EmoInt_Embedding_Comparison.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "EmoInt Embedding Comparison",
      "version": "0.3.2",
      "provenance": [],
      "collapsed_sections": [],
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/nakamura41/emotion_detection_notebooks/blob/master/EmoInt_Embedding_Comparison.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "metadata": {
        "id": "c95s_IW-AeyH",
        "colab_type": "code",
        "outputId": "cb288415-f380-4346-8e81-169e631dc860",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 361
        }
      },
      "cell_type": "code",
      "source": [
        "!pip install gensim"
      ],
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Requirement already satisfied: gensim in /usr/local/lib/python3.6/dist-packages (3.6.0)\n",
            "Requirement already satisfied: scipy>=0.18.1 in /usr/local/lib/python3.6/dist-packages (from gensim) (1.1.0)\n",
            "Requirement already satisfied: six>=1.5.0 in /usr/local/lib/python3.6/dist-packages (from gensim) (1.11.0)\n",
            "Requirement already satisfied: numpy>=1.11.3 in /usr/local/lib/python3.6/dist-packages (from gensim) (1.14.6)\n",
            "Requirement already satisfied: smart-open>=1.2.1 in /usr/local/lib/python3.6/dist-packages (from gensim) (1.7.1)\n",
            "Requirement already satisfied: requests in /usr/local/lib/python3.6/dist-packages (from smart-open>=1.2.1->gensim) (2.18.4)\n",
            "Requirement already satisfied: boto>=2.32 in /usr/local/lib/python3.6/dist-packages (from smart-open>=1.2.1->gensim) (2.49.0)\n",
            "Requirement already satisfied: boto3 in /usr/local/lib/python3.6/dist-packages (from smart-open>=1.2.1->gensim) (1.9.42)\n",
            "Requirement already satisfied: bz2file in /usr/local/lib/python3.6/dist-packages (from smart-open>=1.2.1->gensim) (0.98)\n",
            "Requirement already satisfied: idna<2.7,>=2.5 in /usr/local/lib/python3.6/dist-packages (from requests->smart-open>=1.2.1->gensim) (2.6)\n",
            "Requirement already satisfied: chardet<3.1.0,>=3.0.2 in /usr/local/lib/python3.6/dist-packages (from requests->smart-open>=1.2.1->gensim) (3.0.4)\n",
            "Requirement already satisfied: urllib3<1.23,>=1.21.1 in /usr/local/lib/python3.6/dist-packages (from requests->smart-open>=1.2.1->gensim) (1.22)\n",
            "Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.6/dist-packages (from requests->smart-open>=1.2.1->gensim) (2018.10.15)\n",
            "Requirement already satisfied: jmespath<1.0.0,>=0.7.1 in /usr/local/lib/python3.6/dist-packages (from boto3->smart-open>=1.2.1->gensim) (0.9.3)\n",
            "Requirement already satisfied: botocore<1.13.0,>=1.12.42 in /usr/local/lib/python3.6/dist-packages (from boto3->smart-open>=1.2.1->gensim) (1.12.42)\n",
            "Requirement already satisfied: s3transfer<0.2.0,>=0.1.10 in /usr/local/lib/python3.6/dist-packages (from boto3->smart-open>=1.2.1->gensim) (0.1.13)\n",
            "Requirement already satisfied: docutils>=0.10 in /usr/local/lib/python3.6/dist-packages (from botocore<1.13.0,>=1.12.42->boto3->smart-open>=1.2.1->gensim) (0.14)\n",
            "Requirement already satisfied: python-dateutil<3.0.0,>=2.1; python_version >= \"2.7\" in /usr/local/lib/python3.6/dist-packages (from botocore<1.13.0,>=1.12.42->boto3->smart-open>=1.2.1->gensim) (2.5.3)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "colab_type": "code",
        "id": "OdaHCMv2rGKs",
        "outputId": "4101168e-f670-4d28-ba8f-2403dc197380",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 53
        }
      },
      "cell_type": "code",
      "source": [
        "import os\n",
        "import tensorflow as tf\n",
        "import tensorflow_hub as hub\n",
        "from tensorflow import keras\n",
        "import numpy as np\n",
        "import urllib.request\n",
        "import pandas as pd\n",
        "from gensim.test.utils import common_texts, get_tmpfile\n",
        "from gensim.utils import simple_preprocess\n",
        "from gensim.models import Word2Vec\n",
        "from tempfile import gettempdir\n",
        "import zipfile\n",
        "from keras.preprocessing.text import Tokenizer\n",
        "from keras.preprocessing.sequence import pad_sequences\n",
        "from keras.utils import to_categorical\n",
        "from sklearn.model_selection import train_test_split\n",
        "from sklearn.metrics import confusion_matrix, classification_report, accuracy_score, brier_score_loss\n",
        "import matplotlib.pyplot as plt\n",
        "import seaborn as sns\n",
        "\n",
        "print(tf.__version__)"
      ],
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "1.12.0\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "Using TensorFlow backend.\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "metadata": {
        "id": "wWC0XDwDASUM",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "module_url = \"https://tfhub.dev/google/universal-sentence-encoder-large/3\" #@param [\"https://tfhub.dev/google/universal-sentence-encoder/2\", \"https://tfhub.dev/google/universal-sentence-encoder-large/3\"]"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "IElRtsQwASUP",
        "colab_type": "code",
        "outputId": "d008f4ee-cbe8-4c0f-bac7-69c653bf9abe",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "cell_type": "code",
      "source": [
        "# Import the Universal Sentence Encoder's TF Hub module\n",
        "tf_use_embed = hub.Module(module_url)"
      ],
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "INFO:tensorflow:Using /tmp/tfhub_modules to cache modules.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "0RhkgBlRASUR",
        "colab_type": "code",
        "outputId": "22d4b59d-610c-4e48-b419-7d2f4463ea66",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 161
        }
      },
      "cell_type": "code",
      "source": [
        "sentences = [\"I love you so much\", \"I hate that movie, there are no plotline or even interesting scenes\"]\n",
        "# Reduce logging output.\n",
        "tf.logging.set_verbosity(tf.logging.ERROR)\n",
        "\n",
        "with tf.Session() as session:\n",
        "    session.run([tf.global_variables_initializer(), tf.tables_initializer()])\n",
        "    embedded_sentences = session.run(tf_use_embed(sentences))\n",
        "    for i, embedded_sentence in enumerate(np.array(embedded_sentences).tolist()):\n",
        "        print(\"Sentence: {}\".format(sentences[i]))\n",
        "        print(\"Embedding size: {}\".format(len(embedded_sentence)))\n",
        "        message_embedding_snippet = \", \".join(\n",
        "            (str(x) for x in embedded_sentence[:3]))\n",
