SentimentAnalysis_NeuralNetwork.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "SentimentAnalysis_NeuralNetwork.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "machine_shape": "hm"
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "kDBCAQwLSsXE",
        "colab_type": "text"
      },
      "source": [
        "#Import"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "whTNzty-fFTW",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from keras.layers.wrappers import Bidirectional"
      ],
      "execution_count": 1,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7vH_B55ooAz4",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import tensorflow as tf"
      ],
      "execution_count": 2,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "la6HCGxkAhD0",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import pandas as pd\n",
        "import numpy as np\n",
        "from sklearn.metrics import classification_report, confusion_matrix\n",
        "from IPython.display import Image, display_png\n",
        "from gensim.models import word2vec, KeyedVectors\n",
        "from keras.models import Sequential\n",
        "from keras.layers import Input, Embedding, Dense, Dropout, Flatten, LSTM,Bidirectional,GlobalAveragePooling1D\n",
        "from keras.utils import to_categorical, plot_model\n",
        "from keras.preprocessing.sequence import pad_sequences\n",
        "from keras.optimizers import RMSprop\n",
        "import matplotlib.pyplot as plt\n",
        "plt.style.use('ggplot')"
      ],
      "execution_count": 3,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "0v0_se-r6dms",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "2cdc720c-b281-4bac-cc6d-6af1267fbb0d"
      },
      "source": [
        "!pip install contractions\n",
        "import pandas as pd\n",
        "import contractions\n",
        "import nltk\n",
        "from nltk.tokenize import word_tokenize\n",
        "nltk.download('all')\n",
        "import spacy\n",
        "nlp = spacy.load(\"en_core_web_sm\")"
      ],
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Collecting contractions\n",
            "  Downloading https://files.pythonhosted.org/packages/00/92/a05b76a692ac08d470ae5c23873cf1c9a041532f1ee065e74b374f218306/contractions-0.0.25-py2.py3-none-any.whl\n",
            "Collecting textsearch\n",
            "  Downloading https://files.pythonhosted.org/packages/42/a8/03407021f9555043de5492a2bd7a35c56cc03c2510092b5ec018cae1bbf1/textsearch-0.0.17-py2.py3-none-any.whl\n",
            "Collecting pyahocorasick\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/f4/9f/f0d8e8850e12829eea2e778f1c90e3c53a9a799b7f412082a5d21cd19ae1/pyahocorasick-1.4.0.tar.gz (312kB)\n",
            "\u001b[K     |████████████████████████████████| 317kB 4.2MB/s \n",
            "\u001b[?25hCollecting Unidecode\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/d0/42/d9edfed04228bacea2d824904cae367ee9efd05e6cce7ceaaedd0b0ad964/Unidecode-1.1.1-py2.py3-none-any.whl (238kB)\n",
            "\u001b[K     |████████████████████████████████| 245kB 38.1MB/s \n",
            "\u001b[?25hBuilding wheels for collected packages: pyahocorasick\n",
            "  Building wheel for pyahocorasick (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "  Created wheel for pyahocorasick: filename=pyahocorasick-1.4.0-cp36-cp36m-linux_x86_64.whl size=81698 sha256=1671e317564a181d9d088dee5dfb8ec8172ca1b9377757a9337d118e26e91bfd\n",
            "  Stored in directory: /root/.cache/pip/wheels/0a/90/61/87a55f5b459792fbb2b7ba6b31721b06ff5cf6bde541b40994\n",
            "Successfully built pyahocorasick\n",
            "Installing collected packages: pyahocorasick, Unidecode, textsearch, contractions\n",
            "Successfully installed Unidecode-1.1.1 contractions-0.0.25 pyahocorasick-1.4.0 textsearch-0.0.17\n",
            "[nltk_data] Downloading collection 'all'\n",
            "[nltk_data]    | \n",
            "[nltk_data]    | Downloading package abc to /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/abc.zip.\n",
            "[nltk_data]    | Downloading package alpino to /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/alpino.zip.\n",
            "[nltk_data]    | Downloading package biocreative_ppi to\n",
            "[nltk_data]    |     /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/biocreative_ppi.zip.\n",
            "[nltk_data]    | Downloading package brown to /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/brown.zip.\n",
            "[nltk_data]    | Downloading package brown_tei to /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/brown_tei.zip.\n",
            "[nltk_data]    | Downloading package cess_cat to /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/cess_cat.zip.\n",
            "[nltk_data]    | Downloading package cess_esp to /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/cess_esp.zip.\n",
            "[nltk_data]    | Downloading package chat80 to /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/chat80.zip.\n",
            "[nltk_data]    | Downloading package city_database to\n",
            "[nltk_data]    |     /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/city_database.zip.\n",
            "[nltk_data]    | Downloading package cmudict to /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/cmudict.zip.\n",
            "[nltk_data]    | Downloading package comparative_sentences to\n",
            "[nltk_data]    |     /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/comparative_sentences.zip.\n",
            "[nltk_data]    | Downloading package comtrans to /root/nltk_data...\n",
            "[nltk_data]    | Downloading package conll2000 to /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/conll2000.zip.\n",
            "[nltk_data]    | Downloading package conll2002 to /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/conll2002.zip.\n",
            "[nltk_data]    | Downloading package conll2007 to /root/nltk_data...\n",
            "[nltk_data]    | Downloading package crubadan to /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/crubadan.zip.\n",
            "[nltk_data]    | Downloading package dependency_treebank to\n",
            "[nltk_data]    |     /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/dependency_treebank.zip.\n",
            "[nltk_data]    | Downloading package dolch to /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/dolch.zip.\n",
            "[nltk_data]    | Downloading package europarl_raw to\n",
            "[nltk_data]    |     /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/europarl_raw.zip.\n",
            "[nltk_data]    | Downloading package floresta to /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/floresta.zip.\n",
            "[nltk_data]    | Downloading package framenet_v15 to\n",
            "[nltk_data]    |     /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/framenet_v15.zip.\n",
