Sentiment_analysis_Yelp_ENSAE_TIT_Parziale.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Sentiment_analysis_Yelp_ENSAE_TIT_Parziale.ipynb",
      "provenance": [],
      "collapsed_sections": [
        "F0r3O8VWF2rr",
        "WoGmmE-ysY54"
      ],
      "machine_shape": "hm",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/karimtito/Machine_learning_NLP_Yelp/blob/master/Sentiment_analysis_Yelp_ENSAE_TIT_Parziale.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "on5hMdSZthDa",
        "colab_type": "text"
      },
      "source": [
        "## 1 Importing required librairies, loading and quickly exploring the data \n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "58vol3vgEz64",
        "colab_type": "code",
        "outputId": "4bc32a15-86f0-420f-a9b0-e8e92b3bf18f",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 193
        }
      },
      "source": [
        "import pandas as pd                         #used to load csv files and \n",
        "import numpy as np\n",
        "import re                                   #regular expressions use to clean the text\n",
        "import string\n",
        "from tqdm import tqdm\n",
        "from time import time\n",
        "from wordcloud import WordCloud\n",
        "import nltk\n",
        "from nltk import pos_tag\n",
        "from nltk.corpus import stopwords\n",
        "from nltk.tokenize import WhitespaceTokenizer\n",
        "from nltk.stem import WordNetLemmatizer\n",
        "from gensim.models.phrases import Phrases, Phraser\n",
        "from nltk.tokenize import TreebankWordTokenizer\n",
        "nltk.download('punkt')\n",
        "sent_detector = nltk.data.load('tokenizers/punkt/english.pickle')\n",
        "\n",
        "nltk.download('stopwords')\n",
        "\n",
        "nltk.download('averaged_perceptron_tagger')\n",
        "\n",
        "nltk.download('wordnet')\n",
        "import spacy\n",
        "tokenizer = spacy.load(\"en_core_web_sm\")\n",
        "import os\n",
        "\n",
        "n = 30000                                 #total number of observations\n",
        "s = 15000                                 #number of samples\n",
        "prop_train_test = 0.9                          #proportion of samples for training and testing vs. validation\n",
        "prop_train = 0.8                               #proportion of samples for training vs. testing\n",
        "\n",
        "\n",
        "\n",
        "import numpy as np\n",
        "import seaborn as sns\n",
        "\n",
        "import matplotlib.pyplot as plt\n",
        "%matplotlib inline\n",
        "df = pd.read_csv(\"https://raw.githubusercontent.com/karimtito/Machine_learning_NLP_Yelp/master/yelp_review_30K.csv\",nrows = s)"
      ],
      "execution_count": 45,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[nltk_data] Downloading package punkt to /root/nltk_data...\n",
            "[nltk_data]   Package punkt is already up-to-date!\n",
            "[nltk_data] Downloading package stopwords to /root/nltk_data...\n",
            "[nltk_data]   Package stopwords is already up-to-date!\n",
            "[nltk_data] Downloading package averaged_perceptron_tagger to\n",
            "[nltk_data]     /root/nltk_data...\n",
            "[nltk_data]   Package averaged_perceptron_tagger is already up-to-\n",
            "[nltk_data]       date!\n",
            "[nltk_data] Downloading package wordnet to /root/nltk_data...\n",
            "[nltk_data]   Package wordnet is already up-to-date!\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "oHtIvfLJbyXi",
        "colab_type": "code",
        "outputId": "706d48b5-4419-4b3c-b4fc-c01d36d50f1f",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 52
        }
      },
      "source": [
        "\n",
        "%load_ext autoreload\n",
        "%autoreload 2\n",
        "\n",
        "import torch                                # Importation of PyTorch to train the LSTM\n",
        "import torch.nn as nn\n",
        "import torch.nn.functional as F\n",
        "\n",
        "torch.manual_seed(1)\n",
        "\n",
        "from sklearn.tree import plot_tree                          #used to plot the tree\n",
        "from sklearn.metrics import confusion_matrix                #use the obtain ROC and other metrics for all the models\n"
      ],
      "execution_count": 46,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "The autoreload extension is already loaded. To reload it, use:\n",
            "  %reload_ext autoreload\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Ewzrssc_V3HZ",
        "colab_type": "code",
        "outputId": "30ce3ea6-91f1-4ead-c843-a80213baf0b8",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 228
        }
      },
      "source": [
        "positive_review = lambda x : np.heaviside(x-3,1)         \n",
        "#Nota Bene: to simplify we use '1' for a positive review and '0' for a negative review, even though we are more interested in the \n",
        "#performance for labelling negative reviews i.e. to detect them well\n",
        "labels = np.array(df[\"stars\"].apply(positive_review))\n",
        "df[\"label\"] = labels\n",
        "pos_freq = labels.mean()                         #sample frequency of positive reviews\n",
        "df.head(10),labels.mean()  #we look at the 10 first rows and at the sample frequency of positive reviews\n"
      ],
      "execution_count": 47,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(   Unnamed: 0  stars                                               text  label\n",
              " 0           0      5  Super simple place but amazing nonetheless. It...    1.0\n",
              " 1           1      5  Small unassuming place that changes their menu...    1.0\n",
              " 2           2      5  Lester's is located in a beautiful neighborhoo...    1.0\n",
              " 3           3      4  Love coming here. Yes the place always needs t...    1.0\n",
              " 4           4      4  Had their chocolate almond croissant and it wa...    1.0\n",
              " 5           5      5  Cycle Pub Las Vegas was a blast! Got a groupon...    1.0\n",
              " 6           6      4  Who would have guess that you would be able to...    1.0\n",
              " 7           7      4  Always drove past this coffee house and wonder...    1.0\n",
              " 8           8      3  Not bad!! Love that there is a gluten-free, ve...    1.0\n",
              " 9           9      5  Love this place!\\n\\nPeggy is great with dogs a...    1.0,\n",
              " 0.8195333333333333)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 47
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "bMerSDTXYMkM",
        "colab_type": "text"
      },
      "source": [
        "We see that there is a pretty big imbalance : about 80% of positive reviews vs. 20% of negative reviews\n",
        "\n",
        "\n",
        "> One thing that can be done to improve on that is to sample weights in the training of algorithms\n",
        "\n",
        "The idea is to consider a weighted version of the empirical loss $l_{n,\\omega}(f,D_n) = \\sum_{i :Y_i = 0}\\omega_{0}l(f(X_i),Y_i)+\\sum_{i :Y_i = 1}\\omega_{1}l(f(X_i),Y_i)$, with the condition $\\omega_{0}\\#\\{i :Y_i = 0\\}+\\omega_{1}\\#\\{i :Y_i = 1\\}=1$. For example one can take $\\omega_0 = (1-N_1/n)/N_0$ and $\\omega_1 = \\frac{1}{n}$. We use this approach to train the CART algorithm with Bag-of-words.\n",
        "\n",
        "\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "W5-myrJbmzQz",
        "colab_type": "text"
      },
      "source": [
        "Below we drop unecessary/unusable columns and rows and we get the training, testing and validation labels."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "nGBfgO3QW0GX",
        "colab_type": "code",
        "outputId": "012bd745-007f-4202-eb9b-de10519c49f9",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "df2 =df.drop(labels = ['Unnamed: 0', 'stars'],axis =1)\n",
        "print(df2.shape)\n",
        "df2.dropna()\n",
        "Y = df2.label.to_numpy()\n",
        "n_train_test = int(s*prop_train_test)\n",
        "n_train = int(n_train_test*prop_train)\n",
        "Y_train,Y_test,Y_validation = Y[:n_train],Y[n_train:n_train_test],Y[n_train_test:]\n"
      ],
      "execution_count": 48,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(15000, 2)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "nXNxVQdl2Obd",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "sample_weights = (1./s)*np.ones(s)\n",
        "N_0,N_1 = s-Y.sum(),Y.sum()\n",
        "sample_weights[Y==0] = (1-N_1/n)/N_0"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Dt1Vgl012pC8",
        "colab_type": "text"
      },
      "source": [
        "### Wordcloud visualization:\n",
        "A visualization method that displays how frequently words appear in a given body of text, by making the size of each word proportional to its frequency. All the words are then arranged in a cluster or cloud of words."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vVr6VFax2tXE",
        "colab_type": "code",
        "outputId": "c4ea7bdf-4066-4e52-9814-3d6a9ac78b69",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 473
        }
      },
      "source": [
        "def show_wordcloud(data, title = None):\n",
        "    wordcloud = WordCloud(\n",
        "        background_color = 'white',\n",
        "        max_words = 50,\n",
        "        max_font_size = 40, \n",
        "        scale = 3,\n",
        "        random_state = 42\n",
        "    ).generate(str(data))\n",
        "\n",
        "    fig = plt.figure(1, figsize = (20, 20))\n",
