Modelling Deep Learning Part 1.ipynb

{
 "cells": [
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {
    "id": "No_9ci7iEk4C"
   },
   "source": [
    "# Library"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Requirement already satisfied: pandas in c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages (1.5.3)\n",
      "Requirement already satisfied: numpy>=1.21.0 in c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages (from pandas) (1.23.5)\n",
      "Requirement already satisfied: pytz>=2020.1 in c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages (from pandas) (2023.3)\n",
      "Requirement already satisfied: python-dateutil>=2.8.1 in c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages (from pandas) (2.8.2)\n",
      "Requirement already satisfied: six>=1.5 in c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages (from python-dateutil>=2.8.1->pandas) (1.16.0)\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING: Ignoring invalid distribution -rotobuf (c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages)\n",
      "WARNING: Ignoring invalid distribution -rotobuf (c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages)\n",
      "WARNING: Ignoring invalid distribution -rotobuf (c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages)\n",
      "WARNING: Ignoring invalid distribution -rotobuf (c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages)\n",
      "WARNING: Ignoring invalid distribution -rotobuf (c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages)\n",
      "WARNING: Ignoring invalid distribution -rotobuf (c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages)\n"
     ]
    }
   ],
   "source": [
    "!pip install pandas"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Requirement already satisfied: matplotlib in c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages (3.6.3)\n",
      "Requirement already satisfied: contourpy>=1.0.1 in c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages (from matplotlib) (1.0.7)\n",
      "Requirement already satisfied: cycler>=0.10 in c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages (from matplotlib) (0.11.0)\n",
      "Requirement already satisfied: packaging>=20.0 in c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages (from matplotlib) (23.1)\n",
      "Requirement already satisfied: pillow>=6.2.0 in c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages (from matplotlib) (9.4.0)\n",
      "Requirement already satisfied: python-dateutil>=2.7 in c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages (from matplotlib) (2.8.2)\n",
      "Requirement already satisfied: kiwisolver>=1.0.1 in c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages (from matplotlib) (1.4.4)\n",
      "Requirement already satisfied: fonttools>=4.22.0 in c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages (from matplotlib) (4.39.2)\n",
      "Requirement already satisfied: numpy>=1.19 in c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages (from matplotlib) (1.23.5)\n",
      "Requirement already satisfied: pyparsing>=2.2.1 in c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages (from matplotlib) (3.0.9)\n",
      "Requirement already satisfied: six>=1.5 in c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages (from python-dateutil>=2.7->matplotlib) (1.16.0)\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING: Ignoring invalid distribution -rotobuf (c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages)\n",
      "WARNING: Ignoring invalid distribution -rotobuf (c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages)\n",
      "WARNING: Ignoring invalid distribution -rotobuf (c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages)\n",
      "WARNING: Ignoring invalid distribution -rotobuf (c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages)\n",
      "WARNING: Ignoring invalid distribution -rotobuf (c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages)\n",
      "WARNING: Ignoring invalid distribution -rotobuf (c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages)\n"
     ]
    }
   ],
   "source": [
    "!pip install matplotlib"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Requirement already satisfied: opendatasets in c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages (0.1.22)\n",
      "Requirement already satisfied: click in c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages (from opendatasets) (8.1.3)\n",
      "Requirement already satisfied: kaggle in c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages (from opendatasets) (1.5.13)\n",
      "Requirement already satisfied: tqdm in c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages (from opendatasets) (4.64.1)\n",
      "Requirement already satisfied: colorama in c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages (from click->opendatasets) (0.4.6)\n",
      "Requirement already satisfied: urllib3 in c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages (from kaggle->opendatasets) (1.26.15)\n",
      "Requirement already satisfied: certifi in c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages (from kaggle->opendatasets) (2022.12.7)\n",
