w2v_dl.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "w2v_dl.ipynb",
      "provenance": [],
      "collapsed_sections": [
        "iq9qOwg5KdaH",
        "oXitblIcv8QT",
        "rdEiP82M8dt4",
        "UlnGravi5miL"
      ]
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "code",
      "metadata": {
        "id": "Avd7lq5pA_QY"
      },
      "source": [
        "import numpy as np \n",
        "import pandas as pd\n",
        "import os\n",
        "\n",
        "from sklearn.feature_extraction.text import CountVectorizer\n",
        "from sklearn.metrics import confusion_matrix, accuracy_score\n",
        "from sklearn.model_selection import train_test_split\n",
        "\n",
        "from keras.preprocessing.text import Tokenizer\n",
        "from keras.preprocessing.sequence import pad_sequences\n",
        "from keras.initializers import Constant\n",
        "from keras.models import Model\n",
        "from keras.layers import *\n",
        "from keras.utils.np_utils import to_categorical\n",
        "import re\n",
        "\n",
        "import matplotlib.pyplot as plt\n",
        "%matplotlib inline"
      ],
      "execution_count": 3,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "N0_MOZTzBhnJ",
        "outputId": "835fa6d1-bee6-452c-8aa8-2651439ef0b3"
      },
      "source": [
        "from google.colab import drive\n",
        "drive.mount(\"/content/drive\")"
      ],
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Mounted at /content/drive\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "niVz-N9ewqao",
        "outputId": "28f6b97a-fe68-4b1f-9bed-d32fc49cdb0a"
      },
      "source": [
        "!ls '/content/drive/My Drive/dataset/'"
      ],
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "dataset.xlsx  glove.6B.50d.txt\tglove.6B.50d.txt.zip\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 204
        },
        "id": "f7PsRh9OBf2B",
        "outputId": "bb35e2e7-303d-413a-f0a6-92cf62f81937"
      },
      "source": [
        "df = pd.read_excel('/content/drive/My Drive/dataset/dataset.xlsx', engine='openpyxl')\n",
        "df.head()"
      ],
      "execution_count": 6,
      "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>URL</th>\n",
              "      <th>Text</th>\n",
              "      <th>Sentiment</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>http://www.imdb.com/title/tt0210075/usercomments</td>\n",
              "      <td>Girlfight follows a project dwelling New York ...</td>\n",
              "      <td>POS</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>http://www.imdb.com/title/tt0337640/usercomments</td>\n",
              "      <td>Hollywood North is an euphemism from the movie...</td>\n",
              "      <td>POS</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>http://www.imdb.com/title/tt0303549/usercomments</td>\n",
              "      <td>That '70s Show is definitely the funniest show...</td>\n",
              "      <td>POS</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>http://www.imdb.com/title/tt0716825/usercomments</td>\n",
              "      <td>9/10- 30 minutes of pure holiday terror. Okay,...</td>\n",
              "      <td>POS</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>http://www.imdb.com/title/tt0182225/usercomments</td>\n",
              "      <td>A series of random, seemingly insignificant th...</td>\n",
              "      <td>POS</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                                                URL  ... Sentiment\n",
              "0  http://www.imdb.com/title/tt0210075/usercomments  ...       POS\n",
              "1  http://www.imdb.com/title/tt0337640/usercomments  ...       POS\n",
              "2  http://www.imdb.com/title/tt0303549/usercomments  ...       POS\n",
              "3  http://www.imdb.com/title/tt0716825/usercomments  ...       POS\n",
              "4  http://www.imdb.com/title/tt0182225/usercomments  ...       POS\n",
              "\n",
              "[5 rows x 3 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 6
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "LlZZtuNADUHb"
      },
      "source": [
        "### Removing urls from the reviews:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Njh7xslSCuhw"
      },
      "source": [
        "def clean_str(in_str):\n",
        "    in_str = str(in_str)\n",
        "    # replace urls with 'url'\n",
        "    in_str = re.sub(r\"(https?:\\/\\/(?:www\\.|(?!www))[a-zA-Z0-9][a-zA-Z0-9-]+[a-zA-Z0-9]\\.[^\\s]{2,}|www\\.[a-zA-Z0-9][a-zA-Z0-9-]+[a-zA-Z0-9]\\.[^\\s]{2,}|https?:\\/\\/(?:www\\.|(?!www))[a-zA-Z0-9]\\.[^\\s]{2,}|www\\.[a-zA-Z0-9]\\.[^\\s]{2,})\", \"url\", in_str)\n",
        "    in_str = re.sub(r'([^\\s\\w]|_)+', '', in_str)\n",
        "    return in_str.strip().lower()"
      ],
      "execution_count": 9,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ghh5NpzJC2m8"
      },
      "source": [