        "    plt.axis('off')\n",
        "    if title: \n",
        "        fig.suptitle(title, fontsize = 20)\n",
        "        fig.subplots_adjust(top = 2.3)\n",
        "\n",
        "    plt.imshow(wordcloud)\n",
        "    plt.show()\n",
        "    \n",
        "df[\"text\"]  = df2[\"text\"].apply(lambda x : x.replace('\\n',''))\n",
        "# print wordcloud\n",
        "show_wordcloud(df[\"text\"].to_list())"
      ],
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1440x1440 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "O0RkCYEaHNUt",
        "colab_type": "text"
      },
      "source": [
        "## 2. Implementation of LSTM with Word2vec embedding\n",
        "We begin by cleaning the corpus of reviews and training a Word2Vec embedding."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "IbUBBe5Xoa73",
        "colab_type": "code",
        "outputId": "3391c672-6ec7-491f-f5c3-2e341aeb967a",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 52
        }
      },
      "source": [
        "def tokenized_sentence(corpus):\n",
        "  tokenizer = TreebankWordTokenizer()\n",
        "  tokenized_sentences = []\n",
        "  for sample in tqdm(corpus):\n",
        "    # separating sentences\n",
        "    for sentence in sent_detector.tokenize(sample):\n",
        "      tokens = tokenizer.tokenize(sentence)\n",
        "      tokens = list(map(lambda x: x.lower() , tokens))\n",
        "      tokens = list(map(lambda x: x.replace(\"\\n\",\" \").replace(\"\\r\",\"\").replace(\",\",\" \").replace(\".\",\" \").replace(\"!\",\" \").replace(\"\\\\/\", \" \") , tokens))\n",
        "      tokens = [ t for t in tokens if len(t)>1 ]\n",
        "      tokenized_sentences.append(tokens)\n",
        "  return tokenized_sentences\n",
        "\n",
        "def clean_corpus(corpus, threshold=50):\n",
        "  tokenized_sentences = tokenized_sentence(corpus)\n",
        "  phrases = Phrases(tokenized_sentences, threshold=threshold)\n",
        "\n",
        "  # This lets you use it with less RAM and faster processing.\n",
        "  # But it will no longer be possible to update the detector with new training \n",
        "  # samples\n",
        "  phraser = Phraser(phrases)\n",
        "\n",
        "  # Merging multi-word expressions in the tokenization\n",
        "  clean_corpus = []\n",
        "  for sentence in tokenized_sentences:\n",
        "    clean_corpus.append(phraser[sentence])\n",
        "  \n",
        "  return clean_corpus\n",
        "\n",
        "\n",
        "corpus = clean_corpus(df2.text.array[:n_train])\n",
        "from gensim.models import Word2Vec\n",
        "from multiprocessing import cpu_count\n",
        "\n",
        "cpu = cpu_count()\n",
        "print('The virtual instance has {} cpus, that will be used to train the word2vec model'.format(cpu))\n"
      ],
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "100%|██████████| 10800/10800 [00:14<00:00, 725.49it/s]\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "The virtual instance has 4 cpus, that will be used to train the word2vec model\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "AKi7fbZA8Dnl",
        "colab_type": "code",
        "outputId": "17828ab6-1b66-4109-eac0-ce6b93fd0e90",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 246
        }
      },
      "source": [
        "corpus[1237]"
      ],
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['nothing',\n",
              " 'too',\n",
              " 'outside',\n",
              " 'of',\n",
              " 'the',\n",
              " 'box',\n",
              " 'but',\n",
              " 'that',\n",
              " \"'s\",\n",
              " 'not',\n",
              " 'necessarily',\n",
              " 'bad',\n",
              " 'thing']"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 13
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "NHVLCJ8M2jh4",
        "colab_type": "text"
      },
      "source": [
        "We now an ad hoc train the Word2Vec representation on our cleaned corpus."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "k-Wz7ReH2h6M",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "d= 25\n",
        "model = Word2Vec(corpus, size=d, window=5, min_count=3, workers=cpu)\n",
        "model.train(corpus, total_examples=len(corpus), epochs=10)\n",
        "model_wv = model.wv"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "IYGiBHJKGJWp",
        "colab_type": "code",
        "outputId": "15fb52dd-62fe-49bb-fa0c-7fe348a5dff2",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 72
        }
      },
      "source": [
        "model_wv.save(\"./word2vec50d.model\")"
      ],
      "execution_count": 16,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/smart_open/smart_open_lib.py:410: 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": "markdown",
      "metadata": {
        "id": "NULl7gsQv696",
        "colab_type": "text"
      },
      "source": [
        "We next check qualitatively the relevance of the Word2vec embeding"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_S64t3CoWiHV",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def most_similar(w,n=5):\n",
        "  if w in model_wv:\n",
        "    closest_ws =[w[0] for w in model_wv.most_similar(w, topn=n)]\n",
        "    print(\"According to our ad-hoc Word2vec, the closest {} words to \\\"{}\\\" \" \n",
        "    \"are: \\n {}\".format( n, w, '\"'+'\",\\n then \"'.join(closest_ws)+'\"'))\n",
        "  else:\n",
        "    print(w, \" could not be found in the Word2vec vocabulary\")"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-h0QruNPAjW1",
        "colab_type": "code",
        "outputId": "d8109277-2e53-4216-bd51-aab1a93fd26b",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 389
        }
      },
      "source": [
        "most_similar(\"restaurant\")\n",
        "most_similar(\"amazing\")\n",
        "most_similar(\"confusing\")"
      ],
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "According to our ad-hoc Word2vec, the closest 5 words to \"restaurant\" are: \n",
            " \"location\",\n",
            " then \"place\",\n",
            " then \"area\",\n",
            " then \"theater\",\n",
            " then \"hotel\"\n",
            "According to our ad-hoc Word2vec, the closest 5 words to \"amazing\" are: \n",
            " \"awesome\",\n",
            " then \"fantastic\",\n",
            " then \"excellent\",\n",
            " then \"incredible\",\n",
            " then \"fabulous\"\n",
            "According to our ad-hoc Word2vec, the closest 5 words to \"confusing\" are: \n",
            " \"tight\",\n",
            " then \"tricky\",\n",
            " then \"dated\",\n",
            " then \"sensitive\",\n",
            " then \"sparse\"\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/gensim/matutils.py:737: FutureWarning: Conversion of the second argument of issubdtype from `int` to `np.signedinteger` is deprecated. In future, it will be treated as `np.int64 == np.dtype(int).type`.\n",
            "  if np.issubdtype(vec.dtype, np.int):\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "75w5lzAfHySn",
        "colab_type": "text"
      },
      "source": [
        "The embedding seems good enough to train directly an LSTM network on top of it."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "PE7SoLo9xtJt",
        "colab_type": "text"
      },
      "source": [
        "### 2.2 Training an LSTM with the trainded Word2Vec embedding\n",
        "\n",
        "We first write a function to embed any sequence using the above Word2Vec model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "NhRG6Xv155e4",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def prepare_embedding(seq, embedding =model_wv ):\n",
        "    idxs = [embedding[w] if w in corpus else np.random.normal(size = d) for w in seq]\n",
        "    # note : Word2Vec can't handle out-of-vocabulary, thus we resort to a random vector for out-vocabulary words\n",
        "    idxs = torch.tensor(idxs,dtype =torch.float).unsqueeze(0)\n",
        "    return idxs\n",
        "\n",
        "  "
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "mY8Al7ICQyAN",
        "colab_type": "text"
      },
      "source": [
        "We then clean and tokenize the train,test and validation data using an ad-hoc tokenizer."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "x0iVJmyvQxKV",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from nltk.corpus import wordnet\n",
        "\n",
        "def get_wordnet_pos(pos_tag):\n",
        "    if pos_tag.startswith('J'):\n",
        "        return wordnet.ADJ\n",
        "    elif pos_tag.startswith('V'):\n",
        "        return wordnet.VERB\n",
        "    elif pos_tag.startswith('N'):\n",
        "        return wordnet.NOUN\n",
        "    elif pos_tag.startswith('R'):\n",
        "        return wordnet.ADV\n",
        "    else:\n",
        "        return wordnet.NOUN\n",
        "def clean_tokenized_text(text):\n",
        "    #Ad-hoc cleaning and tokenizing function\n",
        "    # lower text\n",
        "    text = text.lower()\n",
        "    # remove line jumps and tabs\n",
        "    text = text.replace(\"\\n\", '').replace('\\r', '').replace('\\t','').replace(\"\\\\/\",'-').replace(\"!\",' ').replace(\".\",' ').replace(\",\",\" \")\n",
        "    # tokenize text and remove puncutation\n",
        "    text = [word.strip(string.punctuation) for word in text.split(\" \")]           # The character \"'\" is not considered as punctation\n",
        "    # remove words that contain numbers\n",