      "Requirement already satisfied: requests in c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages (from kaggle->opendatasets) (2.28.2)\n",
      "Requirement already satisfied: six>=1.10 in c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages (from kaggle->opendatasets) (1.16.0)\n",
      "Requirement already satisfied: python-dateutil in c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages (from kaggle->opendatasets) (2.8.2)\n",
      "Requirement already satisfied: python-slugify in c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages (from kaggle->opendatasets) (8.0.1)\n",
      "Requirement already satisfied: text-unidecode>=1.3 in c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages (from python-slugify->kaggle->opendatasets) (1.3)\n",
      "Requirement already satisfied: idna<4,>=2.5 in c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages (from requests->kaggle->opendatasets) (3.4)\n",
      "Requirement already satisfied: charset-normalizer<4,>=2 in c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages (from requests->kaggle->opendatasets) (3.1.0)\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING: Ignoring invalid distribution -rotobuf (c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages)\n",
      "WARNING: Ignoring invalid distribution -rotobuf (c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages)\n",
      "WARNING: Ignoring invalid distribution -rotobuf (c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages)\n",
      "WARNING: Ignoring invalid distribution -rotobuf (c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages)\n",
      "WARNING: Ignoring invalid distribution -rotobuf (c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages)\n",
      "WARNING: Ignoring invalid distribution -rotobuf (c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages)\n"
     ]
    }
   ],
   "source": [
    "!pip install opendatasets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Requirement already satisfied: scikit-learn in c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages (1.2.2)\n",
      "Requirement already satisfied: threadpoolctl>=2.0.0 in c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages (from scikit-learn) (3.1.0)\n",
      "Requirement already satisfied: joblib>=1.1.1 in c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages (from scikit-learn) (1.2.0)\n",
      "Requirement already satisfied: scipy>=1.3.2 in c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages (from scikit-learn) (1.9.3)\n",
      "Requirement already satisfied: numpy>=1.17.3 in c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages (from scikit-learn) (1.23.5)\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING: Ignoring invalid distribution -rotobuf (c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages)\n",
      "WARNING: Ignoring invalid distribution -rotobuf (c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages)\n",
      "WARNING: Ignoring invalid distribution -rotobuf (c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages)\n",
      "WARNING: Ignoring invalid distribution -rotobuf (c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages)\n",
      "WARNING: Ignoring invalid distribution -rotobuf (c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages)\n",
      "WARNING: Ignoring invalid distribution -rotobuf (c:\\users\\hasan\\anaconda3\\envs\\tensorflow_gpu_zein\\lib\\site-packages)\n"
     ]
    }
   ],
   "source": [
    "!pip install scikit-learn"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "id": "HqzGEMvPEk4F",
    "tags": []
   },
   "outputs": [],
   "source": [
    "# Library to dataset\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import opendatasets as od\n",
    "\n",
    "# Library tensorflow\n",
    "import tensorflow as tf\n",
    "from tensorflow.keras.preprocessing.text import Tokenizer\n",
    "from tensorflow.keras.preprocessing.sequence import pad_sequences\n",
    "from tensorflow.keras.callbacks import ModelCheckpoint, EarlyStopping, ReduceLROnPlateau\n",
    "\n",
    "# Surpress warning\n",
    "import warnings\n",
    "warnings.filterwarnings(\"ignore\")\n",
    "\n",
    "# Library sklearn\n",
    "from sklearn.preprocessing import LabelEncoder\n",
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "# pickle\n",
    "import pickle\n"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 206
    },
    "id": "ly5YTaNrEk4G",
    "outputId": "9a0d7fcf-2d0b-4ba2-ff17-0ad1e27d4b71",
    "tags": []
   },
   "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>text</th>\n",
       "      <th>class</th>\n",
       "      <th>clean_text</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Ex Wife Threatening SuicideRecently I left my ...</td>\n",
       "      <td>suicide</td>\n",
       "      <td>ex wife threaten suiciderec left wife good che...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Am I weird I don't get affected by compliments...</td>\n",
       "      <td>non-suicide</td>\n",
       "      <td>weird get affect compliment come someon know i...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Finally 2020 is almost over... So I can never ...</td>\n",
