaJZnr"
   },
   "source": [
    "## Callbacks"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "Kl98dyhkJZns"
   },
   "outputs": [],
   "source": [
    "earlystop = EarlyStopping(patience=10)\n",
    "learning_rate_reduction = ReduceLROnPlateau(monitor='val_mae', \n",
    "                                            patience=2, \n",
    "                                            verbose=1, \n",
    "                                            factor=0.5, \n",
    "                                            min_lr=0.00001)\n",
    "\n",
    "callbacks = [earlystop, learning_rate_reduction]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "op3MuPCKJZnx"
   },
   "source": [
    "## Fit the Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 961
    },
    "colab_type": "code",
    "id": "_I34WFwPJZny",
    "outputId": "09157892-f993-4cd8-8382-d90c0dce1387"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/1000\n",
      "37/37 [==============================] - 9s 241ms/step - loss: 22.3452 - mae: 3.8764 - val_loss: 14.8064 - val_mae: 3.2526 - lr: 0.0010\n",
      "Epoch 2/1000\n",
      "37/37 [==============================] - 8s 225ms/step - loss: 13.7956 - mae: 3.1450 - val_loss: 11.9354 - val_mae: 2.9058 - lr: 0.0010\n",
      "Epoch 3/1000\n",
      "37/37 [==============================] - 8s 223ms/step - loss: 9.5904 - mae: 2.6124 - val_loss: 9.3547 - val_mae: 2.3912 - lr: 0.0010\n",
      "Epoch 4/1000\n",
      "37/37 [==============================] - 8s 223ms/step - loss: 6.0659 - mae: 2.0060 - val_loss: 6.1538 - val_mae: 1.7970 - lr: 0.0010\n",
      "Epoch 5/1000\n",
      "37/37 [==============================] - 8s 225ms/step - loss: 4.4369 - mae: 1.6154 - val_loss: 4.4430 - val_mae: 1.5531 - lr: 0.0010\n",
      "Epoch 6/1000\n",
      "37/37 [==============================] - 8s 225ms/step - loss: 3.5309 - mae: 1.4366 - val_loss: 3.6579 - val_mae: 1.4254 - lr: 0.0010\n",
      "Epoch 7/1000\n",
      "37/37 [==============================] - 8s 226ms/step - loss: 2.6832 - mae: 1.2500 - val_loss: 3.7426 - val_mae: 1.4882 - lr: 0.0010\n",
      "Epoch 8/1000\n",
      "37/37 [==============================] - ETA: 0s - loss: 2.2992 - mae: 1.1645\n",
      "Epoch 00008: ReduceLROnPlateau reducing learning rate to 0.0005000000237487257.\n",
      "37/37 [==============================] - 8s 224ms/step - loss: 2.2992 - mae: 1.1645 - val_loss: 4.2020 - val_mae: 1.5354 - lr: 0.0010\n",
      "Epoch 9/1000\n",
      "37/37 [==============================] - 8s 223ms/step - loss: 1.5141 - mae: 0.9270 - val_loss: 3.3952 - val_mae: 1.3770 - lr: 5.0000e-04\n",
      "Epoch 10/1000\n",
      "37/37 [==============================] - 8s 224ms/step - loss: 1.3755 - mae: 0.8874 - val_loss: 3.5139 - val_mae: 1.4101 - lr: 5.0000e-04\n",
      "Epoch 11/1000\n",
      "37/37 [==============================] - 8s 224ms/step - loss: 1.1578 - mae: 0.8104 - val_loss: 3.3887 - val_mae: 1.3532 - lr: 5.0000e-04\n",
      "Epoch 12/1000\n",
      "37/37 [==============================] - 8s 224ms/step - loss: 0.9851 - mae: 0.7462 - val_loss: 3.4476 - val_mae: 1.3817 - lr: 5.0000e-04\n",
      "Epoch 13/1000\n",
      "37/37 [==============================] - 8s 224ms/step - loss: 0.8990 - mae: 0.7087 - val_loss: 3.4394 - val_mae: 1.3453 - lr: 5.0000e-04\n",
      "Epoch 14/1000\n",
      "37/37 [==============================] - 8s 225ms/step - loss: 0.8305 - mae: 0.6869 - val_loss: 3.4471 - val_mae: 1.3808 - lr: 5.0000e-04\n",
      "Epoch 15/1000\n",
      "37/37 [==============================] - ETA: 0s - loss: 0.7095 - mae: 0.6282\n",
      "Epoch 00015: ReduceLROnPlateau reducing learning rate to 0.0002500000118743628.\n",
