BBC_News_classification_with_NN.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "stopped-scale",
   "metadata": {},
   "source": [
    "# BBC News Classification (TF)\n",
    "\n",
    "We will implement a news classification based on [BBC News Classification Dataset](https://www.kaggle.com/c/learn-ai-bbc/overview), which contains 2225 examples of news articles with their respective labels. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "magnetic-rebate",
   "metadata": {
    "id": "gnwiOnGyW5JK",
    "tags": [
     "graded"
    ]
   },
   "outputs": [],
   "source": [
    "import os\n",
    "import io\n",
    "import csv\n",
    "import pandas as pd\n",
    "import tensorflow as tf\n",
    "import numpy as np\n",
    "from tensorflow.keras.preprocessing.text import Tokenizer\n",
    "from tensorflow.keras.preprocessing.sequence import pad_sequences\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.model_selection import train_test_split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "79acd69d",
   "metadata": {},
   "outputs": [],
   "source": [
    "TRAIN_PATH = os.path.join(\"../datasets/bbc-news\", \"BBC News Train.csv\")\n",
    "TEST_PATH = os.path.join(\"../datasets/bbc-news\", \"BBC News Test.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "6d74fbcf",
   "metadata": {},
   "outputs": [
    {
     "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>ArticleId</th>\n",
       "      <th>Text</th>\n",
       "      <th>Category</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1833</td>\n",
       "      <td>worldcom ex-boss launches defence lawyers defe...</td>\n",
       "      <td>business</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>154</td>\n",
       "      <td>german business confidence slides german busin...</td>\n",
       "      <td>business</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1101</td>\n",
       "      <td>bbc poll indicates economic gloom citizens in ...</td>\n",
       "      <td>business</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1976</td>\n",
       "      <td>lifestyle  governs mobile choice  faster  bett...</td>\n",
       "      <td>tech</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>917</td>\n",
       "      <td>enron bosses in $168m payout eighteen former e...</td>\n",
       "      <td>business</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   ArticleId                                               Text  Category\n",
       "0       1833  worldcom ex-boss launches defence lawyers defe...  business\n",
       "1        154  german business confidence slides german busin...  business\n",
       "2       1101  bbc poll indicates economic gloom citizens in ...  business\n",
       "3       1976  lifestyle  governs mobile choice  faster  bett...      tech\n",
       "4        917  enron bosses in $168m payout eighteen former e...  business"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = pd.read_csv(TRAIN_PATH)\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "8724e949",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, 'Examples')"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "df.Category.value_counts().sort_values().plot(kind = 'bar')\n",
    "plt.title(\"BBC News dataset\")\n",
    "plt.xlabel(\"Category\")\n",
    "plt.ylabel(\"Examples\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "balanced-settle",
   "metadata": {},
   "source": [
    "## Defining useful global variables\n",
    "\n",
    "- `NUM_WORDS`: The maximum number of words to keep, based on word frequency. Defaults to 1000.\n",
    "\n",
    "\n",
    "- `EMBEDDING_DIM`: Dimension of the dense embedding, will be used in the embedding layer of the model. Defaults to 16.\n",
    "\n",
    "\n",
    "- `MAXLEN`: Maximum length of all sequences. Defaults to 120.\n",
    "\n",
    "\n",
    "- `PADDING`: Padding strategy (pad either before or after each sequence.). Defaults to 'post'.\n",
    "\n",
    "\n",
    "- `OOV_TOKEN`: Token to replace out-of-vocabulary words during text_to_sequence calls. Defaults to \"\\<OOV>\".\n",
    "\n",
    "    \n",
    "- `TRAINING_SPLIT`: Proportion of data used for training. Defaults to 0.8"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "quantitative-mauritius",