        "        print(\"Embedding: [{}, ...]\\n\".format(message_embedding_snippet))"
      ],
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Sentence: I love you so much\n",
            "Embedding size: 512\n",
            "Embedding: [-0.03638805076479912, -0.07312509417533875, 0.007526352070271969, ...]\n",
            "\n",
            "Sentence: I hate that movie, there are no plotline or even interesting scenes\n",
            "Embedding size: 512\n",
            "Embedding: [0.016185155138373375, 0.059069763869047165, -0.0056126900017261505, ...]\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "EzmR-eCcASUU",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "def embed_use(sentences):\n",
        "    with tf.Session() as session:\n",
        "        session.run([tf.global_variables_initializer(), tf.tables_initializer()])\n",
        "        embedded_sentences = session.run(tf_use_embed(sentences))\n",
        "        return embedded_sentences"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "colab_type": "code",
        "id": "oBynQmEcrNAy",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "from pathlib import Path\n",
        "\n",
        "class EmoIntDataset(object):\n",
        "    BASE_URL = \"http://saifmohammad.com/WebDocs/\"\n",
        "    TRAIN_URI = \"EmoInt%20Train%20Data/{}-ratings-0to1.train.txt\"\n",
        "    TEST_URI = \"EmoInt%20Test%20Gold%20Data/{}-ratings-0to1.test.gold.txt\"\n",
        "    EMOTION_CLASSES = [\"anger\", \"fear\", \"joy\", \"sadness\"]\n",
        "    DATASET_NAME = \"EmoInt\"\n",
        "  \n",
        "    THRESHOLD = 0.33\n",
        "  \n",
        "    def __load_data_per_class(self, category, class_name, url, threshold=0):\n",
        "        filename = \"./{}_{}_{}.csv\".format(category, self.DATASET_NAME, class_name)\n",
        "        if Path(filename).is_file():\n",
        "            df = pd.read_csv(filename)\n",
        "        else:\n",
        "            resource = urllib.request.urlopen(url)\n",
        "            np_array = np.asarray([line.split('\\t') for line in [line.strip() for line in resource.read().decode('utf-8').splitlines()]])\n",
        "            df = pd.DataFrame(np_array, columns=[\"id\", \"text\", \"emotion\", \"emotion_level\"])\n",
        "            df['emotion_level'] = df['emotion_level'].astype(float)\n",
        "            df = df.query('emotion_level>' + str(threshold))\n",
        "            df.to_csv(filename, index=False)\n",
        "        return df[[\"text\", \"emotion\"]]\n",
        "  \n",
        "    def load_data(self, set_threshold=False):\n",
        "        train_data = None\n",
        "        test_data = None\n",
        "    \n",
        "        for emotion in self.EMOTION_CLASSES:\n",
        "            # load train dataset\n",
        "            train_df = self.__load_data_per_class(\"train\", emotion, self.BASE_URL + self.TRAIN_URI.format(emotion), threshold=(self.THRESHOLD if set_threshold else 0))\n",
        "      \n",
        "            # load test dataset\n",
        "            test_df = self.__load_data_per_class(\"test\", emotion, self.BASE_URL + self.TEST_URI.format(emotion), threshold=(self.THRESHOLD if set_threshold else 0))\n",
        "      \n",
        "            train_data = (train_df if train_data is None else train_data.append(train_df))\n",
        "            test_data = (test_df if test_data is None else test_data.append(test_df))\n",
        "      \n",
        "        return train_data, test_data"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "q9HVS3aCY-lt",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "def print_confusion_matrix(confusion_matrix, class_names, figsize = (4,3), fontsize=15):\n",
        "    \"\"\"Prints a confusion matrix, as returned by sklearn.metrics.confusion_matrix, as a heatmap.\n",
        "    \n",
        "    Arguments\n",
        "    ---------\n",
        "    confusion_matrix: numpy.ndarray\n",
        "        The numpy.ndarray object returned from a call to sklearn.metrics.confusion_matrix. \n",
        "        Similarly constructed ndarrays can also be used.\n",
        "    class_names: list\n",
        "        An ordered list of class names, in the order they index the given confusion matrix.\n",
        "    figsize: tuple\n",
        "        A 2-long tuple, the first value determining the horizontal size of the ouputted figure,\n",
        "        the second determining the vertical size. Defaults to (10,7).\n",
        "    fontsize: int\n",
        "        Font size for axes labels. Defaults to 14.\n",
        "        \n",
        "    Returns\n",
        "    -------\n",
        "    matplotlib.figure.Figure\n",
        "        The resulting confusion matrix figure\n",
        "    \"\"\"\n",
        "    df_cm = pd.DataFrame(\n",
        "        confusion_matrix, index=class_names, columns=class_names, \n",
        "    )\n",
        "    fig = plt.figure(figsize=figsize)\n",
        "    try:\n",
        "        heatmap = sns.heatmap(df_cm, annot=True, fmt=\"d\")\n",
        "    except ValueError:\n",
        "        raise ValueError(\"Confusion matrix values must be integers.\")\n",
        "    heatmap.yaxis.set_ticklabels(heatmap.yaxis.get_ticklabels(), rotation=0, ha='right', fontsize=fontsize)\n",
        "    heatmap.xaxis.set_ticklabels(heatmap.xaxis.get_ticklabels(), rotation=45, ha='right', fontsize=fontsize)\n",
        "    plt.ylabel('True label')\n",
        "    plt.xlabel('Predicted label')\n",
        "    return fig"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "colab_type": "code",
        "id": "tevwfq4Tu6fe",
        "outputId": "02ce5baf-211a-41f3-95c9-bf163ae47b94",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 71
        }
      },
      "cell_type": "code",
      "source": [
        "emo_int_dataset = EmoIntDataset()\n",
        "train_data, test_data = emo_int_dataset.load_data(set_threshold=True)\n",
        "train_data, valid_data = train_test_split(train_data, test_size=0.1, random_state=200, stratify=train_data.emotion)\n",
        "\n",
        "print(train_data.shape)\n",
        "print(valid_data.shape)\n",
        "print(test_data.shape)"
      ],
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(2610, 2)\n",
            "(291, 2)\n",
            "(2508, 2)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "colab_type": "code",
        "id": "J-xXkD_4175U",
        "outputId": "f4d5eadd-f63c-4340-a7d6-2adc95d52450",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 53
        }
      },
      "cell_type": "code",
      "source": [
        "emotions = train_data.emotion.unique()\n",