      "37/37 [==============================] - 8s 225ms/step - loss: 0.7095 - mae: 0.6282 - val_loss: 3.4585 - val_mae: 1.3575 - lr: 5.0000e-04\n",
      "Epoch 16/1000\n",
      "37/37 [==============================] - 8s 225ms/step - loss: 0.4907 - mae: 0.5079 - val_loss: 3.5152 - val_mae: 1.3634 - lr: 2.5000e-04\n",
      "Epoch 17/1000\n",
      "37/37 [==============================] - ETA: 0s - loss: 0.4540 - mae: 0.4780\n",
      "Epoch 00017: ReduceLROnPlateau reducing learning rate to 0.0001250000059371814.\n",
      "37/37 [==============================] - 8s 222ms/step - loss: 0.4540 - mae: 0.4780 - val_loss: 3.5271 - val_mae: 1.3676 - lr: 2.5000e-04\n",
      "Epoch 18/1000\n",
      "37/37 [==============================] - 8s 224ms/step - loss: 0.3573 - mae: 0.4162 - val_loss: 3.4911 - val_mae: 1.3712 - lr: 1.2500e-04\n",
      "Epoch 19/1000\n",
      "37/37 [==============================] - ETA: 0s - loss: 0.3311 - mae: 0.3998\n",
      "Epoch 00019: ReduceLROnPlateau reducing learning rate to 6.25000029685907e-05.\n",
      "37/37 [==============================] - 8s 225ms/step - loss: 0.3311 - mae: 0.3998 - val_loss: 3.4779 - val_mae: 1.3738 - lr: 1.2500e-04\n",
      "Epoch 20/1000\n",
      "37/37 [==============================] - 8s 224ms/step - loss: 0.2950 - mae: 0.3660 - val_loss: 3.4929 - val_mae: 1.3764 - lr: 6.2500e-05\n",
      "Epoch 21/1000\n",
      "37/37 [==============================] - ETA: 0s - loss: 0.2854 - mae: 0.3599\n",
      "Epoch 00021: ReduceLROnPlateau reducing learning rate to 3.125000148429535e-05.\n",
      "37/37 [==============================] - 8s 225ms/step - loss: 0.2854 - mae: 0.3599 - val_loss: 3.5019 - val_mae: 1.3812 - lr: 6.2500e-05\n"
     ]
    }
   ],
   "source": [
    "history = model.fit(np.array(x_tr_seq),\n",
    "                    np.array(y_tr),\n",
    "                    batch_size=128,\n",
    "                    epochs=1000,\n",
    "                    validation_data=(np.array(x_val_seq),np.array(y_val)),\n",
    "                    verbose=1,\n",
    "                    callbacks=callbacks)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 52
    },
    "colab_type": "code",
    "id": "ojMu-A3wJZn4",
    "outputId": "7e542430-be09-4e77-f032-8b3a50910ee1"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "37/37 [==============================] - 1s 29ms/step - loss: 3.2136 - mae: 1.3812\n",
      "The val_mae is 1.381.\n"
     ]
    }
   ],
   "source": [
    "#evaluation \n",
    "val_loss, val_mae = model.evaluate(x_val_seq, y_val)\n",
    "\n",
    "print(\"The val_mae is %.3f.\" % val_mae)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 107
    },
    "colab_type": "code",
    "id": "gl3QL2BaU9XQ",
    "outputId": "5a965c20-3aba-4e8c-cab3-3576e6a9daf7"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/tensorflow/python/ops/resource_variable_ops.py:1817: calling BaseResourceVariable.__init__ (from tensorflow.python.ops.resource_variable_ops) with constraint is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "If using Keras pass *_constraint arguments to layers.\n",
      "INFO:tensorflow:Assets written to: /content/drive/My Drive/final_project/lstm_concat_model/assets\n"
     ]
    }
   ],
   "source": [
    "model.save('/content/drive/My Drive/final_project/lstm_concat_model') "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 521
    },
    "colab_type": "code",
    "id": "QX82za-8JZn_",
    "outputId": "ad1f81ec-858e-4da7-c223-24baa5012001"