        "    text = [word for word in text if not any(c.isdigit() for c in word)]\n",
        "    # remove stop words\n",
        "    stop = stopwords.words('english')\n",
        "    text = [x for x in text if x not in stop]\n",
        "    # remove empty tokens\n",
        "    text = [t for t in text if len(t) > 0]      \n",
        "    # pos tag text\n",
        "    pos_tags = pos_tag(text)\n",
        "    # lemmatize text\n",
        "    text = [WordNetLemmatizer().lemmatize(t[0], get_wordnet_pos(t[1])) for t in pos_tags]\n",
        "    # remove words with only one letter\n",
        "    text = [t for t in text if len(t) > 1]   # we are not interested in one-character expressions such as 'a' or '&'\n",
        "    # join all\n",
        "    return(text)   \n",
        "\n",
        "df3 = df2.text.apply(clean_tokenized_text)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "N3xdKnN1SZPn",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "7a4403f1-5f3c-4cf5-fd95-074f98bc2d50"
      },
      "source": [
        "x  = df3.to_numpy()\n",
        "data = []\n",
        "for i in range(len(x)):\n",
        "    data.append((x[i], Y[i]))                 # vector Y is defined above and simply contains the labels\n",
        "training_data = data[:n_train]\n",
        "test_data = data[n_train:n_train_test]\n",
        "validation_data = data[n_train_test:]\n",
        "print(\"Got {} training examples, {} test examples, {} validation examples\".format(len(training_data), len(test_data),len(validation_data)))\n"
      ],
      "execution_count": 52,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Got 10800 training examples, 2700 test examples, 1500 validation examples\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "QHwi8ChS8xhU",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# defining the classifier model\n",
        "\n",
        "class SequenceLabeller(nn.Module):\n",
        "\n",
        "    def __init__(self, embedding_dim, hidden_dim, num_classes=2, sequence_model=\"LSTM\"):\n",
        "        super(SequenceLabeller, self).__init__()\n",
        "\n",
        "        self.hidden_dim = hidden_dim\n",
        "        #self.word_embeddings  /!\\ no embedding layer we use directly a pre-trained word2vec\n",
        "        \n",
        "        if sequence_model == \"LSTM\":\n",
        "            # The LSTM takes word embeddings as inputs, and outputs hidden states\n",
        "            # with dimensionality hidden_dim.\n",
        "            self.seq = nn.LSTM(embedding_dim, hidden_dim,batch_first=True)\n",
        "        else:\n",
        "            raise(Exception(\"Sequence model {} not supported\".format(sequence_model)))\n",
        "\n",
        "        # The linear layer that maps from hidden state space to class space\n",
        "        self.hidden2tag = nn.Linear(hidden_dim, num_classes)\n",
        "\n",
        "    def forward(self, embeded_sentence):\n",
        "        seq_output, (h_n,c_n) = self.seq(embeded_sentence)\n",
        "        tag_space = self.hidden2tag(seq_output)\n",
        "        # Although we will be performing binary classification, this is a \n",
        "        # general implementation displaying what should be done for n_classes>2\n",
        "        tag_scores = F.log_softmax(tag_space, dim=2) # Computationally efficient for the criterion\n",
        "        return tag_scores"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "LOPnFh3NdfuR",
        "colab_type": "text"
      },
      "source": [
        "We run a first pass without training to check that everything works.\n",
        "\n",
        "\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "N0U2PMub-K6h",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "defb022d-b36f-4a1e-e4e0-afe616c58c8f"
      },
      "source": [
        "model = SequenceLabeller(embedding_dim= d, hidden_dim = 30) \n",
        "inputs = prepare_embedding(training_data[0][0])\n",
        "with torch.no_grad():\n",
        "    sent_scores = model(torch.cat([inputs]))\n",
        "    print(\"Input tokens {} scores {} \".format(inputs, np.exp(sent_scores)))"
      ],
      "execution_count": 54,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Input tokens tensor([[[ 1.9112e+00, -2.2103e+00, -4.3647e-01,  2.1100e+00,  9.6652e-01,\n",
            "           1.2168e+00,  5.9171e-01,  4.7506e-01, -3.9234e-01, -1.1213e+00,\n",
            "          -2.4166e-02,  6.5831e-01,  3.0170e-01, -4.9414e-02,  4.9106e-02,\n",
            "          -5.1305e-01, -2.0335e+00,  1.3884e+00, -1.5393e+00, -3.2886e-01,\n",
            "           9.3323e-01, -5.5576e-01,  1.4451e+00,  9.9448e-01, -1.6593e+00],\n",
            "         [-6.9960e-02,  1.1995e+00,  6.8070e-01,  4.6019e-01, -1.5150e+00,\n",
            "           1.0827e+00,  1.8741e+00,  1.6282e+00,  1.2344e-01, -4.6022e-01,\n",
            "          -8.0915e-01,  8.4770e-01,  7.7277e-01,  8.0708e-01,  4.5674e-01,\n",
            "          -1.2088e+00,  1.5459e+00,  1.2602e-01, -9.7572e-01, -1.9069e+00,\n",
            "           4.1210e-02, -1.4017e+00,  1.1802e-01, -2.7362e-01,  4.9176e-02],\n",
            "         [ 1.3796e+00, -5.6618e-01, -3.4138e-01,  9.1363e-01, -4.5381e-01,\n",
            "          -7.9371e-01,  6.1957e-01,  3.6740e-01,  8.9039e-01, -5.0401e-02,\n",
            "           4.7010e-01, -8.7338e-02,  9.5954e-01, -1.2830e+00, -7.7789e-01,\n",
            "           3.2232e-01, -2.8389e-01, -2.2596e+00,  4.4329e-02,  1.1860e+00,\n",
            "           9.9526e-01, -7.2252e-01, -4.9566e-01,  3.7076e-01,  1.6928e-01],\n",
            "         [-1.4608e-01,  4.4094e-01,  1.5165e-01, -2.7846e-01,  2.5536e-01,\n",
            "          -6.0403e-01, -4.6404e-01,  1.3553e+00, -1.4313e+00,  7.3623e-01,\n",
            "          -3.8566e-01,  1.4184e+00, -1.8067e+00, -3.6126e-01,  1.8054e-01,\n",
            "           1.4499e+00, -8.7802e-01, -4.9992e-02,  5.3576e-01,  7.0225e-02,\n",
            "          -5.5842e-01, -1.3598e-03,  8.6752e-01,  9.4409e-01, -2.1233e+00],\n",
            "         [-5.3294e-01,  1.1120e-02,  7.1814e-01, -2.9943e-01,  3.6900e-01,\n",
            "          -2.9993e-01, -2.3578e-01,  5.7018e-01, -7.6496e-02, -4.3369e-01,\n",
            "           4.0170e-02, -1.1821e+00,  9.7243e-01, -4.4728e-01, -6.0972e-01,\n",
            "           8.1448e-01,  5.8341e-02, -1.1421e+00, -3.4933e-01,  1.7008e+00,\n",
            "          -3.9385e-01,  8.4382e-02, -5.3438e-01,  3.8752e-01, -1.3960e-01],\n",
            "         [ 4.9265e-01,  5.9242e-01, -1.4415e-02, -6.2139e-01, -1.2840e-02,\n",
            "           1.0874e+00, -3.0855e+00, -2.1326e-01,  4.6952e-01,  3.0448e-01,\n",
            "          -1.1404e+00, -2.7194e-01, -8.6393e-01, -4.9705e-01, -8.9461e-01,\n",
            "          -1.2115e+00,  8.9507e-01, -8.1901e-01, -1.0745e+00,  5.9800e-01,\n",
            "          -1.5259e+00,  7.4984e-01, -1.0744e+00, -2.3132e-01,  4.8261e-01],\n",
            "         [ 1.5751e+00,  8.0563e-01,  6.2677e-01, -9.2619e-01,  1.0351e+00,\n",
            "           1.5494e+00,  5.3172e-01, -1.9510e-01,  5.6829e-02,  6.7512e-01,\n",
            "           7.7279e-01, -1.0903e+00, -7.9055e-01, -2.5603e+00, -7.7773e-01,\n",
            "           2.2312e+00,  1.4626e-01, -1.1505e-01,  8.8580e-01,  9.9417e-01,\n",
            "           6.6126e-01, -6.6707e-01, -2.5967e-01,  6.4588e-01, -5.5669e-03],\n",
            "         [-3.8505e-01, -1.1815e+00, -1.6480e-02, -1.5778e+00, -8.4851e-01,\n",
            "          -1.2368e+00, -1.4141e-01, -7.4373e-01, -2.3094e-01,  9.3744e-01,\n",
            "           2.0986e-02,  5.1529e-01, -4.2722e-01,  1.0036e+00, -3.3541e-01,\n",
            "          -6.1381e-01, -1.4372e+00, -4.9582e-01, -2.7768e-01, -3.2925e-01,\n",
            "          -1.6549e+00, -3.9728e-01, -4.1473e-01, -1.6837e+00,  3.1945e-02],\n",
            "         [-1.9956e-01,  1.5936e+00, -1.0811e+00, -6.4337e-01,  7.3151e-01,\n",
            "          -1.8837e-01,  2.9959e-02,  4.5443e-01,  1.7957e+00,  1.1152e+00,\n",
            "          -1.1527e+00, -1.1941e+00, -1.9977e+00,  6.7211e-01, -4.8947e-01,\n",
            "           5.0213e-01,  1.7907e-02,  2.0217e-01, -1.3861e+00,  1.1494e-01,\n",
            "           7.8783e-01,  1.8890e+00, -1.7333e-01,  1.3393e-01, -5.5499e-01],\n",
            "         [-4.0883e-01, -3.4920e-01, -1.9985e-01, -1.1188e+00,  3.9337e-01,\n",
            "          -3.6011e-01,  1.6985e+00,  7.8498e-02,  8.0213e-02, -2.0252e+00,\n",
            "           7.4651e-01, -1.0794e+00,  1.2654e+00, -1.0634e+00, -1.5434e+00,\n",
            "           1.8981e-01,  1.1675e+00,  1.4533e+00,  2.0077e+00, -4.9679e-01,\n",
            "           5.2805e-01, -1.5257e+00,  1.0111e+00,  1.5536e+00,  1.9194e-01],\n",
            "         [ 1.8261e+00,  1.6004e+00, -1.3637e-02, -4.6084e-01,  2.0220e-01,\n",
            "          -1.2261e+00, -5.0217e-01, -1.2168e+00, -1.6732e-02,  2.1962e-02,\n",
            "          -1.1661e+00, -1.0337e-01,  9.7315e-01,  1.7144e+00, -4.7844e-01,\n",
            "          -1.1694e+00,  1.6766e+00, -5.4633e-02,  2.2189e+00, -1.6875e+00,\n",
            "           6.3722e-01, -1.6139e-01, -7.8070e-01, -2.0799e+00, -8.4798e-01],\n",
            "         [-4.9207e-01, -2.5964e-01, -9.2045e-01,  4.0285e-01,  1.5826e+00,\n",
            "          -2.7847e+00,  9.6908e-01,  9.4740e-01, -5.3607e-01, -1.0241e+00,\n",
            "          -1.1203e+00,  6.4362e-01, -5.3226e-01, -1.4336e+00,  5.8051e-01,\n",
            "          -2.6927e-01,  6.2265e-01, -1.0527e+00, -5.4905e-01,  5.9113e-01,\n",