        "dic = dict()\n",
        "labels = []\n",
        "for i, emotion in enumerate(emotions):\n",
        "    dic[emotion]=i\n",
        "    labels.append(emotion)\n",
        "print(dic)\n",
        "print(labels)"
      ],
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "{'fear': 0, 'anger': 1, 'sadness': 2, 'joy': 3}\n",
            "['fear', 'anger', 'sadness', 'joy']\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "MtZaeo2JASUh",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "train_labels = np.array(train_data.emotion.apply(lambda x:dic[x]))\n",
        "test_labels = np.array(test_data.emotion.apply(lambda x:dic[x]))"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "0wXUaCQeASUe",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "# Embed with Universal Sentence Encoder"
      ]
    },
    {
      "metadata": {
        "id": "AvcnW45qASUg",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "train_embedded_use = embed_use(train_data.text.tolist())\n",
        "test_embedded_use = embed_use(test_data.text.tolist())"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "-mgHsSbYASUk",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "# Universal Sentence Encoder + SVM Classifier"
      ]
    },
    {
      "metadata": {
        "id": "2l8cnyxSASUk",
        "colab_type": "code",
        "outputId": "d581d4b6-6d56-4fed-c7c8-cc7dd3697e74",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 89
        }
      },
      "cell_type": "code",
      "source": [
        "from sklearn import svm\n",
        "clfSVM = svm.SVC(C=10000.0, gamma='auto', kernel='rbf')\n",
        "clfSVM.fit(train_embedded_use, train_labels)"
      ],
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "SVC(C=10000.0, cache_size=200, class_weight=None, coef0=0.0,\n",
              "  decision_function_shape='ovr', degree=3, gamma='auto', kernel='rbf',\n",
              "  max_iter=-1, probability=False, random_state=None, shrinking=True,\n",
              "  tol=0.001, verbose=False)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 13
        }
      ]
    },
    {
      "metadata": {
        "id": "GKvNJ0g9ASUn",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "test_pred_labels = clfSVM.predict(test_embedded_use)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "UJw3zrZDASUp",
        "colab_type": "code",
        "outputId": "05b57053-8c7c-4877-c3a8-bf1004d3fc96",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 89
        }
      },
      "cell_type": "code",
      "source": [
        "cmSVM = confusion_matrix(test_labels, test_pred_labels)\n",
        "cmSVM"
      ],
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[548, 108,  86,  47],\n",
              "       [137, 415,  62,  29],\n",
              "       [141,  63, 291,  38],\n",
              "       [ 54,  36,  45, 408]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 15
        }
      ]
    },
    {
      "metadata": {
        "id": "Z4jFTBjKZgtl",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "print(print_confusion_matrix(cmSVM, class_names=labels))"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "Tt3rbFReASUr",
        "colab_type": "code",
        "outputId": "3abe8c0c-4e8a-4e87-8560-985a4204cf2c",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 179
        }
      },
      "cell_type": "code",
      "source": [
        "reportSVM = classification_report(test_labels, test_pred_labels, target_names=labels)\n",
        "print(reportSVM)"
      ],
      "execution_count": 16,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "             precision    recall  f1-score   support\n",
            "\n",
            "       fear       0.62      0.69      0.66       789\n",
            "      anger       0.67      0.65      0.66       643\n",
            "    sadness       0.60      0.55      0.57       533\n",
            "        joy       0.78      0.75      0.77       543\n",
            "\n",
            "avg / total       0.66      0.66      0.66      2508\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "suU_Yw-fASUw",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "# TfIdfVectorizer + SVM Classifier"
      ]
    },
    {
      "metadata": {
        "id": "crQ4mZBsASUx",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "from sklearn.feature_extraction.text import TfidfVectorizer\n",
        "\n",
        "tfidf_vectorizer = TfidfVectorizer(max_df=70, min_df=2, ngram_range=(1, 3))\n",
        "train_tfidf_vectors = tfidf_vectorizer.fit_transform(train_data.text.tolist())\n",
        "test_tfidf_vectors = tfidf_vectorizer.transform(test_data.text.tolist())"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "efLiuFqgASUz",
        "colab_type": "code",
        "outputId": "91680863-7a94-4cb0-e979-4f7377c0790e",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 89
        }
      },
      "cell_type": "code",
      "source": [
        "from sklearn import svm\n",
        "clf_SVM_Tfidf = svm.SVC(C=10000.0, gamma='auto', kernel='rbf')\n",
        "clf_SVM_Tfidf.fit(train_tfidf_vectors, train_labels)"
      ],
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "SVC(C=10000.0, cache_size=200, class_weight=None, coef0=0.0,\n",
              "  decision_function_shape='ovr', degree=3, gamma='auto', kernel='rbf',\n",
              "  max_iter=-1, probability=False, random_state=None, shrinking=True,\n",
              "  tol=0.001, verbose=False)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 14
        }
      ]
    },
    {
      "metadata": {
        "id": "MoS_XCtUASU3",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "test_pred_labels_Tfidf = clf_SVM_Tfidf.predict(test_tfidf_vectors)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "eBQ0W7MrASU5",
        "colab_type": "code",
        "outputId": "b281f381-2b2f-451c-c7b2-781ca016faad",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 89
        }
      },
      "cell_type": "code",
      "source": [
        "cmSVM_Tfidf = confusion_matrix(test_labels, test_pred_labels_Tfidf)\n",
        "cmSVM_Tfidf"
      ],
      "execution_count": 16,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[625,  47,  89,  28],\n",
              "       [ 82, 497,  46,  18],\n",
              "       [ 94,  34, 392,  13],\n",
              "       [ 53,  15,  21, 454]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 16
        }
      ]
    },
    {
      "metadata": {
        "id": "A08bNtNhZfbF",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "print(print_confusion_matrix(cmSVM_Tfidf, class_names=labels))"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "N5YwSO_ZASU7",
        "colab_type": "code",