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXAAAAD8CAYAAABuHP8oAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAS4ikqT+P6soVWub3mCZAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXQAAAD8CAYAAABn919SAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nSUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(history.history[\"loss\"], label=\"loss\");\n",
    "plt.plot(history.history[\"val_loss\"], label=\"val_loss\");\n",
    "plt.ylim(0, 25)\n",
    "plt.legend();\n",
    "plt.savefig(\"lstm_custom_loss\")\n",
    "plt.show();\n",
    "plt.close();\n",
    "\n",
    "plt.plot(history.history[\"mae\"], label=\"mae\");\n",
    "plt.plot(history.history[\"val_mae\"], label=\"val_mae\");\n",
    "plt.ylim(0,4)\n",
    "plt.legend();\n",
    "plt.savefig(\"lstm_custom_mae\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "QSOY8AH2JZoF"
   },
   "source": [
    "## [Use Transfer Learning](https://www.analyticsvidhya.com/blog/2020/03/pretrained-word-embeddings-nlp/)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "id": "GdKeDrv9JZoG",
    "outputId": "131ebab6-bc52-4ee6-b499-a40e0865e6e5"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loaded 400000 word vectors.\n"
     ]
    }
   ],
   "source": [
    "# load the whole embedding into memory\n",
    "embeddings_index = dict()\n",
    "\n",
    "with open(\"/content/drive/My Drive/Colab Notebooks/glove.6B/glove.6B.300d.txt\") as f:\n",
    "\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",
    "\n",
    "print('Loaded %s word vectors.' % len(embeddings_index))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "dJqtmI0pJZoL"
   },
   "outputs": [],
   "source": [
    "# create a weight matrix for words in training docs\n",
    "embedding_matrix = np.zeros((size_of_vocabulary, 300))\n",
    "\n",
    "for word, i in tokenizer.word_index.items():\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": 23,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 354
    },
    "colab_type": "code",
    "id": "NGM-hGE4JZoQ",
    "outputId": "116dc794-e1f7-4b80-e881-c1455056a6ba"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_1\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "embedding_1 (Embedding)      (None, 979, 300)          13570500  \n",
      "_________________________________________________________________\n",
      "lstm_1 (LSTM)                (None, 979, 128)          219648    \n",
      "_________________________________________________________________\n",
      "global_max_pooling1d_1 (Glob (None, 128)               0         \n",
      "_________________________________________________________________\n",
      "dense_2 (Dense)              (None, 64)                8256      \n",
      "_________________________________________________________________\n",
      "dense_3 (Dense)              (None, 1)                 65        \n",
      "=================================================================\n",
      "Total params: 13,798,469\n",
      "Trainable params: 227,969\n",
      "Non-trainable params: 13,570,500\n",
      "_________________________________________________________________\n",
      "None\n"
     ]
    }
   ],
   "source": [
    "model2=Sequential()\n",
    "\n",
    "#embedding layer\n",
    "model2.add(Embedding(size_of_vocabulary,\n",
    "                    300,\n",
    "                    weights=[embedding_matrix],\n",
    "                    input_length=max_length,\n",
    "                    trainable=False)) \n",
    "\n",
    "#lstm layer\n",
    "model2.add(LSTM(128,return_sequences=True,dropout=0.2))\n",
    "\n",
    "#Global Maxpooling\n",
    "model2.add(GlobalMaxPooling1D())\n",
    "\n",
    "#Dense Layer\n",
    "model2.add(Dense(64,activation='relu')) \n",
    "model2.add(Dense(1,activation='relu')) \n",
    "\n",
    "model2.compile(optimizer=\"RMSprop\", loss='mean_squared_error', metrics=[\"mae\"]) \n",