        "df['text'] = df['Text'].apply(clean_str)"
      ],
      "execution_count": 10,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "NVvS8MNcDDeC",
        "outputId": "25bfe144-e23e-4d88-91bf-0c67e2c60f84"
      },
      "source": [
        "df.Sentiment.value_counts()"
      ],
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "NEG    1000\n",
              "POS    999 \n",
              "Name: Sentiment, dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 12
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "nQFQncTwDQYD"
      },
      "source": [
        "df_0 = df[df['Sentiment'] == 'POS'].sample(frac=1)\n",
        "df_1 = df[df['Sentiment'] == 'NEG'].sample(frac=1)\n",
        "\n",
        "# we want a balanced set for training against - there are 7072 `0` examples\n",
        "sample_size = 999\n",
        "\n",
        "data = pd.concat([df_0.head(sample_size), df_1.head(sample_size)]).sample(frac=1)"
      ],
      "execution_count": 13,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "rCcLl4CvELgf",
        "outputId": "a168a80a-a414-4292-845b-8818a17e22dd"
      },
      "source": [
        "data.Sentiment.value_counts()"
      ],
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "NEG    999\n",
              "POS    999\n",
              "Name: Sentiment, dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 14
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "YEh3aT-gEIYN",
        "outputId": "6e5cb756-7e0e-48a1-e7f9-0c50f2eb3689"
      },
      "source": [
        "data['l'] = data['text'].apply(lambda x: len(str(x).split(' ')))\n",
        "print(\"mean length of sentence: \" + str(data.l.mean()))\n",
        "print(\"max length of sentence: \" + str(data.l.max()))\n",
        "print(\"std dev length of sentence: \" + str(data.l.std()))"
      ],
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "mean length of sentence: 258.474974974975\n",
            "max length of sentence: 1843\n",
            "std dev length of sentence: 180.6102134381438\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Nt-6ohBSjFZC"
      },
      "source": [
        "## Tokenization Of Sentences"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ooMx0mseETo0"
      },
      "source": [
        "sequence_length = 260 # On an average each review has around 260 words"
      ],
      "execution_count": 16,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "XDgTGc6EEWzM",
        "outputId": "ee3081fc-05ec-4249-d3b8-a0e366708a86"
      },
      "source": [
        "max_features = 25000 # this is the number of words we care about\n",
        "\n",
        "tokenizer = Tokenizer(num_words=max_features, split=' ', oov_token='<unw>')\n",
        "tokenizer.fit_on_texts(df['text'].values)\n",
        "\n",
        "# this takes our sentences and replaces each word with an integer\n",
        "X = tokenizer.texts_to_sequences(df['text'].values)\n",
        "\n",
        "# we then pad the sequences so they're all the same length (sequence_length)\n",
        "X = pad_sequences(X, sequence_length)\n",
        "\n",
        "y = pd.get_dummies(df['Sentiment']).values\n",
        "\n",
        "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3)\n",
        "\n",
        "print(\"test set size \" + str(len(X_test)))"
      ],
      "execution_count": 17,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "test set size 600\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "q217EP_wlcCE"
      },
      "source": [
        "\n",
        "\n",
        "```\n",
        "# This is formatted as code\n",
        "```\n",
        "\n",
        "### Transformation and Vectorization"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "sgaRSIJywBrd",
        "outputId": "d2bb3862-7a22-4337-fda7-26557211a50b"
      },
      "source": [
        "embeddings_index = {}\n",
        "f = open('/content/drive/My Drive/dataset/glove.6B.50d.txt')\n",
        "for line in f:\n",
        "    values = line.split()\n",
        "    word = values[0]\n",
        "    coefs = np.asarray(values[1:], dtype='float32')\n",
        "    embeddings_index[word] = coefs\n",
        "f.close()\n",
        "print('Found %s word vectors.' % len(embeddings_index))"
      ],
      "execution_count": 23,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Found 400000 word vectors.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "O7VJU6P1xXLm",
        "outputId": "ff6bfca9-2b71-4a8b-f895-1d0d176e9dae"
      },
      "source": [
        "word_index = tokenizer.word_index\n",
        "print('Found %s unique tokens.' % len(word_index))"
      ],
      "execution_count": 24,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Found 34724 unique tokens.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "B-uq4bw5yDCi",
        "outputId": "58bced64-b02e-42a2-cc50-52f2eb2b2ffb"
      },
      "source": [
        "num_words = min(max_features, len(word_index)) + 1\n",
        "print(num_words)\n",
        "non_exist_word_count=0\n",