        "outputId": "d737c634-27f6-418c-ab8e-08326f8a53c3",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 179
        }
      },
      "cell_type": "code",
      "source": [
        "reportSVM = classification_report(test_labels, test_pred_labels_Tfidf, target_names=labels)\n",
        "print(reportSVM)"
      ],
      "execution_count": 17,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "             precision    recall  f1-score   support\n",
            "\n",
            "       fear       0.73      0.79      0.76       789\n",
            "      anger       0.84      0.77      0.80       643\n",
            "    sadness       0.72      0.74      0.73       533\n",
            "        joy       0.88      0.84      0.86       543\n",
            "\n",
            "avg / total       0.79      0.78      0.79      2508\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "xY_as_tfASU-",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "# CountVectorizer + SVM Classifier"
      ]
    },
    {
      "metadata": {
        "id": "k0LFEc-xASU-",
        "colab_type": "code",
        "outputId": "8f0af2c4-50e7-4f19-a475-77c305bcb9fb",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 89
        }
      },
      "cell_type": "code",
      "source": [
        "from sklearn.feature_extraction.text import CountVectorizer\n",
        "from sklearn.naive_bayes import MultinomialNB\n",
        "\n",
        "# Initialize a CountVectorizer object: count_vectorizer\n",
        "count_vectorizer = CountVectorizer(stop_words=\"english\", analyzer='word', \n",
        "                                   ngram_range=(1, 3), max_df=70, min_df=2, max_features=None)\n",
        "\n",
        "train_count_vectors = count_vectorizer.fit_transform(train_data.text.tolist())\n",
        "test_count_vectors = count_vectorizer.transform(test_data.text.tolist())\n",
        "\n",
        "clf_NB_count = MultinomialNB(alpha=0.1)\n",
        "clf_SVM_count = svm.SVC(C=10000.0, gamma='auto', kernel='rbf')\n",
        "clf_NB_count.fit(train_count_vectors, train_labels)\n",
        "clf_SVM_count.fit(train_count_vectors, train_labels)"
      ],
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "SVC(C=10000.0, cache_size=200, class_weight=None, coef0=0.0,\n",
              "  decision_function_shape='ovr', degree=3, gamma='auto', kernel='rbf',\n",
              "  max_iter=-1, probability=False, random_state=None, shrinking=True,\n",
              "  tol=0.001, verbose=False)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 18
        }
      ]
    },
    {
      "metadata": {
        "id": "8RGYwlr2ASVB",
        "colab_type": "code",
        "outputId": "9b6d7fc8-8a65-4090-c368-0f4ee7823542",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 197
        }
      },
      "cell_type": "code",
      "source": [
        "# get the feature names as numpy array\n",
        "feature_names = np.array(count_vectorizer.get_feature_names())\n",
        "\n",
        "# Sort the coefficients from the model\n",
        "sorted_coef_index = clf_NB_count.coef_[0].argsort()\n",
        "\n",
        "# Find the 10 smallest and 10 largest coefficients\n",
        "# The 10 largest coefficients are being indexed using [:-11:-1] \n",
        "# so the list returned is in order of largest to smallest\n",
        "print('Smallest Coefs:\\n{}\\n'.format(feature_names[sorted_coef_index[:10]]))\n",
        "print('Largest Coefs: \\n{}'.format(feature_names[sorted_coef_index[:-11:-1]]))"
      ],
      "execution_count": 19,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Smallest Coefs:\n",
            "['00' 'ly broadcast paulzimmer' 'ly broadcast mrgummyking'\n",
            " 'ly broadcast maisiev' 'ly broadcast kaylaaa_sartorious'\n",
            " 'ly broadcast katyy' 'ly broadcast kana_blender'\n",
            " 'ly broadcast jaredhorgan' 'ly broadcast rosannahill'\n",
            " 'ly broadcast izzybuzy365']\n",
            "\n",
            "Largest Coefs: \n",
            "['terrorism' 'fear' 'nightmare' 'terrible' 'awful' 'terror' 'nervous'\n",
            " 'horror' 'panic' 'anxiety']\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "0twrTRK3ASVE",
        "colab_type": "code",
        "outputId": "c42fbe9b-7eba-4f9d-e2e5-827500d4dffa",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 89
        }
      },
      "cell_type": "code",
      "source": [
        "test_pred_labels_count = clf_SVM_count.predict(test_count_vectors)\n",
        "\n",
        "cmSVM_count = confusion_matrix(test_labels, test_pred_labels_count)\n",
        "cmSVM_count"
      ],
      "execution_count": 20,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[644,  37,  92,  16],\n",
              "       [ 65, 516,  53,   9],\n",
              "       [ 76,  32, 409,  16],\n",
              "       [ 56,  13,  26, 448]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 20
        }
      ]
    },
    {
      "metadata": {
        "id": "twCkOLBVZdpf",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 282
        },
        "outputId": "f33d2acf-665f-4823-e331-65b08552927d"
      },
      "cell_type": "code",
      "source": [
        "print(print_confusion_matrix(cmSVM_count, class_names=labels))"
      ],
      "execution_count": 32,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Figure(288x216)\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAATkAAAD3CAYAAABmdCOBAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nJ4RwaxJyQgi3JiEnhHBr/wEkAEVdCmw9jAAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<matplotlib.figure.Figure at 0x7f40f0e5c080>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "metadata": {
        "id": "kjEy36O_ASVL",
        "colab_type": "code",
        "outputId": "9a1c988f-20f6-4d5b-c2c1-f124d2acd241",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 179
        }
      },
      "cell_type": "code",
      "source": [
        "reportSVM = classification_report(test_labels, test_pred_labels_count, target_names=labels)\n",
        "print(reportSVM)"
      ],
      "execution_count": 21,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "             precision    recall  f1-score   support\n",
            "\n",
            "       fear       0.77      0.82      0.79       789\n",
            "      anger       0.86      0.80      0.83       643\n",
            "    sadness       0.71      0.77      0.73       533\n",
            "        joy       0.92      0.83      0.87       543\n",
            "\n",
            "avg / total       0.81      0.80      0.81      2508\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "cunC17ZVASVN",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "# CV 5 fold with CountVectorizer + TfIdfTransformer + SVM Classifier "