            "[nltk_data]    | Downloading package framenet_v17 to\n",
            "[nltk_data]    |     /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/framenet_v17.zip.\n",
            "[nltk_data]    | Downloading package gazetteers to /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/gazetteers.zip.\n",
            "[nltk_data]    | Downloading package genesis to /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/genesis.zip.\n",
            "[nltk_data]    | Downloading package gutenberg to /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/gutenberg.zip.\n",
            "[nltk_data]    | Downloading package ieer to /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/ieer.zip.\n",
            "[nltk_data]    | Downloading package inaugural to /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/inaugural.zip.\n",
            "[nltk_data]    | Downloading package indian to /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/indian.zip.\n",
            "[nltk_data]    | Downloading package jeita to /root/nltk_data...\n",
            "[nltk_data]    | Downloading package kimmo to /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/kimmo.zip.\n",
            "[nltk_data]    | Downloading package knbc to /root/nltk_data...\n",
            "[nltk_data]    | Downloading package lin_thesaurus to\n",
            "[nltk_data]    |     /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/lin_thesaurus.zip.\n",
            "[nltk_data]    | Downloading package mac_morpho to /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/mac_morpho.zip.\n",
            "[nltk_data]    | Downloading package machado to /root/nltk_data...\n",
            "[nltk_data]    | Downloading package masc_tagged to /root/nltk_data...\n",
            "[nltk_data]    | Downloading package moses_sample to\n",
            "[nltk_data]    |     /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping models/moses_sample.zip.\n",
            "[nltk_data]    | Downloading package movie_reviews to\n",
            "[nltk_data]    |     /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/movie_reviews.zip.\n",
            "[nltk_data]    | Downloading package names to /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/names.zip.\n",
            "[nltk_data]    | Downloading package nombank.1.0 to /root/nltk_data...\n",
            "[nltk_data]    | Downloading package nps_chat to /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/nps_chat.zip.\n",
            "[nltk_data]    | Downloading package omw to /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/omw.zip.\n",
            "[nltk_data]    | Downloading package opinion_lexicon to\n",
            "[nltk_data]    |     /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/opinion_lexicon.zip.\n",
            "[nltk_data]    | Downloading package paradigms to /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/paradigms.zip.\n",
            "[nltk_data]    | Downloading package pil to /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/pil.zip.\n",
            "[nltk_data]    | Downloading package pl196x to /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/pl196x.zip.\n",
            "[nltk_data]    | Downloading package ppattach to /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/ppattach.zip.\n",
            "[nltk_data]    | Downloading package problem_reports to\n",
            "[nltk_data]    |     /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/problem_reports.zip.\n",
            "[nltk_data]    | Downloading package propbank to /root/nltk_data...\n",
            "[nltk_data]    | Downloading package ptb to /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/ptb.zip.\n",
            "[nltk_data]    | Downloading package product_reviews_1 to\n",
            "[nltk_data]    |     /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/product_reviews_1.zip.\n",
            "[nltk_data]    | Downloading package product_reviews_2 to\n",
            "[nltk_data]    |     /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/product_reviews_2.zip.\n",
            "[nltk_data]    | Downloading package pros_cons to /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/pros_cons.zip.\n",
            "[nltk_data]    | Downloading package qc to /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/qc.zip.\n",
            "[nltk_data]    | Downloading package reuters to /root/nltk_data...\n",
            "[nltk_data]    | Downloading package rte to /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/rte.zip.\n",
            "[nltk_data]    | Downloading package semcor to /root/nltk_data...\n",
            "[nltk_data]    | Downloading package senseval to /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/senseval.zip.\n",
            "[nltk_data]    | Downloading package sentiwordnet to\n",
            "[nltk_data]    |     /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/sentiwordnet.zip.\n",
            "[nltk_data]    | Downloading package sentence_polarity to\n",
            "[nltk_data]    |     /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/sentence_polarity.zip.\n",
            "[nltk_data]    | Downloading package shakespeare to /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/shakespeare.zip.\n",
            "[nltk_data]    | Downloading package sinica_treebank to\n",
            "[nltk_data]    |     /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/sinica_treebank.zip.\n",
            "[nltk_data]    | Downloading package smultron to /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/smultron.zip.\n",
            "[nltk_data]    | Downloading package state_union to /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/state_union.zip.\n",
            "[nltk_data]    | Downloading package stopwords to /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/stopwords.zip.\n",
            "[nltk_data]    | Downloading package subjectivity to\n",
            "[nltk_data]    |     /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/subjectivity.zip.\n",
            "[nltk_data]    | Downloading package swadesh to /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/swadesh.zip.\n",
            "[nltk_data]    | Downloading package switchboard to /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/switchboard.zip.\n",
            "[nltk_data]    | Downloading package timit to /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/timit.zip.\n",
            "[nltk_data]    | Downloading package toolbox to /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/toolbox.zip.\n",
            "[nltk_data]    | Downloading package treebank to /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/treebank.zip.\n",
            "[nltk_data]    | Downloading package twitter_samples to\n",
            "[nltk_data]    |     /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/twitter_samples.zip.\n",