        "embedding_dim = 50 #Can try 50/100/200\n",
        "\n",
        "# first create a matrix of zeros, this is our embedding matrix\n",
        "embedding_matrix = np.zeros((num_words, embedding_dim))\n",
        "\n",
        "# for each word in our tokenizer lets try to find that work in our w2v model\n",
        "for word, i in word_index.items():\n",
        "    if i > max_features:\n",
        "        continue\n",
        "    embedding_vector = embeddings_index.get(word)\n",
        "    if embedding_vector is not None:\n",
        "        # we found the word - add that words vector to the matrix\n",
        "        embedding_matrix[i] = embedding_vector\n",
        "    else:\n",
        "        non_exist_word_count=non_exist_word_count+1     \n",
        "        # doesn't exist, assign a random vector\n",
        "        embedding_matrix[i] = np.random.randn(embedding_dim)\n",
        "print('Count for non exist words in dict :',non_exist_word_count)"
      ],
      "execution_count": 25,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "25001\n",
            "Count for non exist words in dict : 5489\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "UlnGravi5miL"
      },
      "source": [
        "### CNN"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "dhO7lULP4ffr"
      },
      "source": [
        "from tensorflow.python.keras import regularizers\n",
        "\n",
        "inputs_3 = Input(shape=(sequence_length,), dtype='int32')\n",
        "num_filters = 50\n",
        "embedding_layer_3 = Embedding(num_words,\n",
        "                            embedding_dim,\n",
        "                            embeddings_initializer=Constant(embedding_matrix),\n",
        "                            input_length=sequence_length,\n",
        "                            trainable=True)(inputs_3)\n",
        "reshape_3 = Reshape((sequence_length, embedding_dim, 1))(embedding_layer_3)\n",
        "conv_0_3 = Conv2D(num_filters, kernel_size=(3, embedding_dim), activation='relu', kernel_regularizer=regularizers.l2(1e-3))(reshape_3)\n",
        "maxpool_0_3 = MaxPool2D(pool_size=(sequence_length - 3 + 1, 1), strides=(1,1), padding='valid')(conv_0_3)\n",
        "\n",
        "flatten_3 = Flatten()(maxpool_0_3)\n",
        "\n",
        "# do dropout and predict\n",
        "dropout_3 = Dropout(0.5)(flatten_3)\n",
        "output_3 = Dense(units=2, activation='softmax')(dropout_3)"
      ],
      "execution_count": 30,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "38i9bzAV4nWS",
        "outputId": "c4de8e40-a132-4f09-d8ca-cdb23ab73d8a"
      },
      "source": [
        "model_3 = Model(inputs=inputs_3, outputs=output_3)\n",
        "model_3.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])\n",
        "print(model_3.summary())"
      ],
      "execution_count": 31,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"model_1\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "input_3 (InputLayer)         [(None, 260)]             0         \n",
            "_________________________________________________________________\n",
            "embedding_1 (Embedding)      (None, 260, 50)           1250050   \n",
            "_________________________________________________________________\n",
            "reshape_1 (Reshape)          (None, 260, 50, 1)        0         \n",
            "_________________________________________________________________\n",
            "conv2d_1 (Conv2D)            (None, 258, 1, 50)        7550      \n",
            "_________________________________________________________________\n",
            "max_pooling2d_1 (MaxPooling2 (None, 1, 1, 50)          0         \n",
            "_________________________________________________________________\n",
            "flatten_1 (Flatten)          (None, 50)                0         \n",
            "_________________________________________________________________\n",
            "dropout_1 (Dropout)          (None, 50)                0         \n",
            "_________________________________________________________________\n",
            "dense_1 (Dense)              (None, 2)                 102       \n",
            "=================================================================\n",
            "Total params: 1,257,702\n",
            "Trainable params: 1,257,702\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n",
            "None\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "eZGnRZ2S4uAd",
        "outputId": "0cbc3667-c36e-4bbd-9500-18ba39c74268"
      },
      "source": [
        "batch_size = 32\n",
        "history_3 = model_3.fit(X_train, y_train, epochs=8, batch_size=batch_size, verbose=1, validation_split=0.1, shuffle=True)"
      ],
      "execution_count": 32,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/8\n",
            "40/40 [==============================] - 2s 40ms/step - loss: 0.5984 - accuracy: 0.6607 - val_loss: 0.3130 - val_accuracy: 0.9286\n",