      ]
    },
    {
      "metadata": {
        "id": "PkQTjwcMASVN",
        "colab_type": "code",
        "outputId": "44e75a0b-aa86-4c62-aca0-e01d21ff2962",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 325
        }
      },
      "cell_type": "code",
      "source": [
        "from sklearn.model_selection import GridSearchCV\n",
        "from sklearn.pipeline import Pipeline\n",
        "from sklearn.feature_extraction.text import TfidfTransformer\n",
        "\n",
        "# Define a pipeline combining a text feature extractor with a simple\n",
        "# classifier\n",
        "pipeline = Pipeline([\n",
        "    ('vect', CountVectorizer()),\n",
        "    ('tfidf', TfidfTransformer()),\n",
        "    ('clf', svm.SVC()),\n",
        "])\n",
        "\n",
        "# uncommenting more parameters will give better exploring power but will\n",
        "# increase processing time in a combinatorial way\n",
        "parameters = {\n",
        "    'vect__max_df': (0.5, 0.75, 1.0),\n",
        "    # 'vect__max_features': (None, 5000, 10000, 50000),\n",
        "    'vect__ngram_range': ((1, 1), (1, 2), (1, 3)),  # unigrams or bigrams\n",
        "    'tfidf__use_idf': (True, False),\n",
        "    # 'tfidf__norm': ('l1', 'l2'),\n",
        "    'clf__C': (10000.0,),\n",
        "    'clf__gamma': ('auto',),\n",
        "    'clf__kernel': ('rbf',),\n",
        "}\n",
        "\n",
        "grid_search = GridSearchCV(pipeline, parameters, cv=5, n_jobs=-1, verbose=1)\n",
        "\n",
        "grid_search.fit(train_data.text.tolist(), train_labels)"
      ],
      "execution_count": 22,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Fitting 5 folds for each of 18 candidates, totalling 90 fits\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "[Parallel(n_jobs=-1)]: Done  46 tasks      | elapsed:  1.6min\n",
            "[Parallel(n_jobs=-1)]: Done  90 out of  90 | elapsed:  3.0min finished\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "GridSearchCV(cv=5, error_score='raise',\n",
              "       estimator=Pipeline(memory=None,\n",
              "     steps=[('vect', CountVectorizer(analyzer='word', binary=False, decode_error='strict',\n",
              "        dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\n",
              "        lowercase=True, max_df=1.0, max_features=None, min_df=1,\n",
              "        ngram_range=(1, 1), preprocessor=None, stop_words=None,\n",
              "        strip...,\n",
              "  max_iter=-1, probability=False, random_state=None, shrinking=True,\n",
              "  tol=0.001, verbose=False))]),\n",
              "       fit_params=None, iid=True, n_jobs=-1,\n",
              "       param_grid={'vect__max_df': (0.5, 0.75, 1.0), 'vect__ngram_range': ((1, 1), (1, 2), (1, 3)), 'tfidf__use_idf': (True, False), 'clf__C': (10000.0,), 'clf__gamma': ('auto',), 'clf__kernel': ('rbf',)},\n",
              "       pre_dispatch='2*n_jobs', refit=True, return_train_score='warn',\n",
              "       scoring=None, verbose=1)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 22
        }
      ]
    },
    {
      "metadata": {
        "id": "CwNXm7oTASVQ",
        "colab_type": "code",
        "outputId": "59d55aff-d47e-4ca8-c67e-02070e4b85f9",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 161
        }
      },
      "cell_type": "code",
      "source": [
        "print(\"Best score: %0.3f\" % grid_search.best_score_)\n",
        "print(\"Best parameters set:\")\n",
        "best_parameters = grid_search.best_estimator_.get_params()\n",
        "for param_name in sorted(parameters.keys()):\n",
        "    print(\"\\t%s: %r\" % (param_name, best_parameters[param_name]))"
      ],
      "execution_count": 23,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Best score: 0.832\n",
            "Best parameters set:\n",
            "\tclf__C: 10000.0\n",
            "\tclf__gamma: 'auto'\n",
            "\tclf__kernel: 'rbf'\n",
            "\ttfidf__use_idf: True\n",
            "\tvect__max_df: 0.5\n",
            "\tvect__ngram_range: (1, 1)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "ohWHM2dzASVU",
        "colab_type": "code",
        "outputId": "02c6e4bf-99cb-42fe-bed4-fa04394921c2",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 89
        }
      },
      "cell_type": "code",
      "source": [
        "test_pred_labels_grid = grid_search.predict(test_data.text.tolist())\n",
        "\n",
        "cmSVM_grid = confusion_matrix(test_labels, test_pred_labels_grid)\n",
        "cmSVM_grid"
      ],
      "execution_count": 24,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[630,  49,  86,  24],\n",
              "       [ 85, 483,  57,  18],\n",
              "       [ 89,  41, 387,  16],\n",
              "       [ 49,  18,  21, 455]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 24
        }
      ]
    },
    {
      "metadata": {
        "id": "Z2GtWlo1ZbpG",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 282
        },
        "outputId": "720a1107-1158-495d-b329-271ac8b0c88c"
      },
      "cell_type": "code",
      "source": [
        "print(print_confusion_matrix(cmSVM_grid, class_names=labels))"
      ],
      "execution_count": 33,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Figure(288x216)\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAATkAAAD3CAYAAABmdCOBAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nhBBGTZKcEMKoSZITQhg1SXJCCKP2P+tJ2eo9B62oAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<matplotlib.figure.Figure at 0x7f40f0db1e80>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "metadata": {
        "id": "IOxPmd3tASVW",
        "colab_type": "code",
        "outputId": "c0a15fe4-94f8-42ee-fb39-5a0c01bd05ea",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 179
        }
      },
      "cell_type": "code",
      "source": [
        "reportSVM_grid = classification_report(test_labels, test_pred_labels_grid, target_names=labels)\n",
        "print(reportSVM_grid)"
      ],
      "execution_count": 25,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "             precision    recall  f1-score   support\n",
            "\n",
            "       fear       0.74      0.80      0.77       789\n",
            "      anger       0.82      0.75      0.78       643\n",
            "    sadness       0.70      0.73      0.71       533\n",
            "        joy       0.89      0.84      0.86       543\n",
            "\n",
            "avg / total       0.78      0.78      0.78      2508\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "Sgdaca1cASVZ",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "# CV 5 fold with CountVectorizer + TfIdfTransformer + SGDClassifier "