            "[nltk_data]    | Downloading package udhr to /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/udhr.zip.\n",
            "[nltk_data]    | Downloading package udhr2 to /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/udhr2.zip.\n",
            "[nltk_data]    | Downloading package unicode_samples to\n",
            "[nltk_data]    |     /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/unicode_samples.zip.\n",
            "[nltk_data]    | Downloading package universal_treebanks_v20 to\n",
            "[nltk_data]    |     /root/nltk_data...\n",
            "[nltk_data]    | Downloading package verbnet to /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/verbnet.zip.\n",
            "[nltk_data]    | Downloading package verbnet3 to /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/verbnet3.zip.\n",
            "[nltk_data]    | Downloading package webtext to /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/webtext.zip.\n",
            "[nltk_data]    | Downloading package wordnet to /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/wordnet.zip.\n",
            "[nltk_data]    | Downloading package wordnet_ic to /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/wordnet_ic.zip.\n",
            "[nltk_data]    | Downloading package words to /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/words.zip.\n",
            "[nltk_data]    | Downloading package ycoe to /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/ycoe.zip.\n",
            "[nltk_data]    | Downloading package rslp to /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping stemmers/rslp.zip.\n",
            "[nltk_data]    | Downloading package maxent_treebank_pos_tagger to\n",
            "[nltk_data]    |     /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping taggers/maxent_treebank_pos_tagger.zip.\n",
            "[nltk_data]    | Downloading package universal_tagset to\n",
            "[nltk_data]    |     /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping taggers/universal_tagset.zip.\n",
            "[nltk_data]    | Downloading package maxent_ne_chunker to\n",
            "[nltk_data]    |     /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping chunkers/maxent_ne_chunker.zip.\n",
            "[nltk_data]    | Downloading package punkt to /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping tokenizers/punkt.zip.\n",
            "[nltk_data]    | Downloading package book_grammars to\n",
            "[nltk_data]    |     /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping grammars/book_grammars.zip.\n",
            "[nltk_data]    | Downloading package sample_grammars to\n",
            "[nltk_data]    |     /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping grammars/sample_grammars.zip.\n",
            "[nltk_data]    | Downloading package spanish_grammars to\n",
            "[nltk_data]    |     /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping grammars/spanish_grammars.zip.\n",
            "[nltk_data]    | Downloading package basque_grammars to\n",
            "[nltk_data]    |     /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping grammars/basque_grammars.zip.\n",
            "[nltk_data]    | Downloading package large_grammars to\n",
            "[nltk_data]    |     /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping grammars/large_grammars.zip.\n",
            "[nltk_data]    | Downloading package tagsets to /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping help/tagsets.zip.\n",
            "[nltk_data]    | Downloading package snowball_data to\n",
            "[nltk_data]    |     /root/nltk_data...\n",
            "[nltk_data]    | Downloading package bllip_wsj_no_aux to\n",
            "[nltk_data]    |     /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping models/bllip_wsj_no_aux.zip.\n",
            "[nltk_data]    | Downloading package word2vec_sample to\n",
            "[nltk_data]    |     /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping models/word2vec_sample.zip.\n",
            "[nltk_data]    | Downloading package panlex_swadesh to\n",
            "[nltk_data]    |     /root/nltk_data...\n",
            "[nltk_data]    | Downloading package mte_teip5 to /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/mte_teip5.zip.\n",
            "[nltk_data]    | Downloading package averaged_perceptron_tagger to\n",
            "[nltk_data]    |     /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping taggers/averaged_perceptron_tagger.zip.\n",
            "[nltk_data]    | Downloading package averaged_perceptron_tagger_ru to\n",
            "[nltk_data]    |     /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping\n",
            "[nltk_data]    |       taggers/averaged_perceptron_tagger_ru.zip.\n",
            "[nltk_data]    | Downloading package perluniprops to\n",
            "[nltk_data]    |     /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping misc/perluniprops.zip.\n",
            "[nltk_data]    | Downloading package nonbreaking_prefixes to\n",
            "[nltk_data]    |     /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping corpora/nonbreaking_prefixes.zip.\n",
            "[nltk_data]    | Downloading package vader_lexicon to\n",
            "[nltk_data]    |     /root/nltk_data...\n",
            "[nltk_data]    | Downloading package porter_test to /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping stemmers/porter_test.zip.\n",
            "[nltk_data]    | Downloading package wmt15_eval to /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping models/wmt15_eval.zip.\n",
            "[nltk_data]    | Downloading package mwa_ppdb to /root/nltk_data...\n",
            "[nltk_data]    |   Unzipping misc/mwa_ppdb.zip.\n",
            "[nltk_data]    | \n",
            "[nltk_data]  Done downloading collection all\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "NJ1iTtNi3-9z",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 84
        },
        "outputId": "74996563-0548-494d-c719-5590a205055b"
      },
      "source": [
        "\n",
        "!gdown --id 1bzFpqbhQAaafjwOckOh2abRhrv7CVLeN"
      ],
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Downloading...\n",
            "From: https://drive.google.com/uc?id=1bzFpqbhQAaafjwOckOh2abRhrv7CVLeN\n",
            "To: /content/GoogleNews-vectors-negative300.bin\n",
            "3.64GB [00:37, 97.4MB/s]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "wRw_6qHiGqP9",
        "colab_type": "text"
      },
      "source": [
        "# **Feauture Engineering**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "r7wC3oU0ZFjQ",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 84
        },
        "outputId": "4dc6beee-fe3b-4f60-afc4-34e61afa8095"
      },
      "source": [
        "!gdown --id 1TvPCXiMeBnvsqdUYUzBhzHqok_W5JQwr"
      ],
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Downloading...\n",