    "\n",
    "#Print summary of model\n",
    "print(model2.summary())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 1000
    },
    "colab_type": "code",
    "id": "o4IlOKDvJZoU",
    "outputId": "4838040f-f71b-4cd6-d285-26a74da53518"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/1000\n",
      "37/37 [==============================] - 5s 146ms/step - loss: 20.7488 - mae: 3.6938 - val_loss: 10.6513 - val_mae: 2.7677 - lr: 0.0010\n",
      "Epoch 2/1000\n",
      "37/37 [==============================] - 5s 132ms/step - loss: 7.3238 - mae: 2.2358 - val_loss: 3.6052 - val_mae: 1.4452 - lr: 0.0010\n",
      "Epoch 3/1000\n",
      "37/37 [==============================] - 5s 133ms/step - loss: 3.7520 - mae: 1.5353 - val_loss: 3.1955 - val_mae: 1.3759 - lr: 0.0010\n",
      "Epoch 4/1000\n",
      "37/37 [==============================] - 5s 133ms/step - loss: 3.2194 - mae: 1.3957 - val_loss: 3.0244 - val_mae: 1.3383 - lr: 0.0010\n",
      "Epoch 5/1000\n",
      "37/37 [==============================] - 5s 132ms/step - loss: 2.6532 - mae: 1.2932 - val_loss: 3.4264 - val_mae: 1.4354 - lr: 0.0010\n",
      "Epoch 6/1000\n",
      "37/37 [==============================] - 5s 131ms/step - loss: 2.4888 - mae: 1.2378 - val_loss: 2.4466 - val_mae: 1.1900 - lr: 0.0010\n",
      "Epoch 7/1000\n",
      "37/37 [==============================] - 5s 132ms/step - loss: 2.2813 - mae: 1.1893 - val_loss: 2.2932 - val_mae: 1.1441 - lr: 0.0010\n",
      "Epoch 8/1000\n",
      "37/37 [==============================] - 5s 132ms/step - loss: 2.1760 - mae: 1.1532 - val_loss: 3.0218 - val_mae: 1.3591 - lr: 0.0010\n",
      "Epoch 9/1000\n",
      "37/37 [==============================] - 5s 131ms/step - loss: 1.9431 - mae: 1.0891 - val_loss: 2.2461 - val_mae: 1.1334 - lr: 0.0010\n",
      "Epoch 10/1000\n",
      "37/37 [==============================] - 5s 132ms/step - loss: 1.9189 - mae: 1.0999 - val_loss: 2.0444 - val_mae: 1.0879 - lr: 0.0010\n",
      "Epoch 11/1000\n",
      "37/37 [==============================] - 5s 131ms/step - loss: 1.7703 - mae: 1.0465 - val_loss: 2.1659 - val_mae: 1.1269 - lr: 0.0010\n",
      "Epoch 12/1000\n",
      "37/37 [==============================] - ETA: 0s - loss: 1.6778 - mae: 1.0164\n",
      "Epoch 00012: ReduceLROnPlateau reducing learning rate to 0.0005000000237487257.\n",
      "37/37 [==============================] - 5s 131ms/step - loss: 1.6778 - mae: 1.0164 - val_loss: 2.4442 - val_mae: 1.1845 - lr: 0.0010\n",
      "Epoch 13/1000\n",
      "37/37 [==============================] - 5s 132ms/step - loss: 1.3482 - mae: 0.9079 - val_loss: 2.1803 - val_mae: 1.1133 - lr: 5.0000e-04\n",
      "Epoch 14/1000\n",
      "37/37 [==============================] - 5s 131ms/step - loss: 1.2787 - mae: 0.8838 - val_loss: 1.9733 - val_mae: 1.0771 - lr: 5.0000e-04\n",
      "Epoch 15/1000\n",
      "37/37 [==============================] - 5s 132ms/step - loss: 1.2097 - mae: 0.8622 - val_loss: 2.1999 - val_mae: 1.1148 - lr: 5.0000e-04\n",
      "Epoch 16/1000\n",
      "37/37 [==============================] - 5s 132ms/step - loss: 1.1860 - mae: 0.8532 - val_loss: 1.9240 - val_mae: 1.0538 - lr: 5.0000e-04\n",
      "Epoch 17/1000\n",
      "37/37 [==============================] - 5s 132ms/step - loss: 1.0956 - mae: 0.8211 - val_loss: 1.8833 - val_mae: 1.0503 - lr: 5.0000e-04\n",
      "Epoch 18/1000\n",
      "37/37 [==============================] - 5s 132ms/step - loss: 1.0674 - mae: 0.8029 - val_loss: 2.0720 - val_mae: 1.0884 - lr: 5.0000e-04\n",
      "Epoch 19/1000\n",
      "37/37 [==============================] - ETA: 0s - loss: 1.0081 - mae: 0.7884\n",