      ]
    },
    {
      "metadata": {
        "id": "ARXDnixgASVZ",
        "colab_type": "code",
        "outputId": "55e15073-0be9-44e0-d0c8-e3dd98c0b2a8",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 307
        }
      },
      "cell_type": "code",
      "source": [
        "from sklearn.model_selection import GridSearchCV\n",
        "from sklearn.pipeline import Pipeline\n",
        "from sklearn.linear_model import SGDClassifier\n",
        "from sklearn.feature_extraction.text import TfidfTransformer\n",
        "\n",
        "# Define a pipeline combining a text feature extractor with a simple\n",
        "# classifier\n",
        "pipeline = Pipeline([\n",
        "    ('vect', CountVectorizer()),\n",
        "    ('tfidf', TfidfTransformer()),\n",
        "    ('clf', SGDClassifier()),\n",
        "])\n",
        "\n",
        "# uncommenting more parameters will give better exploring power but will\n",
        "# increase processing time in a combinatorial way\n",
        "parameters = {\n",
        "    'vect__max_df': (0.5, 0.75, 1.0),\n",
        "    # 'vect__max_features': (None, 5000, 10000, 50000),\n",
        "    'vect__ngram_range': ((1, 1), (1, 2)),  # unigrams or bigrams\n",
        "    # 'tfidf__use_idf': (True, False),\n",
        "    # 'tfidf__norm': ('l1', 'l2'),\n",
        "    'clf__max_iter': (5,),\n",
        "    'clf__alpha': (0.00001, 0.000001),\n",
        "    'clf__penalty': ('l2', 'elasticnet'),\n",
        "    # 'clf__max_iter': (10, 50, 80),\n",
        "}\n",
        "\n",
        "grid_search = GridSearchCV(pipeline, parameters, cv=5, n_jobs=-1, verbose=1)\n",
        "\n",
        "grid_search.fit(train_data.text.tolist(), train_labels)"
      ],
      "execution_count": 26,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Fitting 5 folds for each of 24 candidates, totalling 120 fits\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "[Parallel(n_jobs=-1)]: Done  46 tasks      | elapsed:    6.7s\n",
            "[Parallel(n_jobs=-1)]: Done 120 out of 120 | elapsed:   17.8s finished\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "GridSearchCV(cv=5, error_score='raise',\n",
              "       estimator=Pipeline(memory=None,\n",
              "     steps=[('vect', CountVectorizer(analyzer='word', binary=False, decode_error='strict',\n",
              "        dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\n",
              "        lowercase=True, max_df=1.0, max_features=None, min_df=1,\n",
              "        ngram_range=(1, 1), preprocessor=None, stop_words=None,\n",
              "        strip...='l2', power_t=0.5, random_state=None,\n",
              "       shuffle=True, tol=None, verbose=0, warm_start=False))]),\n",
              "       fit_params=None, iid=True, n_jobs=-1,\n",
              "       param_grid={'vect__max_df': (0.5, 0.75, 1.0), 'vect__ngram_range': ((1, 1), (1, 2)), 'clf__max_iter': (5,), 'clf__alpha': (1e-05, 1e-06), 'clf__penalty': ('l2', 'elasticnet')},\n",
              "       pre_dispatch='2*n_jobs', refit=True, return_train_score='warn',\n",
              "       scoring=None, verbose=1)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 26
        }
      ]
    },
    {
      "metadata": {
        "id": "r5_eE1yDASVb",
        "colab_type": "code",
        "outputId": "6b807186-618d-4aab-9346-713ce6f0c3ab",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 143
        }
      },
      "cell_type": "code",
      "source": [
        "print(\"Best score: %0.3f\" % grid_search.best_score_)\n",
        "print(\"Best parameters set:\")\n",
        "best_parameters = grid_search.best_estimator_.get_params()\n",
        "for param_name in sorted(parameters.keys()):\n",
        "    print(\"\\t%s: %r\" % (param_name, best_parameters[param_name]))"
      ],
      "execution_count": 27,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Best score: 0.828\n",
            "Best parameters set:\n",
            "\tclf__alpha: 1e-05\n",
            "\tclf__max_iter: 5\n",
            "\tclf__penalty: 'elasticnet'\n",
            "\tvect__max_df: 0.75\n",
            "\tvect__ngram_range: (1, 2)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "tyT61VJzASVd",
        "colab_type": "code",
        "outputId": "2da566de-6963-463b-e4c2-b837d2048227",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 89
        }
      },
      "cell_type": "code",
      "source": [
        "train_pred_labels_grid = grid_search.predict(train_data.text.tolist())\n",
        "\n",
        "cmSVM_grid = confusion_matrix(train_labels, train_pred_labels_grid)\n",
        "cmSVM_grid"
      ],
      "execution_count": 28,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[814,   2,   7,   0],\n",
              "       [  0, 652,   7,   0],\n",
              "       [  7,   4, 545,   0],\n",
              "       [  0,   0,   0, 572]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 28
        }
      ]
    },
    {
      "metadata": {
        "id": "KrO0xp44ASVg",
        "colab_type": "code",
        "outputId": "72dd61d8-8051-447d-b737-ac80247d462c",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 89
        }
      },
      "cell_type": "code",
      "source": [
        "test_pred_labels_grid = grid_search.predict(test_data.text.tolist())\n",
        "\n",
        "cmSVM_grid = confusion_matrix(test_labels, test_pred_labels_grid)\n",
        "cmSVM_grid"
      ],
      "execution_count": 29,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[596,  55,  92,  46],\n",
              "       [ 65, 494,  46,  38],\n",
              "       [ 82,  45, 387,  19],\n",
              "       [ 39,  14,  32, 458]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 29
        }
      ]
    },
    {
      "metadata": {
        "id": "O_vBiSMbYrQp",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 282
        },
        "outputId": "4e7b9291-e459-43dd-c633-1a568fd5fd9d"
      },
      "cell_type": "code",
      "source": [
        "print(print_confusion_matrix(cmSVM_grid, class_names=labels))"
      ],
      "execution_count": 30,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Figure(288x216)\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAATkAAAD3CAYAAABmdCOBAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nhBBGTZKcEMKoSZITQhg1SXJCCKP2P+tJ2eo9B62oAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<matplotlib.figure.Figure at 0x7f40f2b27e48>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "metadata": {
        "id": "ZBccwVRWASVh",
        "colab_type": "code",