   "metadata": {
    "tags": [
     "graded"
    ]
   },
   "outputs": [],
   "source": [
    "NUM_WORDS = 1000\n",
    "EMBEDDING_DIM = 32\n",
    "MAXLEN = 120\n",
    "PADDING = 'post'\n",
    "OOV_TOKEN = \"<OOV>\"\n",
    "TRAINING_SPLIT = .8"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "synthetic-beijing",
   "metadata": {},
   "source": [
    "## Loading and pre-processing the data\n",
    "\n",
    "Remove stopwords from text and to load the data from a csv file."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "flying-lincoln",
   "metadata": {
    "tags": [
     "graded"
    ]
   },
   "outputs": [],
   "source": [
    "def remove_stopwords(sentence):\n",
    "    \"\"\"\n",
    "    Removes a list of stopwords\n",
    "    \n",
    "    Args:\n",
    "        sentence (string): sentence to remove the stopwords from\n",
    "    \n",
    "    Returns:\n",
    "        sentence (string): lowercase sentence without the stopwords\n",
    "    \"\"\"\n",
    "    # List of stopwords\n",
    "    stopwords = [\"a\", \"about\", \"above\", \"after\", \"again\", \"against\", \"all\", \"am\", \"an\", \"and\", \"any\", \"are\", \"as\", \"at\", \"be\", \"because\", \"been\", \"before\", \"being\", \"below\", \"between\", \"both\", \"but\", \"by\", \"could\", \"did\", \"do\", \"does\", \"doing\", \"down\", \"during\", \"each\", \"few\", \"for\", \"from\", \"further\", \"had\", \"has\", \"have\", \"having\", \"he\", \"he'd\", \"he'll\", \"he's\", \"her\", \"here\", \"here's\", \"hers\", \"herself\", \"him\", \"himself\", \"his\", \"how\", \"how's\", \"i\", \"i'd\", \"i'll\", \"i'm\", \"i've\", \"if\", \"in\", \"into\", \"is\", \"it\", \"it's\", \"its\", \"itself\", \"let's\", \"me\", \"more\", \"most\", \"my\", \"myself\", \"nor\", \"of\", \"on\", \"once\", \"only\", \"or\", \"other\", \"ought\", \"our\", \"ours\", \"ourselves\", \"out\", \"over\", \"own\", \"same\", \"she\", \"she'd\", \"she'll\", \"she's\", \"should\", \"so\", \"some\", \"such\", \"than\", \"that\", \"that's\", \"the\", \"their\", \"theirs\", \"them\", \"themselves\", \"then\", \"there\", \"there's\", \"these\", \"they\", \"they'd\", \"they'll\", \"they're\", \"they've\", \"this\", \"those\", \"through\", \"to\", \"too\", \"under\", \"until\", \"up\", \"very\", \"was\", \"we\", \"we'd\", \"we'll\", \"we're\", \"we've\", \"were\", \"what\", \"what's\", \"when\", \"when's\", \"where\", \"where's\", \"which\", \"while\", \"who\", \"who's\", \"whom\", \"why\", \"why's\", \"with\", \"would\", \"you\", \"you'd\", \"you'll\", \"you're\", \"you've\", \"your\", \"yours\", \"yourself\", \"yourselves\" ]\n",
    "    \n",
    "    # Sentence converted to lowercase-only\n",
    "    sentence = sentence.lower()\n",
    "\n",
    "    words = sentence.split()\n",
    "    no_words = [w for w in words if w not in stopwords]\n",
    "    sentence = \" \".join(no_words)\n",
    "\n",
    "    return sentence"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "1284693a",
   "metadata": {},
   "outputs": [],
   "source": [
    "df['Text'] = df['Text'].map(lambda sentence: remove_stopwords(sentence))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "polished-eagle",
   "metadata": {},
   "source": [
    "## Training - Validation Split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "218b9351",
   "metadata": {},
   "outputs": [],
   "source": [
    "train_sentences, val_sentences, train_labels, val_labels = \\\n",
    "    train_test_split(df['Text'], df['Category'], test_size=1 - TRAINING_SPLIT, random_state=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "crude-mozambique",
   "metadata": {},
   "source": [
    "## Tokenization - Sequences and padding"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "recreational-prince",
   "metadata": {
    "lines_to_next_cell": 2,
    "tags": [
     "graded"
    ]
   },
   "outputs": [],
   "source": [
    "def fit_tokenizer(train_sentences, num_words, oov_token):\n",
    "    \n",
    "    tokenizer = Tokenizer(num_words=num_words, oov_token=oov_token)\n",
    "    tokenizer.fit_on_texts(train_sentences)\n",
    "    \n",
    "    return tokenizer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "great-trading",
   "metadata": {