            "           2.6697e+00, -2.2902e-01,  1.6436e+00,  1.9584e+00,  4.9096e-01],\n",
            "         [ 8.2307e-01,  2.4271e-01, -3.0463e-01,  1.1487e+00, -3.1039e-01,\n",
            "           3.1238e-01,  1.6171e+00,  2.3046e+00, -6.1989e-01,  1.0819e+00,\n",
            "          -1.9043e+00, -5.3404e-02, -1.1561e+00, -2.5769e-01, -1.3264e+00,\n",
            "           7.4616e-01, -8.7248e-01, -1.2062e-01,  1.0158e+00,  9.7910e-01,\n",
            "           7.7605e-01, -4.2759e-01, -1.6691e+00, -1.7492e-01,  1.0967e+00],\n",
            "         [ 1.2509e+00,  2.1547e-01,  1.0871e+00, -1.1452e-01, -4.3526e-01,\n",
            "           7.2519e-02,  2.1549e+00, -3.0489e-01,  4.2262e-01, -6.5194e-02,\n",
            "          -8.5193e-02, -2.8576e+00,  9.9998e-01, -1.2566e+00,  1.2120e-01,\n",
            "          -4.0501e-01, -8.8228e-01,  7.5013e-01,  1.8071e-01, -1.1994e-01,\n",
            "           7.9279e-01, -1.1489e+00, -1.4023e+00, -1.0479e+00,  6.6752e-01],\n",
            "         [ 1.0916e+00,  1.2723e+00, -3.1899e-01,  5.1394e-02,  5.6543e-01,\n",
            "          -4.1473e-01, -5.8257e-01, -8.9858e-01, -1.7270e+00,  7.3894e-03,\n",
            "          -1.6786e+00,  9.7760e-01,  2.1962e-01,  2.3215e-01,  1.4350e+00,\n",
            "           8.3339e-01, -5.0437e-01, -8.2679e-01, -7.7047e-01, -5.3345e-01,\n",
            "           1.0011e+00, -1.3692e+00,  5.1273e-01,  1.8689e+00,  2.8352e+00],\n",
            "         [ 6.0487e-01,  7.9231e-01, -5.1677e-02,  3.1782e-01, -5.3520e-01,\n",
            "          -6.3881e-01, -4.2052e-01, -7.0643e-01, -3.2670e-01,  5.4047e-01,\n",
            "           2.4255e-01,  1.1635e+00,  2.0127e-01, -2.2712e-02, -3.3863e-01,\n",
            "          -4.0780e-01,  1.6159e+00, -1.1580e+00, -2.2230e+00, -1.4496e+00,\n",
            "           2.4433e-01,  1.5259e+00,  1.4611e-01,  9.8108e-01, -3.0175e-01],\n",
            "         [ 1.4950e-01,  1.1893e+00,  1.2066e-01, -1.2049e+00, -4.1031e-01,\n",
            "           9.0788e-01,  1.7368e+00,  1.0016e+00, -5.8667e-01, -3.0554e-01,\n",
            "          -3.6730e-01, -1.3380e+00,  9.0276e-01,  5.9423e-01, -2.5445e-01,\n",
            "          -4.4934e-01, -2.9734e-01, -1.9577e-01,  9.0008e-01,  1.1202e+00,\n",
            "          -6.3600e-02,  1.0452e-01, -2.5229e+00,  4.8245e-01,  3.0121e-01],\n",
            "         [ 9.3186e-01,  1.1088e+00,  7.6828e-01,  8.7398e-01, -1.5214e-01,\n",
            "           4.8488e-01,  5.7004e-01,  1.2185e-01, -2.5573e-01,  2.6094e-01,\n",
            "           9.1575e-01,  4.1242e-01, -6.8965e-01, -1.8958e+00,  1.5298e-01,\n",
            "          -2.1604e+00, -3.4874e-01, -4.2997e-01,  3.3558e-02,  2.8867e-01,\n",
            "           8.5726e-02, -8.8889e-02,  3.4323e-02, -1.4168e+00, -1.4430e+00]]]) scores tensor([[[0.4423, 0.5577],\n",
            "         [0.4708, 0.5292],\n",
            "         [0.4706, 0.5294],\n",
            "         [0.4735, 0.5265],\n",
            "         [0.4952, 0.5048],\n",
            "         [0.4616, 0.5384],\n",
            "         [0.4553, 0.5447],\n",
            "         [0.5269, 0.4731],\n",
            "         [0.4683, 0.5317],\n",
            "         [0.4709, 0.5291],\n",
            "         [0.4838, 0.5162],\n",
            "         [0.4734, 0.5266],\n",
            "         [0.4717, 0.5283],\n",
            "         [0.4501, 0.5499],\n",
            "         [0.4480, 0.5520],\n",
            "         [0.4487, 0.5513],\n",
            "         [0.4367, 0.5633],\n",
            "         [0.4637, 0.5363]]]) \n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "EyWuiPDVd33I",
        "colab_type": "text"
      },
      "source": [
        "As one may expect, without training the network does give much better than a fair coin.\n",
        "We can now train the model, using a stochastic gradient descent."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "AlrqjgmEeIMx",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "loss_function = nn.NLLLoss()\n",
        "optimizer = torch.optim.SGD(model.parameters(), lr=0.08) \n",
        "model"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "MndhR_S4jyK5",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "18eebaca-3c03-42fa-fe81-93044328e018"
      },
      "source": [
        "\n",
        "def prepare_sequence(seq, to_ix):\n",
        "    idxs = [to_ix[w] if w in to_ix else len(to_ix) for w in seq]\n",
        "    idxs = torch.tensor(idxs, dtype=torch.long).unsqueeze(0)\n",
        "    return idxs\n",
        "\n",
        "prepare_sequence([1],tag_to_ix)"
      ],
      "execution_count": 55,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "tensor([[0]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 55
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "u0_dLl0bWdqF",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 405
        },
        "outputId": "44faeb5b-392c-4469-aba7-ac6ebe25d33b"
      },
      "source": [
        "\n",
        "\n",
        "from torch.nn.functional import one_hot\n",
        "n_epochs = 10\n",
        "tag_to_ix = {0.0: 0, 1.0: 1}\n",
        "for epoch in range(n_epochs): \n",
        "    loss_mean_ep = 0\n",
        "    n_sample = 0\n",
        "    t = time()\n",
        "    for sentence, tags in training_data:\n",
        "        if len(sentence) < 2: continue\n",
        "        # Step 1. Remember that Pytorch accumulates gradients.\n",
        "        # We need to clear them out before each instance\n",
        "        model.zero_grad()\n",
        "        # Step 2. Get our inputs ready for the network, that is, turn them into\n",
        "        # Tensors of word indices.\n",
        "        sentence_in = prepare_embedding(sentence,embedding=model_wv)\n",
        "        #import pdb\n",
        "        #pdb.set_trace()\n",
        "        targets = prepare_sequence([tags], tag_to_ix)\n",
        "        one_hot_targets = one_hot(targets.squeeze(0), num_classes=len(tag_to_ix))\n",
        "        # Step 3. Run our forward pass.\n",
        "        tag_scores = model(sentence_in)\n",
        "        # Step 4. Compute the loss, gradients, and update the parameters by\n",
        "        #  calling optimizer.step()\n",
        "        #import pdb\n",
        "        #pdb.set_trace()\n",
        "        # print(tag_scores.shape, one_hot_targets.shape)\n",
        "        loss = loss_function(tag_scores, one_hot_targets)\n",
        "        loss_mean_ep += loss\n",
        "        n_sample += 1\n",
        "        loss.backward()\n",
        "        optimizer.step()\n",
        "    print(\"Epoch {} loss {:0.4f} \".format(epoch, loss_mean_ep/n_sample),\"Taking {} secs\".format(time()-t))"
      ],
      "execution_count": 56,
      "outputs": [
        {
          "output_type": "error",
          "ename": "KeyboardInterrupt",
          "evalue": "ignored",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-56-b68929940a3e>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m     15\u001b[0m         \u001b[0;31m# Step 2. Get our inputs ready for the network, that is, turn them into\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     16\u001b[0m         \u001b[0;31m# Tensors of word indices.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 17\u001b[0;31m         \u001b[0msentence_in\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mprepare_embedding\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0msentence\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0membedding\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mmodel_wv\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     18\u001b[0m         \u001b[0;31m#import pdb\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     19\u001b[0m         \u001b[0;31m#pdb.set_trace()\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m<ipython-input-42-0533ca6d7a87>\u001b[0m in \u001b[0;36mprepare_embedding\u001b[0;34m(seq, embedding)\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mprepare_embedding\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mseq\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0membedding\u001b[0m \u001b[0;34m=\u001b[0m\u001b[0mmodel_wv\u001b[0m \u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 2\u001b[0;31m     \u001b[0midxs\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0membedding\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mw\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mw\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mcorpus\u001b[0m \u001b[0;32melse\u001b[0m \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mrandom\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mnormal\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0msize\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0md\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mw\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mseq\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      3\u001b[0m     \u001b[0;31m# note : Word2Vec can't handle out-of-vocabulary, thus we resort to a random vector for out-vocabulary words\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      4\u001b[0m     \u001b[0midxs\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mtorch\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtensor\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0midxs\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mdtype\u001b[0m \u001b[0;34m=\u001b[0m\u001b[0mtorch\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfloat\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0munsqueeze\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      