        "outputId": "ce1f6c9b-dd32-4136-f013-c7f0aa94bc89",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 179
        }
      },
      "cell_type": "code",
      "source": [
        "reportSVM_grid = classification_report(test_labels, test_pred_labels_grid, target_names=labels)\n",
        "print(reportSVM_grid)"
      ],
      "execution_count": 31,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "             precision    recall  f1-score   support\n",
            "\n",
            "       fear       0.76      0.76      0.76       789\n",
            "      anger       0.81      0.77      0.79       643\n",
            "    sadness       0.69      0.73      0.71       533\n",
            "        joy       0.82      0.84      0.83       543\n",
            "\n",
            "avg / total       0.77      0.77      0.77      2508\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "oDLbHXHUASVk",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "# spaCy vectorization + SVM Classifier"
      ]
    },
    {
      "metadata": {
        "id": "VjGysCs9CjJC",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 271
        },
        "outputId": "a23e9274-08da-40ff-dd84-365fcaca9562"
      },
      "cell_type": "code",
      "source": [
        "!python -m spacy download en_core_web_lg"
      ],
      "execution_count": 36,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Collecting en_core_web_lg==2.0.0 from https://github.com/explosion/spacy-models/releases/download/en_core_web_lg-2.0.0/en_core_web_lg-2.0.0.tar.gz#egg=en_core_web_lg==2.0.0\n",
            "\u001b[?25l  Downloading https://github.com/explosion/spacy-models/releases/download/en_core_web_lg-2.0.0/en_core_web_lg-2.0.0.tar.gz (852.3MB)\n",
            "\u001b[K    100% |████████████████████████████████| 852.3MB 31.5MB/s \n",
            "\u001b[?25hInstalling collected packages: en-core-web-lg\n",
            "  Running setup.py install for en-core-web-lg ... \u001b[?25l-\b \b\\\b \b|\b \b/\b \b-\b \b\\\b \b|\b \b/\b \b-\b \b\\\b \b|\b \bdone\n",
            "\u001b[?25hSuccessfully installed en-core-web-lg-2.0.0\n",
            "\n",
            "\u001b[93m    Linking successful\u001b[0m\n",
            "    /usr/local/lib/python3.6/dist-packages/en_core_web_lg -->\n",
            "    /usr/local/lib/python3.6/dist-packages/spacy/data/en_core_web_lg\n",
            "\n",
            "    You can now load the model via spacy.load('en_core_web_lg')\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "C5Tw3LHsASVl",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "import spacy\n",
        "\n",
        "nlp = spacy.load('en_core_web_lg')\n",
        "train_spacy_vectors = train_data.text.apply(lambda x: nlp(x).vector).tolist()\n",
        "test_spacy_vectors = test_data.text.apply(lambda x: nlp(x).vector).tolist()"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "cVMwEmhsASVo",
        "colab_type": "code",
        "outputId": "e75ddfd6-219b-47e8-f2a2-5bfe33e252f2",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 89
        }
      },
      "cell_type": "code",
      "source": [
        "clf_SVM_spacy = svm.SVC(C=10000.0, gamma='auto', kernel='rbf')\n",
        "clf_SVM_spacy.fit(train_spacy_vectors, train_labels)"
      ],
      "execution_count": 38,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "SVC(C=10000.0, cache_size=200, class_weight=None, coef0=0.0,\n",
              "  decision_function_shape='ovr', degree=3, gamma='auto', kernel='rbf',\n",
              "  max_iter=-1, probability=False, random_state=None, shrinking=True,\n",
              "  tol=0.001, verbose=False)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 38
        }
      ]
    },
    {
      "metadata": {
        "id": "HPESeGFEASVs",
        "colab_type": "code",
        "outputId": "a77a488c-8f34-484d-e452-6d02d0eaf1ea",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 89
        }
      },
      "cell_type": "code",
      "source": [
        "test_pred_labels_spacy = clf_SVM_spacy.predict(test_spacy_vectors)\n",
        "\n",
        "cmSVM_spacy = confusion_matrix(test_labels, test_pred_labels_spacy)\n",
        "cmSVM_spacy"
      ],
      "execution_count": 39,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[536,  97, 113,  43],\n",
              "       [117, 432,  74,  20],\n",
              "       [138,  52, 308,  35],\n",
              "       [ 49,  26,  41, 427]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 39
        }
      ]
    },
    {
      "metadata": {
        "id": "j90TW-aoASVv",
        "colab_type": "code",
        "outputId": "d562e9bd-8ea4-4f38-f69f-204c196491c5",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 179
        }
      },
      "cell_type": "code",
      "source": [
        "reportSVM_spacy = classification_report(test_labels, test_pred_labels_spacy, target_names=labels)\n",
        "print(reportSVM_spacy)"
      ],
      "execution_count": 40,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "             precision    recall  f1-score   support\n",
            "\n",
            "       fear       0.64      0.68      0.66       789\n",
            "      anger       0.71      0.67      0.69       643\n",
            "    sadness       0.57      0.58      0.58       533\n",
            "        joy       0.81      0.79      0.80       543\n",
            "\n",
            "avg / total       0.68      0.68      0.68      2508\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "dYIUwbr5ASVz",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "# Using spaCy TextCategorizer"
      ]
    },
    {
      "metadata": {
        "id": "zbrLaQf_ASVz",
        "colab_type": "code",
        "outputId": "d3e84bd1-38b8-4ef8-eb5c-e297b97551b3",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "cell_type": "code",
      "source": [
        "labels"
      ],
      "execution_count": 41,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['fear', 'anger', 'sadness', 'joy']"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 41
        }
      ]
    },
    {
      "metadata": {
        "id": "qYJWx21QASV2",
        "colab_type": "code",
        "outputId": "e1efe2f9-6110-45c7-a812-e1cd49fcaa33",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "cell_type": "code",
      "source": [
        "from spacy.pipeline import TextCategorizer\n",
        "\n",
        "model = 'en_core_web_lg'\n",
        "nlp = spacy.load(model)  # load existing spaCy model\n",
        "print(\"Loaded model '%s'\" % model)\n",
        "\n",
        "textcat = nlp.create_pipe('textcat')\n",
        "nlp.add_pipe(textcat, last=True)"
      ],