            "Epoch 2/8\n",
            "40/40 [==============================] - 1s 36ms/step - loss: 0.3090 - accuracy: 0.9148 - val_loss: 0.2798 - val_accuracy: 0.9286\n",
            "Epoch 3/8\n",
            "40/40 [==============================] - 1s 36ms/step - loss: 0.3004 - accuracy: 0.9050 - val_loss: 0.2654 - val_accuracy: 0.9286\n",
            "Epoch 4/8\n",
            "40/40 [==============================] - 1s 35ms/step - loss: 0.2641 - accuracy: 0.9146 - val_loss: 0.2587 - val_accuracy: 0.9286\n",
            "Epoch 5/8\n",
            "40/40 [==============================] - 1s 35ms/step - loss: 0.2417 - accuracy: 0.9254 - val_loss: 0.2551 - val_accuracy: 0.9286\n",
            "Epoch 6/8\n",
            "40/40 [==============================] - 1s 35ms/step - loss: 0.2210 - accuracy: 0.9346 - val_loss: 0.2518 - val_accuracy: 0.9214\n",
            "Epoch 7/8\n",
            "40/40 [==============================] - 1s 35ms/step - loss: 0.1965 - accuracy: 0.9411 - val_loss: 0.2496 - val_accuracy: 0.9214\n",
            "Epoch 8/8\n",
            "40/40 [==============================] - 1s 35ms/step - loss: 0.1769 - accuracy: 0.9423 - val_loss: 0.2511 - val_accuracy: 0.9214\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 573
        },
        "id": "lMxChr4M44m9",
        "outputId": "414e5c57-9c1e-4cab-aa2b-a9a66e6a21aa"
      },
      "source": [
        "plt.plot(history_3.history['accuracy'])\n",
        "plt.plot(history_3.history['val_accuracy'])\n",
        "plt.title('model accuracy')\n",
        "plt.ylabel('accuracy')\n",
        "plt.xlabel('epoch')\n",
        "plt.legend(['train', 'validation'], loc='upper left')\n",
        "plt.show()\n",
        "\n",
        "plt.plot(history_3.history['loss'])\n",
        "plt.plot(history_3.history['val_loss'])\n",
        "plt.title('model loss')\n",
        "plt.ylabel('loss')\n",
        "plt.xlabel('epoch')\n",
        "plt.legend(['train', 'validation'], loc='upper left')\n",
        "plt.show()"
      ],
      "execution_count": 33,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8sZsjIOvlLIK"
      },
      "source": [
        "## Helper Function\n",
        "\n",
        "from sklearn.metrics import accuracy_score, confusion_matrix, classification_report\n",
        "from sklearn.metrics import precision_score\n",
        "from sklearn.metrics import recall_score\n",
        "from sklearn.metrics import accuracy_score\n",
        "import pickle\n",
        "\n",
        "def generate_report(pred_list, custom_model, model_name,y_test):\n",
        "  print(f'Confusion Matrix:\\n{confusion_matrix(y_test, pred_list)}')\n",
        "  print(f'\\nClassification Report:\\n{classification_report(y_test,pred_list)}')\n",
        "  acc = accuracy_score(y_test, pred_list) * 100\n",
        "  print(f'Accuracy: {acc}%')\n",
        "  prec = precision_score(y_test, pred_list, average='macro') * 100\n",
        "  print(f'Precision: {prec}%')\n",
        "  rec = recall_score(y_test, pred_list, average='macro') * 100\n",
        "  print(f'Recall: {rec}%')\n",
        "  # pickle.dump(custom_model, open(model_name + '.sav', 'wb'))\n",
        "  print(model_name + ' trained')\n",
        "  return acc, prec, rec"
      ],
      "execution_count": 34,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qgce5YRO5MJV"
      },
      "source": [
        "y_hat_3 = model_3.predict(X_test)"
      ],
      "execution_count": 35,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "TrCXzbU4Ikve",
        "outputId": "fc24e441-c415-44d0-891f-a7880909f9a9"
      },
      "source": [
        "cnn_metrics = [None, None, None]\n",
        "cnn_metrics[0], cnn_metrics[1], cnn_metrics[2] = generate_report(y_hat_3.argmax(axis=1), model_3, 'CNN_model',y_test.argmax(axis=1))"
      ],
      "execution_count": 41,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Confusion Matrix:\n",
            "[[297   8]\n",
            " [ 33 262]]\n",
            "\n",
            "Classification Report:\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.90      0.97      0.94       305\n",
            "           1       0.97      0.89      0.93       295\n",
            "\n",
            "    accuracy                           0.93       600\n",
            "   macro avg       0.94      0.93      0.93       600\n",
            "weighted avg       0.93      0.93      0.93       600\n",
            "\n",
            "Accuracy: 93.16666666666666%\n",
            "Precision: 93.5185185185185%\n",
            "Recall: 93.09530425118089%\n",
            "CNN_model trained\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "B6j-jLzr5rMm"
      },
      "source": [
        "### LSTM"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "P7-u5X1MC1U5"
      },
      "source": [
        "from keras.callbacks import EarlyStopping\n",
        "from keras.models import Sequential\n",