5\u001b[0m     \u001b[0;32mreturn\u001b[0m \u001b[0midxs\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m<ipython-input-42-0533ca6d7a87>\u001b[0m in \u001b[0;36m<listcomp>\u001b[0;34m(.0)\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mprepare_embedding\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mseq\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0membedding\u001b[0m \u001b[0;34m=\u001b[0m\u001b[0mmodel_wv\u001b[0m \u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 2\u001b[0;31m     \u001b[0midxs\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0membedding\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mw\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mw\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mcorpus\u001b[0m \u001b[0;32melse\u001b[0m \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mrandom\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mnormal\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0msize\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0md\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mw\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mseq\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      3\u001b[0m     \u001b[0;31m# note : Word2Vec can't handle out-of-vocabulary, thus we resort to a random vector for out-vocabulary words\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      4\u001b[0m     \u001b[0midxs\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mtorch\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtensor\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0midxs\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mdtype\u001b[0m \u001b[0;34m=\u001b[0m\u001b[0mtorch\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfloat\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0munsqueeze\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      5\u001b[0m     \u001b[0;32mreturn\u001b[0m \u001b[0midxs\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PhFbyZalX8TE",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "torch.save(model.state_dict(),'.\\lstm3.pt')"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "kLBU9iN0X8rI",
        "colab_type": "text"
      },
      "source": [
        "Qualitative evaluation of the model:\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ritJmm_8X1Ir",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "with torch.no_grad():\n",
        "  label_count = [0, 0]\n",
        "  n_samples = 5\n",
        "  for sentence, tag in test_data:\n",
        "          label_count[int(tag)] += 1\n",
        "          if label_count[int(tag)] <= n_samples:\n",
        "            sentence_in = prepare_embedding(sentence)\n",
        "            sentence_tag = 'Positive' if tag==1 else 'Negative'\n",
        "            tag_scores = model(sentence_in)\n",
        "            positive_score = torch.mean(torch.exp(tag_scores.squeeze(0)[:, 1])).item()\n",
        "            print(\"{} sentence scored {} \\\"{}\\\".\"\n",
        "            \"\".format(sentence_tag, positive_score,' '.join(sentence)) )\n",
        "          if (label_count[0] >= n_samples) and (label_count[1] >= n_samples):\n",
        "            break"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "dvivPSJjYMAg",
        "colab_type": "text"
      },
      "source": [
        "Quantitative evaluation of the model:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "v3uqVSNWYF57",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.metrics import roc_auc_score, classification_report\n",
        "\n",
        "ground_truth, scores = [], []\n",
        "with torch.no_grad():\n",
        "  for sentence, tag in test_data:\n",
        "    sentence_in = prepare_embedding(sentence)\n",
        "    ground_truth.append(tag)\n",
        "    tag_scores = model(sentence_in)\n",
        "    scores.append(torch.mean(torch.exp(tag_scores.squeeze(0)[:, 1])).item())\n",
        "\n",
        "print(\"The AUC is \", roc_auc_score(ground_truth, scores))\n",
        "print(\"Classification report:\\n\", classification_report(ground_truth, np.array(scores)>0.5))"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ioFe8eGHxYqc",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "C = confusion_matrix(ground_truth,np.array(scores)>0.5)\n",
        "neg_pred_perf = C[0,0]/(C[0,0]+C[1,0])\n",
        "fals_disc_rate = C[1,1]/(C[1,1]+C[1,0])\n",
        "neg_pred_perf,fals_disc_rate"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "F0r3O8VWF2rr",
        "colab_type": "text"
      },
      "source": [
        "## 3. Classification decision trees with Bag-of-Words features\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "We now use impelment a Bag of Words approach. We first tokenize each review and we also use the tokenized whole corpus to define a dictionnary."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "VfMrrHpCU0A0",
        "colab_type": "code",
        "outputId": "888c0d3a-ad22-42bb-a18b-07eea9eefbe5",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 203
        }
      },
      "source": [
        "\n",
        "#we begin by lowering the review, removing the line breaks and numbers\n",
        "df3 = df2.text.apply(lambda x: re.sub(r'[0-9]+', '', x.lower().rstrip())) \n",
        "\n",
        "x = df3.to_numpy()\n",
        "df2.head()"
      ],
      "execution_count": 292,
      "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>text</th>\n",
              "      <th>label</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>Super simple place but amazing nonetheless. It...</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>Small unassuming place that changes their menu...</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>Lester's is located in a beautiful neighborhoo...</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>Love coming here. Yes the place always needs t...</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>Had their chocolate almond croissant and it wa...</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                                                text  label\n",
              "0  Super simple place but amazing nonetheless. It...    1.0\n",
              "1  Small unassuming place that changes their menu...    1.0\n",
              "2  Lester's is located in a beautiful neighborhoo...    1.0\n",
              "3  Love coming here. Yes the place always needs t...    1.0\n",
              "4  Had their chocolate almond croissant and it wa...    1.0"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 292
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "wQlfEwLjDX7g",
        "colab_type": "text"
      },
      "source": [
        "We then build the Bag-of-words features of the text using a feature extractor from Scikit-learn."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zZlRups1W8CD",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.feature_extraction.text import CountVectorizer\n",
        "nb_max_feat = 5000\n",
        "vectorizer = CountVectorizer(analyzer = \"word\", tokenizer = None, preprocessor = None, stop_words = 'english', max_features = nb_max_feat) \n",
        "train_data_features = vectorizer.fit_transform(x)\n",
        "z =vectorizer.transform([df2.text.array[6]]).toarray()\n",
        "nb_feat = z.shape[0]"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "0B8Cs4gT_b9v",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "X = train_data_features.toarray()\n",
        "X_train,X_test,X_validation= X[:n_train],X[n_train:n_train_test],X[n_train_test:]\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "FwK7-KEPWrVQ",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.tree import DecisionTreeClassifier\n",
        "from sklearn.metrics import confusion_matrix\n",
        "\n",
        "clf = DecisionTreeClassifier(max_depth=5,random_state= 0)\n",
        "\n",
        "clf = clf.fit(X_train,Y_train)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "G4m_xq6SOfHA",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "Y_hat_test2 = clf.predict(X_test)\n",
        "C = confusion_matrix(Y_test,Y_hat_test2)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3d5RtRo2ZU-R",
        "colab_type": "code",
        "outputId": "64910ae6-9be2-4fbf-f9e8-11fec2f05e01",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "neg_pred_perf = C[0,0]/(C[0,0]+C[1,0])\n",
        "fals_disc_rate = C[1,1]/(C[1,1]+C[1,0])\n",
        "fpr = C[0,1]/(C[0,1]+C[1,1])\n",
        "tpr = C[ 1,1]/(C[1,1]+C[0,1])\n",
        "neg_pred_perf,fals_disc_rate"
      ],
      "execution_count": 317,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(0.759493670886076, 0.9732142857142857)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 317
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "5Ycarl1NEicF",
        "colab_type": "text"
      },
      "source": [
        "We see that naively training a decision tree with a maximum depth of 5, we already have pretty good results according to our metrics.\n",
        "Since training the decision trees is not too costly, we can proceed to a model selection on the depth of the tree using the test set."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vHn7O2E2DEcH",
        "colab_type": "code",
        "outputId": "1f326dcd-9fb4-4e9a-d103-b70fca69b0e6",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 567
        }
      },
      "source": [
        "range_depth = np.arange(2,14,2)\n",
        "neg_pred_ = []\n",
        "false_disc_ = []\n",
        "fpr_ = []\n",