            "From: https://drive.google.com/uc?id=1TvPCXiMeBnvsqdUYUzBhzHqok_W5JQwr\n",
            "To: /content/lab4_train.csv\n",
            "\r  0% 0.00/331k [00:00<?, ?B/s]\r100% 331k/331k [00:00<00:00, 46.6MB/s]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Ql3GYT2LhQTi",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "data = pd.read_csv('lab4_train.csv')"
      ],
      "execution_count": 7,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "h54IGT-_i097",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "conflict_word_list = ['but','although','though','despite','spite','nevertheless','however','nonetheless','notwithstanding','contray','contrast']\n",
        "\n",
        "def lemma_filter(string_text):\n",
        "  lemma_filter_list = []\n",
        "  result = []\n",
        "  contracts = contractions.fix(string_text) #เปลี่ยนคำย่อ สมมติว่า didn't เป็น did not \n",
        "  doc = nlp(contracts)\n",
        "  lemma = [token.lemma_ for token in doc]\n",
        "  pos_sentences=nltk.pos_tag(lemma)\n",
        "  for tup in pos_sentences:\n",
        "      if tup[1].startswith(('J','N','R','V')) or tup[0] in conflict_word_list  == True  : #เอา adj noun adv ไว้\n",
        "        lemma_filter_list.append(tup[0])\n",
        "  \n",
        "  return lemma_filter_list\n",
        "\n",
        "data['tokens'] = data['text'].apply(lemma_filter)"
      ],
      "execution_count": 8,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vrHqs9Lc9m6d",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "efb51a27-e0ac-4edf-b429-db8afe166541"
      },
      "source": [
        "lemma_filter(data['text'][0])"
      ],
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['staff', 'be', 'so', 'horrible', '-PRON-']"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 9
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "sagUvJl0jOHg",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 70
        },
        "outputId": "458c796e-bcb2-480c-b927-d621e60e0268"
      },
      "source": [
        "model = KeyedVectors.load_word2vec_format('GoogleNews-vectors-negative300.bin', binary=True)"
      ],
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/smart_open/smart_open_lib.py:254: UserWarning: This function is deprecated, use smart_open.open instead. See the migration notes for details: https://github.com/RaRe-Technologies/smart_open/blob/master/README.rst#migrating-to-the-new-open-function\n",
            "  'See the migration notes for details: %s' % _MIGRATION_NOTES_URL\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "pn0FY7WPhUG9",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "9caef5b3-8ad3-4abf-b60a-1f4cf5fb8362"
      },
      "source": [
        "vocab_in_data = set()\n",
        "check_data_bigram = data['tokens']\n",
        "for row in check_data_bigram:\n",
        "  for bi in row:\n",
        "    vocab_in_data.add(bi)\n",
        "  \n",
        "  \n",
        "print(len(vocab_in_data))"
      ],
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "3544\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_RKI2BVYhc-n",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "vector_dim = model.vector_size"
      ],
      "execution_count": 12,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6RH5RAz9hgki",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "85a84a8b-58b8-432a-c0b2-227278402155"
      },
      "source": [
        "vector_dim"
      ],
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "300"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 13
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "TB5MEboh_FTz",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "max_len = 100"
      ],
      "execution_count": 14,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "a14BR8bxjNm2",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "vocab_size = len(vocab_in_data)  +1"
      ],
      "execution_count": 15,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "HDtS92LSq8-y",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "2640aec4-cf3a-4aa8-9617-ccbdd7434b08"
      },
      "source": [
        "vocab_size"
      ],
      "execution_count": 16,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "3545"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 16
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "XD76fKTYq2pz",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "word_to_index = {'PADDING':0}\n",
        "\n",
        "# make weight matrix of word embedding, vocab size + 1 (for padding)\n",
        "embedding_matrix = np.zeros((vocab_size, vector_dim), dtype=\"float32\")\n",
        "embedding_matrix[0] = np.zeros(vector_dim)\n",
        "\n",
        "for i, word in enumerate(vocab_in_data & set(model.vocab)):\n",
        "    word_to_index[word] = i+1\n",
        "    embedding_matrix[i+1] = model[word] "
      ],
      "execution_count": 17,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "W0_LvbaSIvPH",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import numpy as np\n",
        "train, dev = np.split(data, [int(0.8 * len(data))])"
      ],
      "execution_count": 18,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "o2Da3GMD-38f",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def w2i(tokens):\n",
        "    return [word_to_index.get(word, 0) for word in tokens[:max_len]]\n",
        "\n",
        "# apply functions & convert to np.array\n",
        "input_train = np.array(train['tokens'].apply(w2i).tolist())\n",
        "input_dev = np.array(dev['tokens'].apply(w2i).tolist())\n",
        "\n",
        " \n",
        "# input : zero padding\n",
        "train_x = pad_sequences(input_train, max_len, padding='post')\n",
        "dev_x = pad_sequences(input_dev, max_len, padding='post')"
      ],
      "execution_count": 19,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6v-LFwM4D30D",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "#function for relabel\n",
        "def repolar(polarity): #relabel of polarity\n",
        "    return {'negative':0, 'positive':1, 'conflict':2, 'neutral':3}[polarity]\n",
        "def reaspect(aspect): #relabel of aspect\n",
        "    return {'service':0, 'food':1, 'anecdotes/miscellaneous':2, 'price':3,'ambience':4}[aspect]\n",
        "def index_to_label(polarity_index): #convert back to label\n",
        "    return {0: 'negative', 1: 'positive', 2: 'conflict', 3: 'neutral'}[polarity_index]\n",
        "def index_to_aspect(aspect_index): #convert back to label\n",
        "    return {0:'service', 1:'food', 2:'anecdotes/miscellaneous', 3:'price',4:'ambience'}[aspect_index]"
      ],
      "execution_count": 20,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "2qdltau3BRKv",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# apply functions & convert to np.array\n",
        "label_train = np.array(train['polarity'].apply(repolar).tolist())\n",