    "tags": [
     "graded"
    ]
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Vocabulary contains 22651 words\n",
      "\n"
     ]
    }
   ],
   "source": [
    "tokenizer = fit_tokenizer(train_sentences, NUM_WORDS, OOV_TOKEN)\n",
    "word_index = tokenizer.word_index\n",
    "\n",
    "print(f\"Vocabulary contains {len(word_index)} words\\n\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "fourth-knight",
   "metadata": {
    "tags": [
     "graded"
    ]
   },
   "outputs": [],
   "source": [
    "def seq_and_pad(sentences, tokenizer, padding, maxlen):\n",
    "\n",
    "    sequences = tokenizer.texts_to_sequences(sentences)\n",
    "    padded_sequences = pad_sequences(sequences, maxlen=maxlen, padding=padding)\n",
    "    \n",
    "    return padded_sequences"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "separate-onion",
   "metadata": {
    "tags": [
     "graded"
    ]
   },
   "outputs": [],
   "source": [
    "train_padded_seq = seq_and_pad(train_sentences, tokenizer, PADDING, MAXLEN)\n",
    "val_padded_seq = seq_and_pad(val_sentences, tokenizer, PADDING, MAXLEN)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "id": "active-objective",
   "metadata": {
    "id": "XkWiQ_FKZNp2",
    "lines_to_next_cell": 2,
    "tags": [
     "graded"
    ]
   },
   "outputs": [],
   "source": [
    "def tokenize_labels(label_tokenizer, split_labels):\n",
    "    \"\"\"\n",
    "    Tokenizes the labels\n",
    "    \n",
    "    Args:\n",
    "        all_labels (list of string): labels to generate the word-index from\n",
    "        split_labels (list of string): labels to tokenize\n",
    "    \n",
    "    Returns:\n",
    "        label_seq_np (array of int): tokenized labels\n",
    "    \"\"\"\n",
    "    \n",
    "    # Convert labels to sequences\n",
    "    label_seq = label_tokenizer.texts_to_sequences(split_labels)\n",
    "    \n",
    "    # Convert sequences to a numpy array.\n",
    "    label_seq_np = np.array(label_seq) - 1\n",
    "    \n",
    "    return label_seq_np"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "id": "numerical-certificate",
   "metadata": {
    "tags": [
     "graded"
    ]
   },
   "outputs": [],
   "source": [
    "label_tokenizer = Tokenizer()\n",
    "# Fit the tokenizer on all the labels\n",
    "label_tokenizer.fit_on_texts(df['Category'])  \n",
    "   \n",
    "\n",
    "train_label_seq = tokenize_labels(label_tokenizer, train_labels)\n",
    "val_label_seq = tokenize_labels(label_tokenizer, val_labels)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "37713052",
   "metadata": {},
   "source": [
    "## Training the model for text classification\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "id": "little-bahrain",
   "metadata": {
    "id": "HZ5um4MWZP-W",
    "tags": [
     "graded"
    ]
   },
   "outputs": [],
   "source": [
    "def create_model(num_words, embedding_dim, maxlen):\n",
    "    \n",
    "    tf.random.set_seed(123)\n",
    "       \n",
    "    model = tf.keras.Sequential([ \n",
    "        tf.keras.layers.Embedding(num_words, embedding_dim, input_length=maxlen),\n",
    "        tf.keras.layers.Dropout(0.3),\n",
    "        tf.keras.layers.GlobalAveragePooling1D(),\n",
    "        tf.keras.layers.Dropout(0.3),                 \n",
    "        tf.keras.layers.Dense(24, activation='relu'),\n",
    "        tf.keras.layers.Dropout(0.3), \n",
    "        tf.keras.layers.Dense(5, activation='softmax')\n",
    "    ])\n",
    "    \n",
    "    model.compile(loss='sparse_categorical_crossentropy',\n",
    "                  optimizer='adam',\n",
    "                  metrics=['accuracy']) \n",
    "\n",
    "    return model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "id": "6a06c6d0",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_1\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " embedding_1 (Embedding)     (None, 120, 32)           32000     \n",
      "                                                                 \n",
      " dropout_3 (Dropout)         (None, 120, 32)           0         \n",
      "                                                                 \n",
      " global_average_pooling1d_1   (None, 32)               0         \n",