        "from keras.layers import Embedding, LSTM, Dense, Dropout\n",
        "from keras.optimizers import Adam"
      ],
      "execution_count": 42,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "_HixxGgi51Pt",
        "outputId": "f08e971a-505b-474f-8bd8-a3ef5764acb6"
      },
      "source": [
        "model = Sequential()\n",
        "\n",
        "model.add(Embedding(num_words,embedding_dim,\n",
        "                              embeddings_initializer=Constant(embedding_matrix),\n",
        "                            input_length=sequence_length,\n",
        "                            trainable=True))\n",
        "model.add(LSTM(256, dropout=0.5, recurrent_dropout=0.5, return_sequences=True))\n",
        "model.add(LSTM(128, dropout=0.5, recurrent_dropout=0.5, return_sequences=False))\n",
        "model.add(Dense(100, activation='relu'))\n",
        "model.add(Dropout(0.4))\n",
        "model.add(Dense(2, activation='softmax'))\n",
        "model.compile(loss='categorical_crossentropy', optimizer=Adam(lr=0.005), metrics=['accuracy'])\n",
        "model.summary()"
      ],
      "execution_count": 43,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "embedding_2 (Embedding)      (None, 260, 50)           1250050   \n",
            "_________________________________________________________________\n",
            "lstm (LSTM)                  (None, 260, 256)          314368    \n",
            "_________________________________________________________________\n",
            "lstm_1 (LSTM)                (None, 128)               197120    \n",
            "_________________________________________________________________\n",
            "dense_2 (Dense)              (None, 100)               12900     \n",
            "_________________________________________________________________\n",
            "dropout_2 (Dropout)          (None, 100)               0         \n",
            "_________________________________________________________________\n",
            "dense_3 (Dense)              (None, 2)                 202       \n",
            "=================================================================\n",
            "Total params: 1,774,640\n",
            "Trainable params: 1,774,640\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "7rLkZkVU541x",
        "outputId": "f15fc823-b545-497c-ba47-5a71975c3650"
      },
      "source": [
        "history_5 = model.fit(X_train, y_train, epochs=7, batch_size=256, verbose=1, validation_split=0.1, shuffle=True)"
      ],
      "execution_count": 44,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/7\n",
            "5/5 [==============================] - 47s 8s/step - loss: 0.7500 - accuracy: 0.5221 - val_loss: 0.6797 - val_accuracy: 0.5214\n",
            "Epoch 2/7\n",
            "5/5 [==============================] - 39s 8s/step - loss: 0.6791 - accuracy: 0.5781 - val_loss: 0.5692 - val_accuracy: 0.7143\n",
            "Epoch 3/7\n",
            "5/5 [==============================] - 39s 8s/step - loss: 0.7337 - accuracy: 0.7114 - val_loss: 0.7593 - val_accuracy: 0.4857\n",
            "Epoch 4/7\n",
            "5/5 [==============================] - 39s 8s/step - loss: 0.6907 - accuracy: 0.5408 - val_loss: 0.6711 - val_accuracy: 0.5500\n",
            "Epoch 5/7\n",
            "5/5 [==============================] - 39s 8s/step - loss: 0.6785 - accuracy: 0.5361 - val_loss: 0.6831 - val_accuracy: 0.4857\n",
            "Epoch 6/7\n",
            "5/5 [==============================] - 38s 8s/step - loss: 0.6772 - accuracy: 0.5398 - val_loss: 0.6746 - val_accuracy: 0.5357\n",
            "Epoch 7/7\n",
            "5/5 [==============================] - 39s 8s/step - loss: 0.6807 - accuracy: 0.5033 - val_loss: 0.6723 - val_accuracy: 0.6429\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 573
        },
        "id": "5o0z2dda6CDn",
        "outputId": "dbf9b764-722e-411f-c3f7-b09567c14cd4"
      },
      "source": [
        "plt.plot(history_5.history['accuracy'])\n",
        "plt.plot(history_5.history['val_accuracy'])\n",
        "plt.title('model accuracy')\n",
        "plt.ylabel('accuracy')\n",
        "plt.xlabel('epoch')\n",
        "plt.legend(['train', 'validation'], loc='upper left')\n",
        "plt.show()\n",
        "\n",
        "plt.plot(history_5.history['loss'])\n",
        "plt.plot(history_5.history['val_loss'])\n",
        "plt.title('model loss')\n",
        "plt.ylabel('loss')\n",
        "plt.xlabel('epoch')\n",
        "plt.legend(['train', 'validation'], loc='upper left')\n",
        "plt.show()"
      ],
      "execution_count": 45,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rrLPLk9S6Jkf"
      },
      "source": [
        "y_hat_5 = model.predict(X_test)"
      ],
      "execution_count": 46,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "CeQ5qQ4RGyGh",
        "outputId": "ece69628-56a4-4a26-d2b8-b6e64cea4a01"
      },
      "source": [
        "lstm_metrics = [None, None, None]\n",