        "label_dev = np.array(dev['polarity'].apply(repolar).tolist())\n",
        "\n",
        "aspect_train = np.array(train['aspectCategory'].apply(reaspect).tolist())\n",
        "aspect_dev = np.array(dev['aspectCategory'].apply(reaspect).tolist())\n",
        "\n",
        "# label : one-hot vector\n",
        "train_label = to_categorical(label_train, num_classes=4)\n",
        "dev_label = to_categorical(label_dev, num_classes=4)\n",
        "\n",
        "train_aspect = to_categorical(aspect_train, num_classes=5)\n",
        "dev_aspect = to_categorical(aspect_dev, num_classes=5)"
      ],
      "execution_count": 21,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "OymqvkiBFzeS",
        "colab_type": "text"
      },
      "source": [
        "# **Model for polarity**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "adwDrn6OpcR7",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 353
        },
        "outputId": "34aa213b-b163-4d75-cd23-537851fa5d40"
      },
      "source": [
        "# instantiation\n",
        "p_model = Sequential()\n",
        "\n",
        "# add embedding layer\n",
        "p_model.add(Embedding(input_dim=vocab_size,\n",
        "                    input_length=max_len,\n",
        "                    output_dim=vector_dim, \n",
        "                    weights=[embedding_matrix], \n",
        "                    mask_zero=True,\n",
        "                    trainable=True))\n",
        "\n",
        "# average\n",
        "p_model.add(Bidirectional(LSTM(200,recurrent_dropout = 0.2)))\n",
        "# add hidden layer\n",
        "p_model.add(Dense(250, activation='relu'))\n",
        "p_model.add(Dropout(0.20))\n",
        "p_model.add(Dense(70, activation='relu'))\n",
        "# add output layer\n",
        "p_model.add(Dense(4, activation='softmax'))\n",
        "\n",
        "# compile model\n",
        "p_model.compile(optimizer=\"rmsprop\", loss=\"categorical_crossentropy\", metrics=[\"accuracy\"])\n",
        "\n",
        "p_model.summary()"
      ],
      "execution_count": 22,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "embedding (Embedding)        (None, 100, 300)          1063500   \n",
            "_________________________________________________________________\n",
            "bidirectional (Bidirectional (None, 400)               801600    \n",
            "_________________________________________________________________\n",
            "dense (Dense)                (None, 250)               100250    \n",
            "_________________________________________________________________\n",
            "dropout (Dropout)            (None, 250)               0         \n",
            "_________________________________________________________________\n",
            "dense_1 (Dense)              (None, 70)                17570     \n",
            "_________________________________________________________________\n",
            "dense_2 (Dense)              (None, 4)                 284       \n",
            "=================================================================\n",
            "Total params: 1,983,204\n",
            "Trainable params: 1,983,204\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "NuDa_tANFHWL",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 754
        },
        "outputId": "7055d04d-05fd-4085-f61a-46dd2989712c"
      },
      "source": [
        "plot_model(p_model, show_shapes=True,to_file='model.png')\n",
        "display_png(Image('model.png'))"
      ],
      "execution_count": 23,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n"
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fSZsW0kbJZB3",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 353
        },
        "outputId": "21b26e01-8448-4d67-c0b6-1fed8c6106ac"
      },
      "source": [
        "# train\n",
        "history = p_model.fit(train_x, train_label, batch_size=128, epochs=10, validation_data=(dev_x, dev_label))"
      ],
      "execution_count": 24,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/10\n",
            "20/20 [==============================] - 33s 2s/step - loss: 1.0659 - accuracy: 0.5927 - val_loss: 0.8968 - val_accuracy: 0.6392\n",
            "Epoch 2/10\n",
            "20/20 [==============================] - 33s 2s/step - loss: 0.8090 - accuracy: 0.6823 - val_loss: 0.8046 - val_accuracy: 0.6804\n",
            "Epoch 3/10\n",
            "20/20 [==============================] - 33s 2s/step - loss: 0.6941 - accuracy: 0.7338 - val_loss: 0.7689 - val_accuracy: 0.7009\n",
            "Epoch 4/10\n",
            "20/20 [==============================] - 32s 2s/step - loss: 0.5925 - accuracy: 0.7777 - val_loss: 0.7553 - val_accuracy: 0.6994\n",
            "Epoch 5/10\n",
            "20/20 [==============================] - 32s 2s/step - loss: 0.4829 - accuracy: 0.8142 - val_loss: 0.8290 - val_accuracy: 0.6867\n",
            "Epoch 6/10\n",
            "20/20 [==============================] - 32s 2s/step - loss: 0.4012 - accuracy: 0.8475 - val_loss: 0.8845 - val_accuracy: 0.6835\n",
            "Epoch 7/10\n",
            "20/20 [==============================] - 40s 2s/step - loss: 0.3487 - accuracy: 0.8704 - val_loss: 0.9152 - val_accuracy: 0.7025\n",
            "Epoch 8/10\n",
            "20/20 [==============================] - 33s 2s/step - loss: 0.2772 - accuracy: 0.8922 - val_loss: 1.0079 - val_accuracy: 0.6851\n",
            "Epoch 9/10\n",
            "20/20 [==============================] - 32s 2s/step - loss: 0.2487 - accuracy: 0.8974 - val_loss: 1.2826 - val_accuracy: 0.6155\n",
            "Epoch 10/10\n",
            "20/20 [==============================] - 32s 2s/step - loss: 0.2206 - accuracy: 0.9109 - val_loss: 1.0449 - val_accuracy: 0.6946\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "y6Rf6YELJ3tk",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 282
        },
        "outputId": "cff53fe3-e359-42eb-eaa4-ed8ca287d088"
      },
      "source": [
        "plt.plot(history.history['accuracy'])\n",
        "plt.plot(history.history['val_accuracy'])\n",
        "plt.ylabel('accuracy')\n",
        "plt.xlabel('epoch')\n",
        "plt.legend(['train', 'dev'], loc='best')\n",
        "plt.show()"
      ],
      "execution_count": 25,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "sTfluLwbn8vH",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "polar_prediction = p_model.predict(dev_x)\n",
        "polar_prediction = [index_to_label(np.argmax(i)) for i in polar_prediction]\n"
      ],
      "execution_count": 27,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "uECDdDiBGQM8",
        "colab_type": "text"
      },
      "source": [
        "# **Model for aspect**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1KeWbD-l3V4E",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 353
        },
        "outputId": "9d513112-1a21-498e-f0e8-a9dc72415f0c"
      },
      "source": [
        "#Aspect\n",
        "# instantiation\n",