      " (GlobalAveragePooling1D)                                        \n",
      "                                                                 \n",
      " dropout_4 (Dropout)         (None, 32)                0         \n",
      "                                                                 \n",
      " dense_2 (Dense)             (None, 24)                792       \n",
      "                                                                 \n",
      " dropout_5 (Dropout)         (None, 24)                0         \n",
      "                                                                 \n",
      " dense_3 (Dense)             (None, 5)                 125       \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 32,917\n",
      "Trainable params: 32,917\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model = create_model(NUM_WORDS, EMBEDDING_DIM, MAXLEN)\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "id": "resident-productivity",
   "metadata": {
    "scrolled": true,
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/100\n",
      "38/38 [==============================] - 0s 4ms/step - loss: 1.6032 - accuracy: 0.2651 - val_loss: 1.5969 - val_accuracy: 0.3087\n",
      "Epoch 2/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 1.5829 - accuracy: 0.3607 - val_loss: 1.5704 - val_accuracy: 0.4228\n",
      "Epoch 3/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 1.5403 - accuracy: 0.4396 - val_loss: 1.5210 - val_accuracy: 0.4497\n",
      "Epoch 4/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 1.4703 - accuracy: 0.4958 - val_loss: 1.4304 - val_accuracy: 0.5067\n",
      "Epoch 5/100\n",
      "38/38 [==============================] - 0s 3ms/step - loss: 1.3509 - accuracy: 0.5747 - val_loss: 1.2994 - val_accuracy: 0.5638\n",
      "Epoch 6/100\n",
      "38/38 [==============================] - 0s 3ms/step - loss: 1.2115 - accuracy: 0.6275 - val_loss: 1.1447 - val_accuracy: 0.6275\n",
      "Epoch 7/100\n",
      "38/38 [==============================] - 0s 3ms/step - loss: 1.0531 - accuracy: 0.6837 - val_loss: 0.9929 - val_accuracy: 0.7282\n",
      "Epoch 8/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.8917 - accuracy: 0.7836 - val_loss: 0.8506 - val_accuracy: 0.8289\n",
      "Epoch 9/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.7883 - accuracy: 0.8205 - val_loss: 0.7276 - val_accuracy: 0.8691\n",
      "Epoch 10/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.6882 - accuracy: 0.8398 - val_loss: 0.6256 - val_accuracy: 0.9094\n",
      "Epoch 11/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.5904 - accuracy: 0.8725 - val_loss: 0.5443 - val_accuracy: 0.9161\n",
      "Epoch 12/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.5151 - accuracy: 0.8951 - val_loss: 0.4790 - val_accuracy: 0.9195\n",
      "Epoch 13/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.4628 - accuracy: 0.8968 - val_loss: 0.4288 - val_accuracy: 0.9262\n",
      "Epoch 14/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.4178 - accuracy: 0.9027 - val_loss: 0.3828 - val_accuracy: 0.9295\n",
      "Epoch 15/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.3623 - accuracy: 0.9144 - val_loss: 0.3510 - val_accuracy: 0.9295\n",
      "Epoch 16/100\n",
      "38/38 [==============================] - 0s 3ms/step - loss: 0.3170 - accuracy: 0.9362 - val_loss: 0.3205 - val_accuracy: 0.9396\n",
      "Epoch 17/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.2890 - accuracy: 0.9371 - val_loss: 0.2995 - val_accuracy: 0.9430\n",
      "Epoch 18/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.2714 - accuracy: 0.9413 - val_loss: 0.2819 - val_accuracy: 0.9463\n",
      "Epoch 19/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.2564 - accuracy: 0.9455 - val_loss: 0.2695 - val_accuracy: 0.9463\n",
      "Epoch 20/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.2267 - accuracy: 0.9530 - val_loss: 0.2558 - val_accuracy: 0.9430\n",
      "Epoch 21/100\n",
      "38/38 [==============================] - 0s 3ms/step - loss: 0.2195 - accuracy: 0.9471 - val_loss: 0.2441 - val_accuracy: 0.9463\n",
      "Epoch 22/100\n",
      "38/38 [==============================] - 0s 3ms/step - loss: 0.1942 - accuracy: 0.9631 - val_loss: 0.2369 - val_accuracy: 0.9497\n",