      "execution_count": 42,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Loaded model 'en_core_web_lg'\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "8Iq1guxgASV4",
        "colab_type": "code",
        "outputId": "0b65d501-2d4f-4833-f155-af00456ff8db",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "cell_type": "code",
      "source": [
        "# assign label\n",
        "for label in labels:\n",
        "    textcat.add_label(label)\n",
        "    \n",
        "textcat.labels"
      ],
      "execution_count": 43,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['fear', 'anger', 'sadness', 'joy']"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 43
        }
      ]
    },
    {
      "metadata": {
        "id": "_6YDUTDuASV6",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "train_texts = train_data.text.tolist()\n",
        "valid_texts = valid_data.text.tolist()\n",
        "test_texts = test_data.text.tolist()\n",
        "\n",
        "train_labels = np.array(train_data.emotion.apply(lambda x:dic[x]))\n",
        "valid_labels = np.array(valid_data.emotion.apply(lambda x:dic[x]))\n",
        "test_labels = np.array(test_data.emotion.apply(lambda x:dic[x]))"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "Ni-mxu8FASV-",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "def to_spacy_categorical(label_data, labels):\n",
        "    new_data = dict()\n",
        "    for label in labels:\n",
        "        new_data[label] = (label_data == label)\n",
        "    return new_data"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "q3i7mn0zASWA",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "train_encoded_labels = np.array(train_data.emotion.apply(lambda x:to_spacy_categorical(x, labels)))\n",
        "valid_encoded_labels = np.array(valid_data.emotion.apply(lambda x:to_spacy_categorical(x, labels)))\n",
        "test_encoded_labels = np.array(test_data.emotion.apply(lambda x:to_spacy_categorical(x, labels)))"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "D4b4BSk6ASWD",
        "colab_type": "code",
        "outputId": "7944220f-5dd6-4dab-f705-e44a31fd45c0",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 161
        }
      },
      "cell_type": "code",
      "source": [
        "train_encoded_labels"
      ],
      "execution_count": 47,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([{'fear': True, 'anger': False, 'sadness': False, 'joy': False},\n",
              "       {'fear': True, 'anger': False, 'sadness': False, 'joy': False},\n",
              "       {'fear': False, 'anger': True, 'sadness': False, 'joy': False},\n",
              "       ...,\n",
              "       {'fear': False, 'anger': True, 'sadness': False, 'joy': False},\n",
              "       {'fear': False, 'anger': False, 'sadness': False, 'joy': True},\n",
              "       {'fear': False, 'anger': False, 'sadness': False, 'joy': True}],\n",
              "      dtype=object)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 47
        }
      ]
    },
    {
      "metadata": {
        "id": "5MNtZ2hFASWF",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "train_spacy_data = list(zip(train_texts, [{'cats': cats} for cats in train_encoded_labels]))\n",
        "valid_spacy_data = list(zip(valid_texts, [{'cats': cats} for cats in valid_encoded_labels]))\n",
        "test_spacy_data = list(zip(test_texts, [{'cats': cats} for cats in test_encoded_labels]))"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "k5XLQpIoASWG",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "from spacy.util import minibatch\n",
        "\n",
        "def evaluate(tokenizer, textcat, texts, cats):\n",
        "    docs = (tokenizer(text) for text in texts)\n",
        "    tp = 1e-8  # True positives\n",
        "    fp = 1e-8  # False positives\n",
        "    fn = 1e-8  # False negatives\n",
        "    tn = 1e-8  # True negatives\n",
        "    for i, doc in enumerate(textcat.pipe(docs)):\n",
        "        gold = cats[i]\n",
        "        for label, score in doc.cats.items():\n",
        "            if label not in gold:\n",
        "                continue\n",
        "            if score >= 0.5 and gold[label] >= 0.5:\n",
        "                tp += 1.\n",
        "            elif score >= 0.5 and gold[label] < 0.5:\n",
        "                fp += 1.\n",
        "            elif score < 0.5 and gold[label] < 0.5:\n",
        "                tn += 1\n",
        "            elif score < 0.5 and gold[label] >= 0.5:\n",
        "                fn += 1\n",
        "    precision = tp / (tp + fp)\n",
        "    recall = tp / (tp + fn)\n",
        "    f_score = 2 * (precision * recall) / (precision + recall)\n",
        "    return {'textcat_p': precision, 'textcat_r': recall, 'textcat_f': f_score}\n",
        "\n",
        "\n",
        "def training(spacy_nlp, train_data, valid_data, epoch=20):\n",
        "    valid_texts, valid_labels = zip(*valid_data)\n",
        "    \n",
        "    other_pipes = [pipe for pipe in spacy_nlp.pipe_names if pipe != 'textcat']\n",
        "    with spacy_nlp.disable_pipes(*other_pipes):  # only train textcat\n",
        "        optimizer = spacy_nlp.begin_training()\n",
        "        print(\"Training the model...\")\n",
        "        print('{:^5}\\t{:^5}\\t{:^5}\\t{:^5}\\t{:^5}'.format('EPOCH', 'LOSS', 'P', 'R', 'F'))\n",
        "        for i in range(epoch):\n",
        "            losses = {}\n",
        "            # batch up the examples using spaCy's minibatch\n",
        "            batches = minibatch(train_data, size=4)\n",
        "\n",
        "            for batch in batches:\n",
        "                texts, annotations = zip(*batch)\n",
        "                spacy_nlp.update(texts, annotations, sgd=optimizer, drop=0.2, losses=losses)\n",
        "\n",
        "            with textcat.model.use_params(optimizer.averages):\n",
        "                # evaluate on the dev data split off in load_data()\n",
        "                scores = evaluate(spacy_nlp.tokenizer, textcat, valid_texts, valid_labels)\n",
        "\n",
        "            print('{0:.0f}\\t{1:.3f}\\t{2:.3f}\\t{3:.3f}\\t{4:.3f}'  # print a simple table\n",
        "                  .format(i + 1, losses['textcat'], scores['textcat_p'],\n",
        "                          scores['textcat_r'], scores['textcat_f']))\n",
        "\n",
        "    return spacy_nlp"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "xWxmVAB6ASWJ",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "nlp = training(nlp, train_spacy_data, valid_spacy_data, epoch=20)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "LDcKudmEASWK",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    }
  ]
}