        "tpr_ = []\n",
        "T = np.linspace(0,1,100)\n",
        "plt.figure(figsize=( 13,9))\n",
        "plt.subplot(1,2,1)\n",
        "plt.plot(T,T,'b--')\n",
        "for d in range_depth:\n",
        "  clf = DecisionTreeClassifier(max_depth=d,random_state=0)\n",
        "  clf = clf.fit(X_train,Y_train)\n",
        "  Y_hat_test = clf.predict(X_test)\n",
        "  C = confusion_matrix(Y_test,Y_hat_test)\n",
        "  neg_pred_perf = C[0,0]/(C[0,0]+C[1,0])\n",
        "  fals_disc_rate = C[1,1]/(C[1,1]+C[1,0])\n",
        "  fpr = C[0,1]/(C[0,1]+C[1,1])\n",
        "  tpr = C[ 1,1]/(C[1,1]+C[0,1])\n",
        "  neg_pred_.append(neg_pred_perf)\n",
        "  false_disc_.append(fals_disc_rate)\n",
        "  fpr_.append(fpr)\n",
        "  tpr_.append(tpr)\n",
        "  plt.plot(fpr,tpr,'o',label = \"Depth of tree = {}\".format(d))\n",
        "\n",
        "plt.legend()\n",
        "plt.title(\"ROC for decision trees with different depths\")\n",
        "plt.savefig(\"ROC_CART\")\n",
        "plt.subplot(1,2,2)\n",
        "plt.xlabel(\"Negative predictive performance\")\n",
        "plt.ylabel(\"False discovery rate\")\n",
        "for i in range(len(range_depth)):\n",
        "  d=range_depth[i]\n",
        "  plt.plot(neg_pred_[i],false_disc_[i],'o',label = \"Depth of tree = {}\".format(d))\n",
        "\n",
        "plt.legend()\n",
        "plt.title(\"Ah-doc metrics performances for decision trees with different depths\")\n",
        "plt.savefig(\"adhocCART\")\n",
        "\n",
        "\n",
        "\n",
        "\n"
      ],
      "execution_count": 318,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 936x648 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "1u75jSID58Jj",
        "colab_type": "text"
      },
      "source": [
        "We see that overall the decision trees have pretty good perfomances. We see that the ROC improves gradually as depth is increased. On the others hand, for the metrics we chose to focus on, we see that the perfomances seems to decrease for larger models. Given the perfomances, we choose to use a depth of *4* for the validation. \\\\\n",
        "We next plot the decision tree obtained for $depth = 4$."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "hQuCbbrjskvd",
        "colab_type": "code",
        "outputId": "fdbdf886-8293-4a51-846d-81c7f529991e",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 969
        }
      },
      "source": [
        "\n",
        "clf = DecisionTreeClassifier(max_depth=4,random_state=0)\n",
        "plt.figure(figsize=(15, 15))\n",
        "plot_tree(clf.fit(X_train,Y_train),max_depth=1) "
      ],
      "execution_count": 326,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[Text(418.5, 679.5, 'X[4942] <= 0.5\\ngini = 0.293\\nsamples = 10800\\nvalue = [1923, 8877]'),\n",
              " Text(209.25, 407.69999999999993, 'X[218] <= 0.5\\ngini = 0.276\\nsamples = 10603\\nvalue = [1751, 8852]'),\n",
              " Text(104.625, 135.89999999999998, '\\n  (...)  \\n'),\n",
              " Text(313.875, 135.89999999999998, '\\n  (...)  \\n'),\n",
              " Text(627.75, 407.69999999999993, 'X[2592] <= 1.5\\ngini = 0.222\\nsamples = 197\\nvalue = [172, 25]'),\n",
              " Text(523.125, 135.89999999999998, '\\n  (...)  \\n'),\n",
              " Text(732.375, 135.89999999999998, '\\n  (...)  \\n')]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 326
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1080x1080 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "nIRLWRbj9Yx9",
        "colab_type": "text"
      },
      "source": [
        "To interprete the tree plot we determine the word associated with each of the 3 first nodes. "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "NwrQpDA5IHYh",
        "colab_type": "code",
        "outputId": "b29ab48e-3c8b-4722-bd83-28f0dfb05b43",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "X = np.zeros(5000)\n",
        "X[4942] = 1\n",
        "vectorizer.inverse_transform(X)  # we use the inverse transform of the Bag-of-words feature extractor above"
      ],
      "execution_count": 323,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[array(['worst'], dtype='<U19')]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 323
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "IFnmJYgJQi5G",
        "colab_type": "code",
        "outputId": "87b4cf81-2a28-4f7d-fea6-05326afc730d",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "X = np.zeros(5000)\n",
        "X[218] = 1\n",
        "vectorizer.inverse_transform(X)"
      ],
      "execution_count": 324,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[array(['asked'], dtype='<U19')]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 324
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "eTqJoZtsQqDW",
        "colab_type": "code",
        "outputId": "06022ebf-b07c-4405-a8c1-d7c5f252054b",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "X = np.zeros(5000)\n",
        "X[2592] = 1\n",
        "vectorizer.inverse_transform(X)"
      ],
      "execution_count": 325,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[array(['love'], dtype='<U19')]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 325
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "dpY7yGaJZJi9",
        "colab_type": "text"
      },
      "source": [
        "Here we see that the first words selected by the model are \"worsted\", \"asked\" and \"love\", which makes sense qualitatively. \\\\\n",
        "We next use test samples to verify qualitatively the performance of the DT.\n",
        "Finally we compute the performance of the selected tree on the validation set."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "imxUEqTPRA7X",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "Y_hat_validation = clf.predict(X_validation)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "IK7CZK1AsdOU",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "d6a8f903-bcc3-4656-9bd0-0f6f43629922"
      },
      "source": [
        "\n",
        "C = confusion_matrix(Y_validation,Y_hat_validation)\n",
        "neg_pred_perf = C[0,0]/(C[0,0]+C[1,0])\n",
        "fals_disc_rate = C[1,1]/(C[1,1]+C[1,0])\n",
        "fpr = C[0,1]/(C[0,1]+C[1,1])\n",
        "tpr = C[ 1,1]/(C[1,1]+C[0,1])\n",
        "neg_pred_perf,fals_disc_rate"
      ],
      "execution_count": 333,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(0.625, 0.9790372670807453)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 333
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "cGSTWfalNY4q",
        "colab_type": "text"
      },
      "source": [
        "## 4 Training a LSTM nn with it's own embedding layer"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "VUJkMmNjVPG7",
        "colab_type": "text"
      },
      "source": [
        "We initally wanted to train a bidirectional LSTM but we had trouble with training it properly. We at least get a good comparison between an LSTM with a pre-trained Word2vec layer, and an LSTM with it's own embedding layer.\n",
        "\n",
        "To keep things fare with the LSTM above, we give the same dimensions for it's embedding and hidden layers."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "WoGmmE-ysY54",
        "colab_type": "text"
      },
      "source": [
        "### Building the features for the model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PnyDIwa9BN2k",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from nltk.corpus import wordnet\n",
        "\n",
        "def get_wordnet_pos(pos_tag):\n",
        "    if pos_tag.startswith('J'):\n",
        "        return wordnet.ADJ\n",
        "    elif pos_tag.startswith('V'):\n",
        "        return wordnet.VERB\n",
        "    elif pos_tag.startswith('N'):\n",
        "        return wordnet.NOUN\n",
        "    elif pos_tag.startswith('R'):\n",
        "        return wordnet.ADV\n",
        "    else:\n",
        "        return wordnet.NOUN\n",
        "\n",
        "def clean_tokenized_text(text):\n",
        "    #Ad-hoc cleaning and tokenizing of the text\n",
        "    # lower text\n",
        "    text = text.lower()\n",
        "    # remove line jumps and tabs\n",
        "    text = text.replace(\"\\n\", '').replace('\\r', '').replace('\\t','').replace(\"\\\\/\",'-').replace(\"!\",' ').replace(\".\",' ').replace(\",\",\" \")\n",
        "    # tokenize text and remove puncutation\n",
        "    text = [word.strip(string.punctuation) for word in text.split(\" \")]           # The character \"'\" is not considered as punctation\n",
        "    # remove words that contain numbers\n",
        "    text = [word for word in text if not any(c.isdigit() for c in word)]\n",
        "    # remove stop words\n",
        "    stop = stopwords.words('english')\n",
        "    text = [x for x in text if x not in stop]\n",
        "    # remove empty tokens\n",
        "    text = [t for t in text if len(t) > 0]      \n",
        "    # pos tag text\n",
        "    pos_tags = pos_tag(text)\n",
        "    # lemmatize text\n",
        "    text = [WordNetLemmatizer().lemmatize(t[0], get_wordnet_pos(t[1])) for t in pos_tags]\n",
        "    # remove words with only one letter\n",
        "    text = [t for t in text if len(t) > 1]   # we are not interested in one-character expressions such as 'a' or '&'\n",
        "    # join all\n",
        "    return(text)   \n",
        "\n",
        "\n",
        "\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "2g2VDsAaFDap",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "df2.text=df[\"text\"].apply(lambda x: clean_tokenized_text(x))\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "2f3khej8iJJZ",