        "a_model = Sequential()\n",
        "\n",
        "# add embedding layer\n",
        "a_model.add(Embedding(input_dim=vocab_size,\n",
        "                    input_length=max_len,\n",
        "                    output_dim=vector_dim, \n",
        "                    weights=[embedding_matrix], \n",
        "                    mask_zero=True,\n",
        "                    trainable=True))\n",
        "\n",
        "# average\n",
        "a_model.add(Bidirectional(LSTM(200,recurrent_dropout = 0.2)))\n",
        "\n",
        "# add hidden layer\n",
        "a_model.add(Dense(250, activation='relu'))\n",
        "a_model.add(Dropout(0.20))\n",
        "a_model.add(Dense(70, activation='relu'))\n",
        "\n",
        "\n",
        "# add output layer\n",
        "a_model.add(Dense(5, activation='softmax'))\n",
        "\n",
        "# compile model\n",
        "a_model.compile(optimizer=\"rmsprop\", loss=\"categorical_crossentropy\", metrics=[\"accuracy\"])\n",
        "\n",
        "a_model.summary()"
      ],
      "execution_count": 32,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential_2\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "embedding_2 (Embedding)      (None, 100, 300)          1063500   \n",
            "_________________________________________________________________\n",
            "bidirectional_2 (Bidirection (None, 400)               801600    \n",
            "_________________________________________________________________\n",
            "dense_6 (Dense)              (None, 250)               100250    \n",
            "_________________________________________________________________\n",
            "dropout_2 (Dropout)          (None, 250)               0         \n",
            "_________________________________________________________________\n",
            "dense_7 (Dense)              (None, 70)                17570     \n",
            "_________________________________________________________________\n",
            "dense_8 (Dense)              (None, 5)                 355       \n",
            "=================================================================\n",
            "Total params: 1,983,275\n",
            "Trainable params: 1,983,275\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "YTHmcdJ93j-s",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 754
        },
        "outputId": "a120b362-eef9-4868-8ce5-b7b51ca12536"
      },
      "source": [
        "plot_model(a_model, show_shapes=True,to_file='model.png')\n",
        "display_png(Image('model.png'))"
      ],
      "execution_count": 33,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n"
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "sDaEHA6a3uOL",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 353
        },
        "outputId": "3ddbaf03-8507-4cc7-f923-00215e70bb4d"
      },
      "source": [
        "history = a_model.fit(train_x, train_aspect, batch_size=128, epochs=10, validation_data=(dev_x, dev_aspect))"
      ],
      "execution_count": 34,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/10\n",
            "20/20 [==============================] - 32s 2s/step - loss: 1.2595 - accuracy: 0.5218 - val_loss: 1.0290 - val_accuracy: 0.6424\n",
            "Epoch 2/10\n",
            "20/20 [==============================] - 32s 2s/step - loss: 0.8767 - accuracy: 0.6763 - val_loss: 1.2680 - val_accuracy: 0.5190\n",
            "Epoch 3/10\n",
            "20/20 [==============================] - 32s 2s/step - loss: 0.7480 - accuracy: 0.7056 - val_loss: 0.8604 - val_accuracy: 0.6630\n",
            "Epoch 4/10\n",
            "20/20 [==============================] - 32s 2s/step - loss: 0.6123 - accuracy: 0.7496 - val_loss: 0.8796 - val_accuracy: 0.6851\n",
            "Epoch 5/10\n",
            "20/20 [==============================] - 31s 2s/step - loss: 0.5551 - accuracy: 0.7639 - val_loss: 0.8858 - val_accuracy: 0.6756\n",
            "Epoch 6/10\n",
            "20/20 [==============================] - 33s 2s/step - loss: 0.4993 - accuracy: 0.7710 - val_loss: 0.9072 - val_accuracy: 0.6725\n",
            "Epoch 7/10\n",
            "20/20 [==============================] - 31s 2s/step - loss: 0.4712 - accuracy: 0.7722 - val_loss: 0.9339 - val_accuracy: 0.6661\n",
            "Epoch 8/10\n",
            "20/20 [==============================] - 31s 2s/step - loss: 0.4270 - accuracy: 0.7916 - val_loss: 0.9246 - val_accuracy: 0.6835\n",
            "Epoch 9/10\n",
            "20/20 [==============================] - 31s 2s/step - loss: 0.4165 - accuracy: 0.7865 - val_loss: 0.8987 - val_accuracy: 0.6741\n",
            "Epoch 10/10\n",
            "20/20 [==============================] - 32s 2s/step - loss: 0.3919 - accuracy: 0.7991 - val_loss: 0.9809 - val_accuracy: 0.6851\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "YGLGw5rH332L",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 282
        },
        "outputId": "471eaf56-6f9e-4265-ffc0-100eef0693c7"
      },
      "source": [
        "plt.plot(history.history['accuracy'])\n",
        "plt.plot(history.history['val_accuracy'])\n",
        "plt.ylabel('accuracy')\n",
        "plt.xlabel('epoch')\n",
        "plt.legend(['train', 'dev'], loc='best')\n",
        "plt.show()"
      ],
      "execution_count": 35,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8lnnR_Si4F3I",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "aspect_prediction = a_model.predict(dev_x)\n",
        "aspect_prediction = [index_to_aspect(np.argmax(i)) for i in aspect_prediction]"
      ],
      "execution_count": 36,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "z6hPtalbTrHm",
        "colab_type": "text"
      },
      "source": [
        "#Generate file to evalute"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zRPG6Pluk_VE",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 84
        },
        "outputId": "9d63231f-d806-4f4f-f15e-1b8e8c9e4116"
      },
      "source": [
        "!gdown --id 1YtAHCzeZUXGZQ9cimdkkUq4lUk3ZH-I_"
      ],
      "execution_count": 51,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Downloading...\n",
            "From: https://drive.google.com/uc?id=1YtAHCzeZUXGZQ9cimdkkUq4lUk3ZH-I_\n",
            "To: /content/evaluate.py\n",
            "\r  0% 0.00/7.03k [00:00<?, ?B/s]\r100% 7.03k/7.03k [00:00<00:00, 13.6MB/s]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "EHdJOAm44RF9",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 402
        },
        "outputId": "6420632e-551d-436d-e9a8-ba8d96ddb0d3"
      },
      "source": [
        "file = pd.DataFrame()\n",
        "file['id'] = dev['id']\n",
        "file['polarity'] = polar_prediction\n",
        "file['aspectCategory'] = aspect_prediction\n",
        "file.to_csv('pred.csv', index=None)\n",
        "file"
      ],
      "execution_count": 52,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>id</th>\n",