       "      <td>non-suicide</td>\n",
       "      <td>final almost never hear bad year ever swear fu...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>i need helpjust help me im crying so hard</td>\n",
       "      <td>suicide</td>\n",
       "      <td>need helpjust help im cri hard</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>I’m so lostHello, my name is Adam (16) and I’v...</td>\n",
       "      <td>suicide</td>\n",
       "      <td>losthello name adam struggl year afraid past y...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                text        class  \\\n",
       "0  Ex Wife Threatening SuicideRecently I left my ...      suicide   \n",
       "1  Am I weird I don't get affected by compliments...  non-suicide   \n",
       "2  Finally 2020 is almost over... So I can never ...  non-suicide   \n",
       "3          i need helpjust help me im crying so hard      suicide   \n",
       "4  I’m so lostHello, my name is Adam (16) and I’v...      suicide   \n",
       "\n",
       "                                          clean_text  \n",
       "0  ex wife threaten suiciderec left wife good che...  \n",
       "1  weird get affect compliment come someon know i...  \n",
       "2  final almost never hear bad year ever swear fu...  \n",
       "3                     need helpjust help im cri hard  \n",
       "4  losthello name adam struggl year afraid past y...  "
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Read dataset\n",
    "data = pd.read_csv('clean_text.csv')\n",
    "data.head()"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Encoding"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "id": "tI4zGvPDEk4H",
    "tags": []
   },
   "outputs": [],
   "source": [
    "X = data['clean_text'].astype(str)\n",
    "y = data['class']\n",
    "\n",
    "# encode class values as integers\n",
    "# Define label encoder\n",
    "encoder = LabelEncoder()\n",
    "# fit and transform\n",
    "encoded_y = encoder.fit_transform(y)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Test Train Split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "id": "m7z7S5FdEk4H",
    "tags": []
   },
   "outputs": [],
   "source": [
    "# Test and train split\n",
    "X_train, X_test, y_train, y_test = train_test_split(X,encoded_y, test_size=.2, random_state=42)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {
    "id": "YksTB-KzEk4I"
   },
   "source": [
    "# Data Preprocessing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "id": "sIEVlUS7Ek4I",
    "tags": []
   },
   "outputs": [],
   "source": [
    "vocab_size = 10_000\n",
    "max_length = 10_000\n",
    "trunc_type='post'\n",
    "padding_type='post'\n",
    "oov_tok = \"<OOV>\"\n",
    "\n",
    "# Initialize the Tokenizer class\n",
    "tokenizer = Tokenizer(num_words=vocab_size, oov_token=oov_tok)\n",
    "\n",
    "# Generate the word index dictionary\n",
    "tokenizer.fit_on_texts(X_train)\n",
    "word_index = tokenizer.word_index\n",
    "\n",
    "# Generate and pad the training sequences\n",
    "training_sequences = tokenizer.texts_to_sequences(X_train)\n",
    "training_padded = pad_sequences(training_sequences, maxlen=max_length, padding=padding_type, truncating=trunc_type)\n",
    "\n",
    "# Generate and pad the testing sequences\n",
    "testing_sequences = tokenizer.texts_to_sequences(X_test)\n",
    "testing_padded = pad_sequences(testing_sequences, maxlen=max_length, padding=padding_type, truncating=trunc_type)\n",
    "\n",
    "# Convert the labels lists into numpy arrays\n",
    "training_labels = np.array(y_train)\n",
    "testing_labels = np.array(y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Save the word index dictionary\n",
    "np.save('word_index.npy', word_index)\n",
    "\n",
    "# Save Tokenizer\n",
    "with open('tokenizer.pickle', 'wb') as handle:\n",
    "    pickle.dump(tokenizer, handle, protocol=pickle.HIGHEST_PROTOCOL)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {
    "id": "uw8lhhWEEk4J"
   },
   "source": [
    "# Basic Model Embeding Layers"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "of8FCYLUEk4J",
    "outputId": "c84ae2f6-9b35-4989-e7ae-80118688de8c",
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " embedding (Embedding)       (None, 10000, 16)         160000    \n",
      "                                                                 \n",
      " flatten (Flatten)           (None, 160000)            0         \n",
      "                                                                 \n",
      " dense (Dense)               (None, 128)               20480128  \n",
      "                                                                 \n",
      " dense_1 (Dense)             (None, 128)               16512     \n",
      "                                                                 \n",
      " dense_2 (Dense)             (None, 128)               16512     \n",
      "                                                                 \n",
      " dense_3 (Dense)             (None, 1)                 129       \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 20,673,281\n",