        "colab_type": "code",
        "outputId": "82286576-ae71-4b86-bf4e-694fa68c4de4",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "\"eerj j \\n stuff\\n\".replace(\"\\n\",'')"
      ],
      "execution_count": 25,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "'eerj j  stuff'"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 25
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "DTezzIz_P-4o",
        "colab_type": "code",
        "outputId": "39bfdab2-ff12-4ee9-dd9f-ddc6a0f4c828",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 210
        }
      },
      "source": [
        "df2.text.head(10)"
      ],
      "execution_count": 26,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0    [super, simple, place, amaze, nonetheless, aro...\n",
              "1    [small, unassuming, place, change, menu, every...\n",
              "2    [lester's, locate, beautiful, neighborhood, si...\n",
              "3    [love, come, yes, place, always, need, floor, ...\n",
              "4    [chocolate, almond, croissant, amazing, light,...\n",
              "5    [cycle, pub, la, vega, blast, get, groupon, re...\n",
              "6    [would, guess, would, able, get, fairly, decen...\n",
              "7    [always, drive, past, coffee, house, wonder, b...\n",
              "8    [bad, love, gluten-free, vegan, version, chees...\n",
              "9    [love, place, peggy, great, dog, great, job, p...\n",
              "Name: text, dtype: object"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 26
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qnp4yIqbNnax",
        "colab_type": "code",
        "outputId": "12e1b1e4-10f6-47c3-af82-e388c493d36e",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "x = df2.text.to_numpy()\n",
        "data = []\n",
        "for i in range(len(x)):\n",
        "    data.append((x[i], Y[i]))\n",
        "training_data = data[:n_train]\n",
        "test_data = data[n_train:n_train_test]\n",
        "validation_data = data[n_train_test:]\n",
        "print(\"Got {} training samples, {} test samples, {} validation samples\".format(len(training_data), len(test_data),len(validation_data)))\n"
      ],
      "execution_count": 27,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Got 10800 training samples, 2700 test samples, 1500 validation samples\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vvKV8-1xfiea",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def prepare_sequence(seq, to_ix):\n",
        "    idxs = [to_ix[w] if w in to_ix else len(to_ix) for w in seq]\n",
        "    idxs = torch.tensor(idxs, dtype=torch.long).unsqueeze(0)\n",
        "    return idxs\n",
        "\n",
        "word_to_ix = {}\n",
        "tag_to_ix = {}\n",
        "for review, tag in training_data:\n",
        "    if tag not in tag_to_ix:\n",
        "        tag_to_ix[tag] = len(tag_to_ix)\n",
        "    for word in review:\n",
        "        if word not in word_to_ix:\n",
        "            word_to_ix[word] = len(word_to_ix)\n",
        "\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_kVQBQvngC0H",
        "colab_type": "code",
        "outputId": "dbe07f7d-e209-420f-9b42-9cdd6ee03b61",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        }
      },
      "source": [
        "training_data[6]"
      ],
      "execution_count": 29,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(['would',\n",
              "  'guess',\n",
              "  'would',\n",
              "  'able',\n",
              "  'get',\n",
              "  'fairly',\n",
              "  'decent',\n",
              "  'vietnamese',\n",
              "  'restaurant',\n",
              "  'east',\n",
              "  'york',\n",
              "  'quite',\n",
              "  'chinatown',\n",
              "  'term',\n",
              "  'price',\n",
              "  'slightly',\n",
              "  'higher',\n",
              "  'definitely',\n",
              "  'one',\n",
              "  'good',\n",
              "  'vietnamese',\n",
              "  'restaurant',\n",
              "  'outside',\n",
              "  'neighbourhood',\n",
              "  'time',\n",
              "  'go',\n",
              "  'chinatown',\n",
              "  'next',\n",
              "  'best',\n",
              "  'thing',\n",
              "  'street',\n",
              "  'far',\n",
              "  'item',\n",
              "  'try',\n",
              "  'phos',\n",
              "  'beef',\n",
              "  'chicken',\n",
              "  'vegetarian',\n",
              "  'disappoint',\n",
              "  'especially',\n",
              "  'chicken',\n",
              "  'pho',\n",
              "  'next',\n",
              "  'time',\n",
              "  'go',\n",
              "  'back',\n",
              "  \"i'm\",\n",
              "  'go',\n",
              "  'try',\n",
              "  'banh',\n",
              "  'cuon',\n",
              "  'steam',\n",
              "  'rice',\n",
              "  'noodle',\n",
              "  'vermicelli'],\n",
              " 1.0)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 29
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PSK6wAj4mbXU",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# defining the classifier model\n",
        "\n",
        "class SequenceLabeller(nn.Module):\n",
        "\n",
        "    def __init__(self, embedding_dim, hidden_dim, vocab_size, num_classes, sequence_model=\"LSTM\"):\n",
        "        super(SequenceLabeller, self).__init__()\n",
        "\n",
        "        self.hidden_dim = hidden_dim\n",
        "        self.word_embeddings = nn.Embedding(vocab_size, embedding_dim)\n",
        "        \n",
        "        if sequence_model == \"LSTM\":\n",
        "            # The LSTM takes word embeddings as inputs, and outputs hidden states\n",
        "            # with dimensionality hidden_dim.\n",
        "            self.seq = nn.LSTM(embedding_dim, hidden_dim,batch_first=True)\n",
        "        else:\n",
        "            raise(Exception(\"Sequence model {} not supported\".format(sequence_model)))\n",
        "\n",
        "        # The linear layer that maps from hidden state space to class space\n",
        "        self.hidden2tag = nn.Linear(hidden_dim, num_classes)\n",
        "\n",
        "    def forward(self, sentence):\n",
        "        embeds = self.word_embeddings(sentence)\n",
        "        seq_output, (h_n,c_n) = self.seq(embeds)\n",
        "        tag_space = self.hidden2tag(seq_output)\n",
        "        # Although we will be performing binary classification, this is a \n",
        "        # general implementation displaying what should be done for n_classes>2\n",
        "        tag_scores = F.log_softmax(tag_space, dim=2) # Computationally efficient for the criterion\n",
        "        return tag_scores"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "shbethv7vlEk",
        "colab_type": "text"
      },
      "source": [
        "We run once the model on the first review."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "VA9wLOLbjNaE",
        "colab_type": "code",
        "outputId": "7c7f8787-6b12-4ae2-976e-6d06a5f8a0cf",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 354
        }
      },
      "source": [
        "# the additional vocabulary element is to account for unkown words\n",
        "model = SequenceLabeller(25, 30, len(word_to_ix)+1, len(tag_to_ix)) \n",
        "inputs = prepare_sequence(training_data[0][0], word_to_ix)\n",
        "with torch.no_grad():\n",
        "    sent_scores = model(torch.cat([inputs]))\n",
        "    print(\"Input tokens {} scores {} \".format(inputs, np.exp(sent_scores)))"
      ],
      "execution_count": 33,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Input tokens tensor([[ 0,  1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11, 12, 13, 14, 15, 16, 17]]) scores tensor([[[0.4584, 0.5416],\n",
            "         [0.4502, 0.5498],\n",
            "         [0.4591, 0.5409],\n",
            "         [0.4378, 0.5622],\n",
            "         [0.4440, 0.5560],\n",
            "         [0.4920, 0.5080],\n",
            "         [0.5070, 0.4930],\n",
            "         [0.4825, 0.5175],\n",
            "         [0.4824, 0.5176],\n",
            "         [0.4917, 0.5083],\n",
            "         [0.4924, 0.5076],\n",
            "         [0.4636, 0.5364],\n",
            "         [0.4497, 0.5503],\n",
            "         [0.4342, 0.5658],\n",
            "         [0.4372, 0.5628],\n",
            "         [0.4608, 0.5392],\n",
            "         [0.4472, 0.5528],\n",
            "         [0.4678, 0.5322]]]) \n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "aP5Q1rNzjWVB",
        "colab_type": "code",
        "outputId": "b3e5536a-8af1-4bea-90ac-fd83d7ded32d",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 105
        }
      },
      "source": [
        "model"
      ],
      "execution_count": 34,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "SequenceLabeller(\n",
              "  (word_embeddings): Embedding(32528, 25)\n",
              "  (seq): LSTM(25, 30, batch_first=True)\n",
              "  (hidden2tag): Linear(in_features=30, out_features=2, bias=True)\n",
              ")"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 34
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "HMycGDlIjPVp",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "loss_function = nn.NLLLoss()\n",
        "optimizer = torch.optim.SGD(model.parameters(), lr=0.01) "