              "      <th>polarity</th>\n",
              "      <th>aspectCategory</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>2524</th>\n",
              "      <td>3235</td>\n",
              "      <td>positive</td>\n",
              "      <td>service</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2525</th>\n",
              "      <td>3480</td>\n",
              "      <td>neutral</td>\n",
              "      <td>anecdotes/miscellaneous</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2526</th>\n",
              "      <td>1459</td>\n",
              "      <td>positive</td>\n",
              "      <td>food</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2527</th>\n",
              "      <td>1459</td>\n",
              "      <td>positive</td>\n",
              "      <td>food</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2528</th>\n",
              "      <td>1597</td>\n",
              "      <td>positive</td>\n",
              "      <td>food</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3151</th>\n",
              "      <td>2853</td>\n",
              "      <td>positive</td>\n",
              "      <td>anecdotes/miscellaneous</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3152</th>\n",
              "      <td>1163</td>\n",
              "      <td>positive</td>\n",
              "      <td>service</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3153</th>\n",
              "      <td>216</td>\n",
              "      <td>positive</td>\n",
              "      <td>anecdotes/miscellaneous</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3154</th>\n",
              "      <td>1109</td>\n",
              "      <td>positive</td>\n",
              "      <td>food</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3155</th>\n",
              "      <td>899</td>\n",
              "      <td>neutral</td>\n",
              "      <td>food</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>632 rows × 3 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "        id  polarity           aspectCategory\n",
              "2524  3235  positive                  service\n",
              "2525  3480   neutral  anecdotes/miscellaneous\n",
              "2526  1459  positive                     food\n",
              "2527  1459  positive                     food\n",
              "2528  1597  positive                     food\n",
              "...    ...       ...                      ...\n",
              "3151  2853  positive  anecdotes/miscellaneous\n",
              "3152  1163  positive                  service\n",
              "3153   216  positive  anecdotes/miscellaneous\n",
              "3154  1109  positive                     food\n",
              "3155   899   neutral                     food\n",
              "\n",
              "[632 rows x 3 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 52
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Biyy-4Sj4YVt",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 403
        },
        "outputId": "88056089-872a-4d47-9621-dba55c93fdc7"
      },
      "source": [
        "!python3 evaluate.py lab4_train.csv pred.csv"
      ],
      "execution_count": 53,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "=== CLASSIFICATION : ASPECT ===\n",
            "                class name  precision  recall  F1-score support\n",
            "0                     food      0.891   0.847     0.869     203\n",
            "1                    price      0.818   0.300     0.439      60\n",
            "2                  service      0.897   0.673     0.769     104\n",
            "3                 ambience      0.850   0.479     0.613      71\n",
            "4  anecdotes/miscellaneous      0.848   0.716     0.777     194\n",
            "5                MACRO AVG      0.861   0.603     0.693     632\n",
            "6                MICRO AVG      0.871   0.685     0.767     632 \n",
            "\n",
            "=== CLASSIFICATION : SENTIMENT ===\n",
            "  class name  precision  recall  F1-score support\n",
            "0   positive      0.814   0.830     0.822     306\n",
            "1   negative      0.600   0.552     0.575     125\n",
            "2    neutral      0.452   0.378     0.412      74\n",
            "3   conflict      0.125   0.042     0.062      24\n",
            "4  MACRO AVG      0.498   0.451     0.468     529\n",
            "5  MICRO AVG      0.708   0.665     0.686     529 \n",
            "\n",
            "=== CLASSIFICATION : OVERALL ===\n",
            "              precision  recall  F1-score support\n",
            "0  MICRO AVG      0.616   0.484     0.542     632 \n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "oJg6XGncTjTP",
        "colab_type": "text"
      },
      "source": [
        "#Generate TEST file"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "U22SaNi1THp0",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 84
        },
        "outputId": "765c500b-8626-4335-bc3f-08af71f4666e"
      },
      "source": [
        "!gdown --id 1EacvwnOHfwa4FiZy2K8mFpFjmpb4Mt-t"
      ],
      "execution_count": 54,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Downloading...\n",
            "From: https://drive.google.com/uc?id=1EacvwnOHfwa4FiZy2K8mFpFjmpb4Mt-t\n",
            "To: /content/lab4_test.csv\n",
            "\r  0% 0.00/36.2k [00:00<?, ?B/s]\r100% 36.2k/36.2k [00:00<00:00, 13.3MB/s]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "RDh26pSZTVTj",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "test = pd.read_csv('lab4_test.csv')"
      ],
      "execution_count": 55,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "sshK121QTK4E",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "test['tokens'] = test['text'].apply(lemma_filter)"
      ],
      "execution_count": 56,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lqd6aw0yYzkP",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def w2i(tokens):\n",
        "    return [word_to_index.get(word, 0) for word in tokens[:max_len]]\n",
        "\n",
        "# apply functions & convert to np.array\n",
        "input_test = np.array(test['tokens'].apply(w2i).tolist()) \n",
        "# input : zero padding\n",
        "\n",
        "test_x = pad_sequences(input_test, max_len, padding='post')"
      ],
      "execution_count": 57,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "57YkzHDWdPBy",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "label_train = np.array(train['polarity'].apply(repolar).tolist())\n",
        " # label_dev = np.array(dev['polarity'].apply(repolar).tolist())\n",
        "\n",
        "aspect_train = np.array(train['aspectCategory'].apply(reaspect).tolist())\n",
        "#aspect_dev = np.array(dev['aspectCategory'].apply(reaspect).tolist())\n",
        "\n",
        "# label : one-hot vector\n",
        "train_label = to_categorical(label_train, num_classes=4)\n",