      "Epoch 00019: ReduceLROnPlateau reducing learning rate to 0.0002500000118743628.\n",
      "37/37 [==============================] - 5s 132ms/step - loss: 1.0081 - mae: 0.7884 - val_loss: 1.9470 - val_mae: 1.0647 - lr: 5.0000e-04\n",
      "Epoch 20/1000\n",
      "37/37 [==============================] - 5s 131ms/step - loss: 0.8824 - mae: 0.7341 - val_loss: 1.8752 - val_mae: 1.0433 - lr: 2.5000e-04\n",
      "Epoch 21/1000\n",
      "37/37 [==============================] - 5s 132ms/step - loss: 0.8558 - mae: 0.7198 - val_loss: 1.9465 - val_mae: 1.0550 - lr: 2.5000e-04\n",
      "Epoch 22/1000\n",
      "37/37 [==============================] - 5s 131ms/step - loss: 0.8126 - mae: 0.7037 - val_loss: 1.8606 - val_mae: 1.0403 - lr: 2.5000e-04\n",
      "Epoch 23/1000\n",
      "37/37 [==============================] - 5s 131ms/step - loss: 0.7882 - mae: 0.6940 - val_loss: 1.9643 - val_mae: 1.0629 - lr: 2.5000e-04\n",
      "Epoch 24/1000\n",
      "37/37 [==============================] - ETA: 0s - loss: 0.7587 - mae: 0.6826\n",
      "Epoch 00024: ReduceLROnPlateau reducing learning rate to 0.0001250000059371814.\n",
      "37/37 [==============================] - 5s 131ms/step - loss: 0.7587 - mae: 0.6826 - val_loss: 1.8841 - val_mae: 1.0467 - lr: 2.5000e-04\n",
      "Epoch 25/1000\n",
      "37/37 [==============================] - 5s 131ms/step - loss: 0.7107 - mae: 0.6563 - val_loss: 1.8509 - val_mae: 1.0443 - lr: 1.2500e-04\n",
      "Epoch 26/1000\n",
      "37/37 [==============================] - 5s 131ms/step - loss: 0.6984 - mae: 0.6553 - val_loss: 1.8494 - val_mae: 1.0393 - lr: 1.2500e-04\n",
      "Epoch 27/1000\n",
      "37/37 [==============================] - 5s 132ms/step - loss: 0.6726 - mae: 0.6400 - val_loss: 1.9073 - val_mae: 1.0508 - lr: 1.2500e-04\n",
      "Epoch 28/1000\n",
      "37/37 [==============================] - ETA: 0s - loss: 0.6605 - mae: 0.6385\n",
      "Epoch 00028: ReduceLROnPlateau reducing learning rate to 6.25000029685907e-05.\n",
      "37/37 [==============================] - 5s 132ms/step - loss: 0.6605 - mae: 0.6385 - val_loss: 1.8587 - val_mae: 1.0421 - lr: 1.2500e-04\n",
      "Epoch 29/1000\n",
      "37/37 [==============================] - 5s 132ms/step - loss: 0.6532 - mae: 0.6331 - val_loss: 1.8963 - val_mae: 1.0497 - lr: 6.2500e-05\n",
      "Epoch 30/1000\n",
      "37/37 [==============================] - ETA: 0s - loss: 0.6238 - mae: 0.6193\n",
      "Epoch 00030: ReduceLROnPlateau reducing learning rate to 3.125000148429535e-05.\n",
      "37/37 [==============================] - 5s 132ms/step - loss: 0.6238 - mae: 0.6193 - val_loss: 1.8828 - val_mae: 1.0458 - lr: 6.2500e-05\n",
      "Epoch 31/1000\n",
      "37/37 [==============================] - 5s 131ms/step - loss: 0.6165 - mae: 0.6125 - val_loss: 1.8763 - val_mae: 1.0447 - lr: 3.1250e-05\n",
      "Epoch 32/1000\n",
      "37/37 [==============================] - ETA: 0s - loss: 0.6219 - mae: 0.6160\n",
      "Epoch 00032: ReduceLROnPlateau reducing learning rate to 1.5625000742147677e-05.\n",
      "37/37 [==============================] - 5s 132ms/step - loss: 0.6219 - mae: 0.6160 - val_loss: 1.8811 - val_mae: 1.0451 - lr: 3.1250e-05\n",
      "Epoch 33/1000\n",
      "37/37 [==============================] - 5s 132ms/step - loss: 0.6102 - mae: 0.6122 - val_loss: 1.8522 - val_mae: 1.0424 - lr: 1.5625e-05\n",
      "Epoch 34/1000\n",
      "37/37 [==============================] - ETA: 0s - loss: 0.6004 - mae: 0.6101\n",
      "Epoch 00034: ReduceLROnPlateau reducing learning rate to 1e-05.\n",
      "37/37 [==============================] - 5s 132ms/step - loss: 0.6004 - mae: 0.6101 - val_loss: 1.8697 - val_mae: 1.0443 - lr: 1.5625e-05\n",