        "lstm_metrics[0], lstm_metrics[1], lstm_metrics[2] = generate_report(y_hat_5.argmax(axis=1), model, 'LSTM_model',y_test.argmax(axis=1))"
      ],
      "execution_count": 49,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Confusion Matrix:\n",
            "[[164 141]\n",
            " [ 75 220]]\n",
            "\n",
            "Classification Report:\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.69      0.54      0.60       305\n",
            "           1       0.61      0.75      0.67       295\n",
            "\n",
            "    accuracy                           0.64       600\n",
            "   macro avg       0.65      0.64      0.64       600\n",
            "weighted avg       0.65      0.64      0.64       600\n",
            "\n",
            "Accuracy: 64.0%\n",
            "Precision: 64.78053755838616%\n",
            "Recall: 64.17338149485968%\n",
            "LSTM_model trained\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "zSZr_eFOlLIM"
      },
      "source": [
        "### Decision Tree Classifier"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "LJalYXkmlLIN"
      },
      "source": [
        "from sklearn.tree import DecisionTreeClassifier\n",
        "from sklearn.model_selection import cross_val_score"
      ],
      "execution_count": 50,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jFtwVxwilLIN"
      },
      "source": [
        "dt_clf = DecisionTreeClassifier()\n",
        "\n",
        "# Train the Model\n",
        "dt_clf.fit(X_train,y_train)   \n",
        "dt_pred = dt_clf.predict(X_test)"
      ],
      "execution_count": 51,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "e_NrShfwlLIO",
        "outputId": "73f8dd21-b4b9-40b2-e5e4-f567bc176e33"
      },
      "source": [
        "dt_metrics = [None, None, None]\n",
        "dt_metrics[0], dt_metrics[1], dt_metrics[2] = generate_report(dt_pred.argmax(axis=1), dt_clf, 'DecisionTree_model',y_test.argmax(axis=1))"
      ],
      "execution_count": 52,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Confusion Matrix:\n",
            "[[213  92]\n",
            " [ 84 211]]\n",
            "\n",
            "Classification Report:\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.72      0.70      0.71       305\n",
            "           1       0.70      0.72      0.71       295\n",
            "\n",
            "    accuracy                           0.71       600\n",
            "   macro avg       0.71      0.71      0.71       600\n",
            "weighted avg       0.71      0.71      0.71       600\n",
            "\n",
            "Accuracy: 70.66666666666667%\n",
            "Precision: 70.67706770677067%\n",
            "Recall: 70.68074465129203%\n",
            "DecisionTree_model trained\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "_jnxM8eUlLIQ"
      },
      "source": [
        "### Bernoulli Naive Bayes:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "bBeeRlTtlLIR"
      },
      "source": [
        "from sklearn.naive_bayes import BernoulliNB\n",
        "bnb_clf = BernoulliNB()\n",
        "bnb_clf.fit(X_train,y_train.argmax(axis=1))\n",
        "BNBpred=bnb_clf.predict(X_test)"
      ],
      "execution_count": 53,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "K9KNBoKGlLIR",
        "outputId": "1871baef-940d-4e9d-d6be-02d024fcd3a9"
      },
      "source": [
        "bnb_metrics = [None, None, None]\n",
        "bnb_metrics[0], bnb_metrics[1], bnb_metrics[2] = generate_report(BNBpred, bnb_clf, 'BernoulliNB_model',y_test.argmax(axis=1))"
      ],
      "execution_count": 54,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Confusion Matrix:\n",
            "[[165 140]\n",
            " [154 141]]\n",
            "\n",
            "Classification Report:\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.52      0.54      0.53       305\n",
            "           1       0.50      0.48      0.49       295\n",
            "\n",
            "    accuracy                           0.51       600\n",
            "   macro avg       0.51      0.51      0.51       600\n",
            "weighted avg       0.51      0.51      0.51       600\n",
            "\n",
            "Accuracy: 51.0%\n",
            "Precision: 50.9510369370475%\n",
            "Recall: 50.947485412614625%\n",
            "BernoulliNB_model trained\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Kv7sddtolLIS"
      },
      "source": [
        "### Multinomial Naive Bayes:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "pSmc7Ot0lLIS"
      },
      "source": [
        "from sklearn.naive_bayes import MultinomialNB"
      ],
      "execution_count": 55,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zfQtE8iJlLIT"
      },
      "source": [
        "mnb_clf = MultinomialNB()\n",
        "\n",
        "#train the model\n",
        "mnb_clf.fit(X_train,y_train.argmax(axis=1))\n",
        "MNBpred=mnb_clf.predict(X_test)"