        "#dev_label = to_categorical(label_dev, num_classes=4)\n",
        "\n",
        "train_aspect = to_categorical(aspect_train, num_classes=5)\n",
        "#dev_aspect = to_categorical(aspect_dev, num_classes=5)"
      ],
      "execution_count": 58,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "DwhXsaIxS8BE",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "polar_p = p_model.predict(test_x)\n",
        "polar_p = [index_to_label(np.argmax(i)) for i in polar_p]"
      ],
      "execution_count": 59,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "XGnQMpUUbfXz",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 195
        },
        "outputId": "680c42d5-5fc7-4e39-e35e-081a31cc221a"
      },
      "source": [
        "test.head()"
      ],
      "execution_count": 60,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>id</th>\n",
              "      <th>text</th>\n",
              "      <th>tokens</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>899</td>\n",
              "      <td>Food and service was okay.</td>\n",
              "      <td>[food, service, be, okay]</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>1349</td>\n",
              "      <td>great for a romantic evening, or a fun evening...</td>\n",
              "      <td>[great, romantic, evening, fun, evening, friend]</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>934</td>\n",
              "      <td>The lamb meat was under-cooked and EXTRMELY CH...</td>\n",
              "      <td>[lamb, meat, be, cook, extrmely, CHEWY]</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>2199</td>\n",
              "      <td>The best pad thai i've ever had.</td>\n",
              "      <td>[good, pad, thai, -PRON-, have, ever, have]</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>188</td>\n",
              "      <td>Over time, the food quality has decreased subs...</td>\n",
              "      <td>[time, food, quality, have, decrease, substant...</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "     id  ...                                             tokens\n",
              "0   899  ...                          [food, service, be, okay]\n",
              "1  1349  ...   [great, romantic, evening, fun, evening, friend]\n",
              "2   934  ...            [lamb, meat, be, cook, extrmely, CHEWY]\n",
              "3  2199  ...        [good, pad, thai, -PRON-, have, ever, have]\n",
              "4   188  ...  [time, food, quality, have, decrease, substant...\n",
              "\n",
              "[5 rows x 3 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 60
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "gLyu2KrVb4E2",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "e332a792-59a3-4f9a-a120-a91e94861eef"
      },
      "source": [
        "len(polar_p)"
      ],
      "execution_count": 61,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "461"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 61
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4Co6AJBFb72_",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "385971f1-4a85-4d70-c2f6-9b49e066b424"
      },
      "source": [
        "len(test)"
      ],
      "execution_count": 62,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "461"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 62
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "t_2HbNGuTAik",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "aspect_p = a_model.predict(test_x)\n",
        "aspect_p = [index_to_aspect(np.argmax(i)) for i in aspect_p]"
      ],
      "execution_count": 63,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "FJLHn9BlW2i2",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 402
        },
        "outputId": "f910543f-7277-4d93-ffd2-6cfbb8002689"
      },
      "source": [
        "nfile = pd.DataFrame()\n",
        "nfile['id'] = test['id']\n",
        "nfile['polarity'] = polar_p\n",
        "nfile['aspectCategory'] = aspect_p\n",
        "nfile.to_csv('test_pred.csv', index=None)\n",
        "nfile"
      ],
      "execution_count": 64,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>id</th>\n",
              "      <th>polarity</th>\n",
              "      <th>aspectCategory</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>899</td>\n",
              "      <td>neutral</td>\n",
              "      <td>food</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>1349</td>\n",
              "      <td>positive</td>\n",
              "      <td>anecdotes/miscellaneous</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>934</td>\n",
              "      <td>positive</td>\n",
              "      <td>food</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>2199</td>\n",
              "      <td>positive</td>\n",
              "      <td>food</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>188</td>\n",
              "      <td>negative</td>\n",
              "      <td>food</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>456</th>\n",
              "      <td>1063</td>\n",
              "      <td>positive</td>\n",
              "      <td>anecdotes/miscellaneous</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>457</th>\n",
              "      <td>777</td>\n",
              "      <td>positive</td>\n",
              "      <td>food</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>458</th>\n",
              "      <td>875</td>\n",
              "      <td>negative</td>\n",
              "      <td>anecdotes/miscellaneous</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>459</th>\n",
              "      <td>671</td>\n",
              "      <td>positive</td>\n",
              "      <td>food</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>460</th>\n",
              "      <td>617</td>\n",
              "      <td>neutral</td>\n",
              "      <td>anecdotes/miscellaneous</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>461 rows × 3 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "       id  polarity           aspectCategory\n",
              "0     899   neutral                     food\n",
              "1    1349  positive  anecdotes/miscellaneous\n",
              "2     934  positive                     food\n",
              "3    2199  positive                     food\n",
              "4     188  negative                     food\n",
              "..    ...       ...                      ...\n",
              "456  1063  positive  anecdotes/miscellaneous\n",
              "457   777  positive                     food\n",
              "458   875  negative  anecdotes/miscellaneous\n",
              "459   671  positive                     food\n",
              "460   617   neutral  anecdotes/miscellaneous\n",
              "\n",
              "[461 rows x 3 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 64
        }
      ]
    }
  ]
}