      "Epoch 35/1000\n",
      "37/37 [==============================] - 5s 131ms/step - loss: 0.6045 - mae: 0.6085 - val_loss: 1.8642 - val_mae: 1.0437 - lr: 1.0000e-05\n",
      "Epoch 36/1000\n",
      "37/37 [==============================] - 5s 132ms/step - loss: 0.6008 - mae: 0.6071 - val_loss: 1.8678 - val_mae: 1.0443 - lr: 1.0000e-05\n"
     ]
    }
   ],
   "source": [
    "history2 = model2.fit(np.array(x_tr_seq),\n",
    "                    np.array(y_tr),\n",
    "                    batch_size=128,\n",
    "                    epochs=1000,\n",
    "                    validation_data=(np.array(x_val_seq),np.array(y_val)),\n",
    "                    verbose=1,\n",
    "                    callbacks=callbacks)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 52
    },
    "colab_type": "code",
    "id": "88gF08odJZoY",
    "outputId": "c531887b-5f94-489a-e15b-b1beae2cf374"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "37/37 [==============================] - 1s 27ms/step - loss: 1.7923 - mae: 1.0443\n",
      "The val_mae is 1.044.\n"
     ]
    }
   ],
   "source": [
    "#evaluation \n",
    "_, val_mae = model2.evaluate(x_val_seq, y_val)\n",
    "\n",
    "print(\"The val_mae is %.3f.\" % val_mae)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "id": "xmPRr1m6bUFI",
    "outputId": "4027468b-77e6-48a9-8bf7-b1cfe4efb664"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: /content/drive/My Drive/final_project/lstm_concat_transfer_model/assets\n"
     ]
    }
   ],
   "source": [
    "model.save('/content/drive/My Drive/final_project/lstm_concat_transfer_model') "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 521
    },
    "colab_type": "code",
    "id": "FY3cjzn3JZoe",
    "outputId": "5ecee93a-b6db-4213-a7ee-cbfa8f365d9a"
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXAAAAD8CAYAAABuHP8oAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nIgH1/wFg+K6gFXW1SwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXQAAAD8CAYAAABn919SAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nwgLdGGMyhAW6McZkCAt0Y4zJEP8fEDNXHPT1tMEAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(history2.history[\"loss\"], label=\"loss\");\n",
    "plt.plot(history2.history[\"val_loss\"], label=\"val_loss\");\n",
    "plt.ylim(0,25)\n",
    "plt.legend();\n",
    "plt.savefig(\"lstm_pretrained_loss\")\n",
    "plt.show();\n",
    "plt.close();\n",
    "\n",
    "plt.plot(history2.history[\"mae\"], label=\"mae\");\n",
    "plt.plot(history2.history[\"val_mae\"], label=\"val_mae\");\n",
    "plt.ylim(0, 4)\n",
    "plt.legend();\n",
    "plt.savefig(\"lstm_pretrained_mae\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "KJ8CiHmPtftz"
   },
   "outputs": [],
   "source": [
    "a = model.predict(x_val_seq)\n",
    "b = model2.predict(x_val_seq)\n",
    "#a = min_max_scaler_target.inverse_transform(a)\n",
    "#b = min_max_scaler_target.inverse_transform(b)\n",
    "\n",
    "csm_lstm_predictions_df = pd.DataFrame({\"concat_custom\": list(a), \"concat_transfer\": list(b)},\n",
    "                              columns = [\"concat_custom\", \"concat_transfer\"], \n",
    "                              index=test_data.index)\n",
    "\n",
    "csm_lstm_predictions_df.to_csv('/content/drive/My Drive/final_project/lstm_concat.csv', index=\"False\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "T1vMqXpnF0ZL"
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "collapsed_sections": [],
   "name": "10_lstm_concat.ipynb",
   "provenance": [],
   "toc_visible": true
  },
  "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": "3.7.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}