      ],
      "execution_count": 56,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ORbwC8R6lLIT",
        "outputId": "e06d5bb0-ba09-4c76-bb4b-12debe9a664b"
      },
      "source": [
        "mnb_metrics = [None, None, None]\n",
        "mnb_metrics[0], mnb_metrics[1], mnb_metrics[2] = generate_report(MNBpred, mnb_clf, 'MultinomialNB_model',y_test.argmax(axis=1))"
      ],
      "execution_count": 57,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Confusion Matrix:\n",
            "[[145 160]\n",
            " [127 168]]\n",
            "\n",
            "Classification Report:\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.53      0.48      0.50       305\n",
            "           1       0.51      0.57      0.54       295\n",
            "\n",
            "    accuracy                           0.52       600\n",
            "   macro avg       0.52      0.52      0.52       600\n",
            "weighted avg       0.52      0.52      0.52       600\n",
            "\n",
            "Accuracy: 52.166666666666664%\n",
            "Precision: 52.264167862266866%\n",
            "Recall: 52.24506807446513%\n",
            "MultinomialNB_model trained\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "8c2qtQK5lLIU"
      },
      "source": [
        "### Maximum Entropy (Logistic Regression):"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jbIZ2FXPlLIU"
      },
      "source": [
        "from sklearn.linear_model import LogisticRegression"
      ],
      "execution_count": 58,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Qj5yMj2AlLIV",
        "outputId": "bfd1bd18-9f2d-45be-cc29-135afe9ab060"
      },
      "source": [
        "lr_clf = LogisticRegression(max_iter=300)\n",
        "\n",
        "# Train the Model\n",
        "lr_clf.fit(X_train,y_train.argmax(axis=1))   \n",
        "LRpred = lr_clf.predict(X_test)\n",
        "\n",
        "lr_metrics = [None, None, None]\n",
        "lr_metrics[0], lr_metrics[1], lr_metrics[2] = generate_report(LRpred, lr_clf, 'LinearRegression_model',y_test.argmax(axis=1))"
      ],
      "execution_count": 59,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Confusion Matrix:\n",
            "[[204 101]\n",
            " [106 189]]\n",
            "\n",
            "Classification Report:\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.66      0.67      0.66       305\n",
            "           1       0.65      0.64      0.65       295\n",
            "\n",
            "    accuracy                           0.66       600\n",
            "   macro avg       0.65      0.65      0.65       600\n",
            "weighted avg       0.65      0.66      0.65       600\n",
            "\n",
            "Accuracy: 65.5%\n",
            "Precision: 65.48943270300333%\n",
            "Recall: 65.47652125590442%\n",
            "LinearRegression_model trained\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "FB_VXXhOlLIV"
      },
      "source": [
        "### Support Vector Machine:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "BjgtQ6BnlLIW"
      },
      "source": [
        "from sklearn.svm import SVC\n",
        "from sklearn import svm"
      ],
      "execution_count": 60,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "uPkyVvtrlLIW"
      },
      "source": [
        "SVMclassifier = svm.SVC()\n",
        "SVMclassifier.fit(X_train,y_train.argmax(axis=1))\n",
        "SVMpred = SVMclassifier.predict(X_test)"
      ],
      "execution_count": 61,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "TTl-UcNPlLIX",
        "outputId": "73a8b121-dcb3-4dbc-c885-1885140dcaab"
      },
      "source": [
        "svm_metrics = [None, None, None]\n",
        "svm_metrics[0], svm_metrics[1], svm_metrics[2] = generate_report(SVMpred, SVMclassifier, 'SVM_model',y_test.argmax(axis=1))"
      ],
      "execution_count": 62,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Confusion Matrix:\n",
            "[[264  41]\n",
            " [ 76 219]]\n",
            "\n",
            "Classification Report:\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.78      0.87      0.82       305\n",
            "           1       0.84      0.74      0.79       295\n",
            "\n",
            "    accuracy                           0.81       600\n",
            "   macro avg       0.81      0.80      0.80       600\n",
            "weighted avg       0.81      0.81      0.80       600\n",
            "\n",
            "Accuracy: 80.5%\n",
            "Precision: 80.93891402714932%\n",
            "Recall: 80.39733259238677%\n",
            "SVM_model trained\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "zJdO5kQzJSoR"
      },
      "source": [
        "## Results Comparisons plots and tables:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "WFqAg2OkJZQ6",
        "outputId": "6ac186fb-b326-4f27-9adc-c9551b67984d"
      },
      "source": [
        "!pip3 install tabulate\n",