      "Epoch 23/100\n",
      "38/38 [==============================] - 0s 3ms/step - loss: 0.1815 - accuracy: 0.9589 - val_loss: 0.2320 - val_accuracy: 0.9396\n",
      "Epoch 24/100\n",
      "38/38 [==============================] - 0s 3ms/step - loss: 0.1727 - accuracy: 0.9597 - val_loss: 0.2236 - val_accuracy: 0.9463\n",
      "Epoch 25/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.1612 - accuracy: 0.9698 - val_loss: 0.2202 - val_accuracy: 0.9497\n",
      "Epoch 26/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.1450 - accuracy: 0.9698 - val_loss: 0.2166 - val_accuracy: 0.9463\n",
      "Epoch 27/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.1478 - accuracy: 0.9698 - val_loss: 0.2111 - val_accuracy: 0.9497\n",
      "Epoch 28/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.1285 - accuracy: 0.9748 - val_loss: 0.2083 - val_accuracy: 0.9497\n",
      "Epoch 29/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.1184 - accuracy: 0.9773 - val_loss: 0.2052 - val_accuracy: 0.9463\n",
      "Epoch 30/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.1099 - accuracy: 0.9807 - val_loss: 0.2040 - val_accuracy: 0.9497\n",
      "Epoch 31/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.1197 - accuracy: 0.9824 - val_loss: 0.2008 - val_accuracy: 0.9497\n",
      "Epoch 32/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.1186 - accuracy: 0.9790 - val_loss: 0.1994 - val_accuracy: 0.9497\n",
      "Epoch 33/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.1051 - accuracy: 0.9832 - val_loss: 0.2002 - val_accuracy: 0.9463\n",
      "Epoch 34/100\n",
      "38/38 [==============================] - 0s 3ms/step - loss: 0.1010 - accuracy: 0.9807 - val_loss: 0.1981 - val_accuracy: 0.9497\n",
      "Epoch 35/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0912 - accuracy: 0.9841 - val_loss: 0.1961 - val_accuracy: 0.9463\n",
      "Epoch 36/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0838 - accuracy: 0.9841 - val_loss: 0.1942 - val_accuracy: 0.9497\n",
      "Epoch 37/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0789 - accuracy: 0.9866 - val_loss: 0.1955 - val_accuracy: 0.9463\n",
      "Epoch 38/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0708 - accuracy: 0.9883 - val_loss: 0.1928 - val_accuracy: 0.9463\n",
      "Epoch 39/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0748 - accuracy: 0.9866 - val_loss: 0.1972 - val_accuracy: 0.9430\n",
      "Epoch 40/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0610 - accuracy: 0.9958 - val_loss: 0.1948 - val_accuracy: 0.9430\n",
      "Epoch 41/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0698 - accuracy: 0.9891 - val_loss: 0.1937 - val_accuracy: 0.9463\n",
      "Epoch 42/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0596 - accuracy: 0.9883 - val_loss: 0.1947 - val_accuracy: 0.9430\n",
      "Epoch 43/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0619 - accuracy: 0.9916 - val_loss: 0.1965 - val_accuracy: 0.9430\n",
      "Epoch 44/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0553 - accuracy: 0.9916 - val_loss: 0.1950 - val_accuracy: 0.9396\n",
      "Epoch 45/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0510 - accuracy: 0.9933 - val_loss: 0.1964 - val_accuracy: 0.9463\n",
      "Epoch 46/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0503 - accuracy: 0.9958 - val_loss: 0.1942 - val_accuracy: 0.9463\n",
      "Epoch 47/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0406 - accuracy: 0.9958 - val_loss: 0.1969 - val_accuracy: 0.9430\n",
      "Epoch 48/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0433 - accuracy: 0.9950 - val_loss: 0.1973 - val_accuracy: 0.9430\n",
      "Epoch 49/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0461 - accuracy: 0.9908 - val_loss: 0.1969 - val_accuracy: 0.9463\n",
      "Epoch 50/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0411 - accuracy: 0.9933 - val_loss: 0.1976 - val_accuracy: 0.9430\n",
      "Epoch 51/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0483 - accuracy: 0.9924 - val_loss: 0.1996 - val_accuracy: 0.9430\n",
      "Epoch 52/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0443 - accuracy: 0.9908 - val_loss: 0.2000 - val_accuracy: 0.9430\n",