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "toKCnq0-uLBz",
        "colab_type": "text"
      },
      "source": [
        "### 4.3 Training of the model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "dkeefLZXuI1o",
        "colab_type": "code",
        "outputId": "2417ecf4-7c04-48b1-9054-ef5bd591033d",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 369
        }
      },
      "source": [
        "from torch.nn.functional import one_hot\n",
        "n_epochs = 20\n",
        "for epoch in range(n_epochs):  # again, normally you would NOT do 300 epochs, it is toy data\n",
        "    loss_mean_ep = 0\n",
        "    n_sample = 0\n",
        "    for sentence, tags in training_data:\n",
        "        if len(sentence) < 2: continue\n",
        "        # Step 1. Remember that Pytorch accumulates gradients.\n",
        "        # We need to clear them out before each instance\n",
        "        model.zero_grad()\n",
        "        # Step 2. Get our inputs ready for the network, that is, turn them into\n",
        "        # Tensors of word indices.\n",
        "        sentence_in = prepare_sequence(sentence, word_to_ix)\n",
        "        #import pdb\n",
        "        #pdb.set_trace()\n",
        "        targets = prepare_sequence([tags], tag_to_ix)\n",
        "        one_hot_targets = one_hot(targets.squeeze(0), num_classes=len(tag_to_ix))\n",
        "        # Step 3. Run our forward pass.\n",
        "        tag_scores = model(sentence_in)\n",
        "        # Step 4. Compute the loss, gradients, and update the parameters by\n",
        "        #  calling optimizer.step()\n",
        "        #import pdb\n",
        "        #pdb.set_trace()\n",
        "        # print(tag_scores.shape, one_hot_targets.shape)\n",
        "        loss = loss_function(tag_scores, one_hot_targets)\n",
        "        loss_mean_ep += loss\n",
        "        n_sample += 1\n",
        "        loss.backward()\n",
        "        optimizer.step()\n",
        "    print(\"Epoch {} loss {:0.4f} \".format(epoch, loss_mean_ep/n_sample))"
      ],
      "execution_count": 37,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 0 loss 0.5832 \n",
            "Epoch 1 loss 0.4684 \n",
            "Epoch 2 loss 0.4611 \n",
            "Epoch 3 loss 0.4570 \n",
            "Epoch 4 loss 0.4534 \n",
            "Epoch 5 loss 0.4500 \n",
            "Epoch 6 loss 0.4468 \n",
            "Epoch 7 loss 0.4439 \n",
            "Epoch 8 loss 0.4410 \n",
            "Epoch 9 loss 0.4382 \n",
            "Epoch 10 loss 0.4354 \n",
            "Epoch 11 loss 0.4327 \n",
            "Epoch 12 loss 0.4299 \n",
            "Epoch 13 loss 0.4272 \n",
            "Epoch 14 loss 0.4245 \n",
            "Epoch 15 loss 0.4218 \n",
            "Epoch 16 loss 0.4190 \n",
            "Epoch 17 loss 0.4162 \n",
            "Epoch 18 loss 0.4134 \n",
            "Epoch 19 loss 0.4105 \n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "81tv3IxW9aUc",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "torch.save(model.state_dict(),'.\\lstm2')"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ekV7zPQ1bWiM",
        "colab_type": "text"
      },
      "source": [
        "Qualititative evaluation of the model:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "HZ3GZ6Lfa9NW",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 213
        },
        "outputId": "42a4d905-3117-42a5-e132-bc129a1f54e6"
      },
      "source": [
        "with torch.no_grad():\n",
        "  label_count = [0, 0]\n",
        "  n_samples = 5\n",
        "  for sentence, tag in test_data:\n",
        "          label_count[int(tag)] += 1\n",
        "          if label_count[int(tag)] <= n_samples:\n",
        "            sentence_in = prepare_sequence(sentence, word_to_ix)\n",
        "            sentence_tag = 'Positive' if tag==1 else 'Negative'\n",
        "            tag_scores = model(sentence_in)\n",
        "            positive_score = torch.mean(torch.exp(tag_scores.squeeze(0)[:, 1])).item()\n",
        "            print(\"{} sentence scored {} \\\"{}\\\".\"\n",
        "            \"\".format(sentence_tag, positive_score,' '.join(sentence)) )\n",
        "          if (label_count[0] >= n_samples) and (label_count[1] >= n_samples):\n",
        "            break"
      ],
      "execution_count": 39,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Negative sentence scored 0.024647405371069908 \"call criss angel masturbate live show stroke ego good time spend watch mashups old video see youtube elsewhere criss appear put effort whatsoever performance literally say show get pay whether applauded trick seem like magic show part rest make worth spent best part show fifi maestro even become tiresome vulgur awhile tell afterwards show close april renew tell criss do bad fan continue spend money disappoint\".\n",
            "Positive sentence scored 0.03982580080628395 \"great company they're reliable honest owner man word move bed damage however owner promptly take care problem cost would definitely recommend company others guarantee service\".\n",
            "Negative sentence scored 0.02169165574014187 \"see movie split people theater minute leave movie right major climax light turn never saw end movie people theater walk around minute look manager place completely empty manager find fiancé call manager helpful plan take way top since manager call fiancé liar tell true also refund regular price seat vip seat purchase sound like bad management one floor manage confusion\".\n",
            "Positive sentence scored 0.03761202469468117 \"love place remind trip italy come home husband buy expobar house love make coffee however love experience go traditional italian coffee shop stand around table enjoy daily cappuccino\".\n",
            "Negative sentence scored 0.026872003450989723 \"ask membership close since may continue charge card month even though repetitively email call email call contact hq inform cannot help dollar charge save time energy go anywhere contact pnc fraud local police station report theft\".\n",
            "Positive sentence scored 0.08556883782148361 \"best music great staff staple north shore baseball-football game around fun atmosphere\".\n",
            "Positive sentence scored 0.060871656984090805 \"like think charge miss class stuck traffic little rough overall enjoy workout good get back work\".\n",
            "Positive sentence scored 0.03726540878415108 \"costumer service lack time high food presentation good put together well customer issue order one seem care help-the person make food roll eye old lady ask spoon soup customer service poor can't expect service smile place food still good\".\n",
            "Negative sentence scored 0.11171727627515793 \"mean actual place hot mess express put misery twice\".\n",
            "Negative sentence scored 0.021743150427937508 \"dollar mimosa make cheap sparkling wine would justify product mid range dollar bad steak frozen fry think steak egg picture meal seem like make leftover dinner night server great staff great place clean cool concept place little rat highly price brunch tab two tip one do trip\".\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "kWwv0r7MbQOm",
        "colab_type": "text"
      },
      "source": [
        "Quantitative evaluation of the model:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ddXZPKSmbOeU",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 266
        },
        "outputId": "e4ae09f1-c1f4-4fa2-ac63-3f8ce22c128b"
      },
      "source": [
        "from sklearn.metrics import roc_auc_score, classification_report\n",
        "\n",
        "ground_truth, scores = [], []\n",
        "with torch.no_grad():\n",
        "  for sentence, tag in test_data:\n",
        "    sentence_in = prepare_sequence(sentence, word_to_ix)\n",
        "    ground_truth.append(tag)\n",
        "    tag_scores = model(sentence_in)\n",
        "    scores.append(torch.mean(torch.exp(tag_scores.squeeze(0)[:, 1])).item())\n",
        "\n",
        "print(\"The AUC is \", roc_auc_score(ground_truth, scores))\n",
        "print(\"Classification report:\\n\", classification_report(ground_truth, np.array(scores)>0.5))"
      ],
      "execution_count": 40,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "The AUC is  0.5153867514064883\n",
            "Classification report:\n",
            "               precision    recall  f1-score   support\n",
            "\n",
            "         0.0       0.21      1.00      0.35       572\n",
            "         1.0       0.00      0.00      0.00      2128\n",
            "\n",
            "    accuracy                           0.21      2700\n",
            "   macro avg       0.11      0.50      0.17      2700\n",
            "weighted avg       0.04      0.21      0.07      2700\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/sklearn/metrics/_classification.py:1272: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
            "  _warn_prf(average, modifier, msg_start, len(result))\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "syYLXyQOwQmP",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "a9da651a-4bc0-4066-fd5a-7d716fb49e02"
      },
      "source": [
        "C = confusion_matrix(ground_truth,np.array(scores)>0.5)\n",
        "neg_pred_perf = C[0,0]/(C[0,0]+C[1,0])\n",
        "fals_disc_rate = C[1,1]/(C[1,1]+C[1,0])\n",
        "neg_pred_perf"
      ],
      "execution_count": 61,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.21185185185185185"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 61
        }
      ]
    }
  ]
}