        "from tabulate import tabulate\n",
        "table = [['Model Name', 'Accuracy', 'Precision', 'Recall'], \n",
        "         ['Multinomial Naïve Bayes', mnb_metrics[0], mnb_metrics[1], mnb_metrics[2]], \n",
        "         ['Bernoulli Naïve Bayes', bnb_metrics[0], bnb_metrics[1], bnb_metrics[2]],\n",
        "         ['Support Vector Machine', svm_metrics[0], svm_metrics[1], svm_metrics[2]],\n",
        "         ['Maximum Entropy', lr_metrics[0], lr_metrics[1], lr_metrics[2]],\n",
        "         ['Decision Tree', dt_metrics[0], dt_metrics[1], dt_metrics[2]],\n",
        "         ['CNN', cnn_metrics[0], cnn_metrics[1], cnn_metrics[2]],\n",
        "         ['LSTM', lstm_metrics[0], lstm_metrics[1], lstm_metrics[2]]]\n",
        "\n",
        "print(tabulate(table, headers='firstrow', tablefmt='fancy_grid'))"
      ],
      "execution_count": 63,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Requirement already satisfied: tabulate in /usr/local/lib/python3.7/dist-packages (0.8.9)\n",
            "╒═════════════════════════╤════════════╤═════════════╤══════════╕\n",
            "│ Model Name              │   Accuracy │   Precision │   Recall │\n",
            "╞═════════════════════════╪════════════╪═════════════╪══════════╡\n",
            "│ Multinomial Naïve Bayes │    52.1667 │     52.2642 │  52.2451 │\n",
            "├─────────────────────────┼────────────┼─────────────┼──────────┤\n",
            "│ Bernoulli Naïve Bayes   │    51      │     50.951  │  50.9475 │\n",
            "├─────────────────────────┼────────────┼─────────────┼──────────┤\n",
            "│ Support Vector Machine  │    80.5    │     80.9389 │  80.3973 │\n",
            "├─────────────────────────┼────────────┼─────────────┼──────────┤\n",
            "│ Maximum Entropy         │    65.5    │     65.4894 │  65.4765 │\n",
            "├─────────────────────────┼────────────┼─────────────┼──────────┤\n",
            "│ Decision Tree           │    70.6667 │     70.6771 │  70.6807 │\n",
            "├─────────────────────────┼────────────┼─────────────┼──────────┤\n",
            "│ CNN                     │    93.1667 │     93.5185 │  93.0953 │\n",
            "├─────────────────────────┼────────────┼─────────────┼──────────┤\n",
            "│ LSTM                    │    64      │     64.7805 │  64.1734 │\n",
            "╘═════════════════════════╧════════════╧═════════════╧══════════╛\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 346
        },
        "id": "vsqBbaAHJ77L",
        "outputId": "79798e3b-b924-4773-cde3-7a0aa9a2e72d"
      },
      "source": [
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "  \n",
        "N = 7\n",
        "ind = np.arange(N) \n",
        "width = 0.25\n",
        "  \n",
        "acc = [mnb_metrics[0], bnb_metrics[0], svm_metrics[0], lr_metrics[0], dt_metrics[0], cnn_metrics[0], lstm_metrics[0]]\n",
        "bar1 = plt.bar(ind, acc, width)\n",
        "  \n",
        "prec = [mnb_metrics[1], bnb_metrics[1], svm_metrics[1], lr_metrics[1], dt_metrics[1], cnn_metrics[1], lstm_metrics[1]]\n",
        "bar2 = plt.bar(ind+width, prec, width)\n",
        "  \n",
        "recall = [mnb_metrics[2], bnb_metrics[2], svm_metrics[2], lr_metrics[2], dt_metrics[2], cnn_metrics[2], lstm_metrics[2]]\n",
        "bar3 = plt.bar(ind+width*2, recall, width)\n",
        "  \n",
        "# plt.xlabel(\"Dates\")\n",
        "plt.ylabel(\"Percentage\")\n",
        "plt.title(\"Comparison\")\n",
        "  \n",
        "plt.xticks(ind+width,['Multinomial NB','Bernoulli NB','SVM','Logistic Regression', 'Decision Tree', 'CNN', 'LSTM'])\n",
        "plt.xticks(rotation = 45)\n",
        "plt.legend( (bar1, bar2, bar3), ('Accuracy', 'Precision', 'Recall'), bbox_to_anchor = (1.05, 0.6))\n",
        "plt.show()"
      ],
      "execution_count": 64,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 346
        },
        "id": "9jiwMcrjKRpM",
        "outputId": "67c90d1f-6f19-499f-f6cb-0f6a19c0e38c"
      },
      "source": [
        "import matplotlib.pyplot as plt\n",
        "models = ['Multinomial NB','Bernoulli NB','SVM','Logistic Regression', 'Decision Tree', 'CNN', 'LSTM']\n",
        "recall = [mnb_metrics[2], bnb_metrics[2], svm_metrics[2], lr_metrics[2], dt_metrics[2], cnn_metrics[2], lstm_metrics[2]]\n",
        "precision = [mnb_metrics[1], bnb_metrics[1], svm_metrics[1], lr_metrics[1], dt_metrics[1], cnn_metrics[1], lstm_metrics[1]]\n",
        "line_chart1 = plt.plot(models, recall, 'o-g')\n",
        "line_chart2 = plt.plot(models, precision, 'o-y')\n",
        "plt.title('Recall vs Precision')\n",
        "plt.ylabel('Score')\n",
        "plt.xticks(rotation = 45)\n",
        "plt.legend(['Recall', 'Precision'], bbox_to_anchor = (1.3, 0.6))\n",
        "plt.show()"
      ],
      "execution_count": 65,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "UQlsE0nUkuDZ"
      },
      "source": [
        ""
      ],
      "execution_count": 65,
      "outputs": []
    }
  ]
}