      "Epoch 53/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0385 - accuracy: 0.9958 - val_loss: 0.2023 - val_accuracy: 0.9430\n",
      "Epoch 54/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0412 - accuracy: 0.9966 - val_loss: 0.2052 - val_accuracy: 0.9396\n",
      "Epoch 55/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0361 - accuracy: 0.9950 - val_loss: 0.2013 - val_accuracy: 0.9430\n",
      "Epoch 56/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0307 - accuracy: 0.9992 - val_loss: 0.2042 - val_accuracy: 0.9430\n",
      "Epoch 57/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0365 - accuracy: 0.9916 - val_loss: 0.2046 - val_accuracy: 0.9430\n",
      "Epoch 58/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0365 - accuracy: 0.9975 - val_loss: 0.2032 - val_accuracy: 0.9430\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 59/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0333 - accuracy: 0.9966 - val_loss: 0.2059 - val_accuracy: 0.9430\n",
      "Epoch 60/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0339 - accuracy: 0.9916 - val_loss: 0.2100 - val_accuracy: 0.9396\n",
      "Epoch 61/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0350 - accuracy: 0.9933 - val_loss: 0.2076 - val_accuracy: 0.9396\n",
      "Epoch 62/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0332 - accuracy: 0.9950 - val_loss: 0.2083 - val_accuracy: 0.9430\n",
      "Epoch 63/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0330 - accuracy: 0.9941 - val_loss: 0.2116 - val_accuracy: 0.9396\n",
      "Epoch 64/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0305 - accuracy: 0.9933 - val_loss: 0.2105 - val_accuracy: 0.9430\n",
      "Epoch 65/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0256 - accuracy: 0.9966 - val_loss: 0.2136 - val_accuracy: 0.9430\n",
      "Epoch 66/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0246 - accuracy: 0.9975 - val_loss: 0.2122 - val_accuracy: 0.9396\n",
      "Epoch 67/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0291 - accuracy: 0.9941 - val_loss: 0.2194 - val_accuracy: 0.9396\n",
      "Epoch 68/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0312 - accuracy: 0.9950 - val_loss: 0.2154 - val_accuracy: 0.9396\n",
      "Epoch 69/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0228 - accuracy: 0.9975 - val_loss: 0.2156 - val_accuracy: 0.9430\n",
      "Epoch 70/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0251 - accuracy: 0.9975 - val_loss: 0.2186 - val_accuracy: 0.9430\n",
      "Epoch 71/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0233 - accuracy: 0.9975 - val_loss: 0.2230 - val_accuracy: 0.9430\n",
      "Epoch 72/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0207 - accuracy: 0.9950 - val_loss: 0.2197 - val_accuracy: 0.9396\n",
      "Epoch 73/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0259 - accuracy: 0.9958 - val_loss: 0.2212 - val_accuracy: 0.9430\n",
      "Epoch 74/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0195 - accuracy: 0.9966 - val_loss: 0.2260 - val_accuracy: 0.9396\n",
      "Epoch 75/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0272 - accuracy: 0.9958 - val_loss: 0.2239 - val_accuracy: 0.9362\n",
      "Epoch 76/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0241 - accuracy: 0.9975 - val_loss: 0.2237 - val_accuracy: 0.9362\n",
      "Epoch 77/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0221 - accuracy: 0.9975 - val_loss: 0.2231 - val_accuracy: 0.9430\n",
      "Epoch 78/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0205 - accuracy: 0.9958 - val_loss: 0.2280 - val_accuracy: 0.9396\n",
      "Epoch 79/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0236 - accuracy: 0.9975 - val_loss: 0.2216 - val_accuracy: 0.9430\n",
      "Epoch 80/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0176 - accuracy: 1.0000 - val_loss: 0.2221 - val_accuracy: 0.9396\n",
      "Epoch 81/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0194 - accuracy: 0.9958 - val_loss: 0.2243 - val_accuracy: 0.9430\n",
      "Epoch 82/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0154 - accuracy: 0.9983 - val_loss: 0.2257 - val_accuracy: 0.9430\n",
      "Epoch 83/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0196 - accuracy: 0.9966 - val_loss: 0.2316 - val_accuracy: 0.9396\n",
      "Epoch 84/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0164 - accuracy: 0.9975 - val_loss: 0.2265 - val_accuracy: 0.9396\n",
      "Epoch 85/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0189 - accuracy: 0.9975 - val_loss: 0.2283 - val_accuracy: 0.9362\n",
      "Epoch 86/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0186 - accuracy: 0.9992 - val_loss: 0.2281 - val_accuracy: 0.9396\n",
      "Epoch 87/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0148 - accuracy: 1.0000 - val_loss: 0.2323 - val_accuracy: 0.9396\n",
      "Epoch 88/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0144 - accuracy: 0.9983 - val_loss: 0.2344 - val_accuracy: 0.9396\n",
      "Epoch 89/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0165 - accuracy: 0.9975 - val_loss: 0.2352 - val_accuracy: 0.9396\n",
      "Epoch 90/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0234 - accuracy: 0.9950 - val_loss: 0.2378 - val_accuracy: 0.9396\n",
      "Epoch 91/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0168 - accuracy: 0.9958 - val_loss: 0.2334 - val_accuracy: 0.9430\n",
      "Epoch 92/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0147 - accuracy: 0.9992 - val_loss: 0.2373 - val_accuracy: 0.9430\n",
      "Epoch 93/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0180 - accuracy: 0.9992 - val_loss: 0.2364 - val_accuracy: 0.9430\n",
      "Epoch 94/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0134 - accuracy: 0.9992 - val_loss: 0.2364 - val_accuracy: 0.9396\n",
      "Epoch 95/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0142 - accuracy: 0.9983 - val_loss: 0.2398 - val_accuracy: 0.9396\n",
      "Epoch 96/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0116 - accuracy: 0.9992 - val_loss: 0.2395 - val_accuracy: 0.9396\n",
      "Epoch 97/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0145 - accuracy: 0.9966 - val_loss: 0.2415 - val_accuracy: 0.9362\n",
      "Epoch 98/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0194 - accuracy: 0.9950 - val_loss: 0.2483 - val_accuracy: 0.9396\n",
      "Epoch 99/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0161 - accuracy: 0.9958 - val_loss: 0.2416 - val_accuracy: 0.9430\n",
      "Epoch 100/100\n",
      "38/38 [==============================] - 0s 2ms/step - loss: 0.0103 - accuracy: 0.9992 - val_loss: 0.2417 - val_accuracy: 0.9430\n"
     ]
    }
   ],
   "source": [
    "history = model.fit(train_padded_seq, train_label_seq, epochs=100, validation_data=(val_padded_seq, val_label_seq))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "id": "rural-sheffield",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "def plot_graphs(history, metric):\n",
    "    plt.plot(history.history[metric])\n",
    "    plt.plot(history.history[f'val_{metric}'])\n",
    "    plt.xlabel(\"Epochs\")\n",
    "    plt.ylabel(metric)\n",
    "    plt.legend([metric, f'val_{metric}'])\n",
    "    plt.show()\n",
    "    \n",
    "plot_graphs(history, \"accuracy\")\n",
    "plot_graphs(history, \"loss\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "material-breast",
   "metadata": {},
   "source": [
    "### Model Evaluation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "id": "40641380",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "10/10 [==============================] - 0s 759us/step\n"
     ]
    }
   ],
   "source": [
    "preds = model.predict(val_padded_seq)\n",
    "indices, values = tf.nn.top_k(preds, k=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "id": "4149c1c6",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix\n",
    "import seaborn as sns\n",
    "\n",
    "labels = label_tokenizer.word_index.keys()\n",
    "conf_mat = confusion_matrix(val_label_seq, values)\n",
    "sns.heatmap(conf_mat, annot=True, fmt='d',\n",
    "                       xticklabels=labels, yticklabels=labels, cmap='Blues')\n",
    "plt.ylabel('Actual')\n",
    "plt.xlabel('Predicted')\n",
    "plt.show()"
   ]
  }
 ],
 "metadata": {
  "dlai_version": "1.2.0",
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}