      "Trainable params: 20,673,281\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "embedding_dim = 16\n",
    "\n",
    "# Initialize the model\n",
    "model = tf.keras.Sequential([\n",
    "    tf.keras.layers.Embedding(vocab_size, embedding_dim, input_length=max_length),\n",
    "    tf.keras.layers.Flatten(),\n",
    "    tf.keras.layers.Dense(128, activation='relu'),\n",
    "    tf.keras.layers.Dense(128, activation='relu'),\n",
    "    tf.keras.layers.Dense(128, activation='relu'),\n",
    "    tf.keras.layers.Dense(1, activation='sigmoid')\n",
    "])\n",
    "\n",
    "# Compile the model\n",
    "model.compile(loss='binary_crossentropy',optimizer='adam',metrics=['accuracy'])\n",
    "\n",
    "# Print the model summary\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Difine the callbacks\n",
    "# Save the best model\n",
    "checkpoint_callback = ModelCheckpoint(filepath=\"Basic Model Embeding Layers.h5\", save_best_only=True, monitor=\"val_accuracy\", mode=\"max\", verbose=1)\n",
    "\n",
    "# Early stopping\n",
    "early_stopping_callback = EarlyStopping(monitor=\"val_accuracy\", mode=\"max\", patience=10, verbose=1, restore_best_weights=True)\n",
    "\n",
    "# Reduce learning rate\n",
    "reduce_lr_callback = ReduceLROnPlateau(monitor=\"val_loss\", factor=0.1, patience=5, verbose=1, mode=\"min\", min_delta=0.0001, cooldown=0, min_lr=0)\n",
    "\n",
    "# Callbacks list\n",
    "callbacks=[checkpoint_callback, early_stopping_callback, reduce_lr_callback]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "id": "wt_f2RcWEk4J",
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/100\n",
      "5799/5802 [============================>.] - ETA: 0s - loss: 0.2011 - accuracy: 0.9216\n",
      "Epoch 1: val_accuracy improved from -inf to 0.93967, saving model to Basic Model Embeding Layers.h5\n",
      "5802/5802 [==============================] - 51s 8ms/step - loss: 0.2011 - accuracy: 0.9216 - val_loss: 0.1629 - val_accuracy: 0.9397 - lr: 0.0010\n",
      "Epoch 2/100\n",
      "5798/5802 [============================>.] - ETA: 0s - loss: 0.1565 - accuracy: 0.9416\n",
      "Epoch 2: val_accuracy improved from 0.93967 to 0.94030, saving model to Basic Model Embeding Layers.h5\n",
      "5802/5802 [==============================] - 49s 8ms/step - loss: 0.1565 - accuracy: 0.9416 - val_loss: 0.1603 - val_accuracy: 0.9403 - lr: 0.0010\n",
      "Epoch 3/100\n",
      "5796/5802 [============================>.] - ETA: 0s - loss: 0.1397 - accuracy: 0.9486\n",
      "Epoch 3: val_accuracy improved from 0.94030 to 0.94099, saving model to Basic Model Embeding Layers.h5\n",
      "5802/5802 [==============================] - 49s 8ms/step - loss: 0.1397 - accuracy: 0.9486 - val_loss: 0.1588 - val_accuracy: 0.9410 - lr: 0.0010\n",
      "Epoch 4/100\n",
      "5799/5802 [============================>.] - ETA: 0s - loss: 0.1279 - accuracy: 0.9530\n",
      "Epoch 4: val_accuracy did not improve from 0.94099\n",
      "5802/5802 [==============================] - 49s 8ms/step - loss: 0.1279 - accuracy: 0.9530 - val_loss: 0.1623 - val_accuracy: 0.9402 - lr: 0.0010\n",
      "Epoch 5/100\n",
      "5797/5802 [============================>.] - ETA: 0s - loss: 0.1115 - accuracy: 0.9592\n",
      "Epoch 5: val_accuracy did not improve from 0.94099\n",
      "5802/5802 [==============================] - 49s 8ms/step - loss: 0.1116 - accuracy: 0.9592 - val_loss: 0.1823 - val_accuracy: 0.9332 - lr: 0.0010\n",
      "Epoch 6/100\n",
      "5802/5802 [==============================] - ETA: 0s - loss: 0.0933 - accuracy: 0.9662\n",
      "Epoch 6: val_accuracy did not improve from 0.94099\n",
      "5802/5802 [==============================] - 47s 8ms/step - loss: 0.0933 - accuracy: 0.9662 - val_loss: 0.1857 - val_accuracy: 0.9371 - lr: 0.0010\n",
      "Epoch 7/100\n",
      "5802/5802 [==============================] - ETA: 0s - loss: 0.0780 - accuracy: 0.9725\n",
      "Epoch 7: val_accuracy did not improve from 0.94099\n",
      "5802/5802 [==============================] - 46s 8ms/step - loss: 0.0780 - accuracy: 0.9725 - val_loss: 0.2044 - val_accuracy: 0.9334 - lr: 0.0010\n",
      "Epoch 8/100\n",
      "5798/5802 [============================>.] - ETA: 0s - loss: 0.0643 - accuracy: 0.9775\n",
      "Epoch 8: val_accuracy did not improve from 0.94099\n",
      "\n",
      "Epoch 8: ReduceLROnPlateau reducing learning rate to 0.00010000000474974513.\n",
      "5802/5802 [==============================] - 46s 8ms/step - loss: 0.0643 - accuracy: 0.9775 - val_loss: 0.2474 - val_accuracy: 0.9305 - lr: 0.0010\n",
      "Epoch 9/100\n",
      "5798/5802 [============================>.] - ETA: 0s - loss: 0.0426 - accuracy: 0.9861\n",
      "Epoch 9: val_accuracy did not improve from 0.94099\n",
      "5802/5802 [==============================] - 46s 8ms/step - loss: 0.0426 - accuracy: 0.9860 - val_loss: 0.2867 - val_accuracy: 0.9270 - lr: 1.0000e-04\n",
      "Epoch 10/100\n",
      "5801/5802 [============================>.] - ETA: 0s - loss: 0.0381 - accuracy: 0.9877\n",
      "Epoch 10: val_accuracy did not improve from 0.94099\n",
      "5802/5802 [==============================] - 46s 8ms/step - loss: 0.0381 - accuracy: 0.9877 - val_loss: 0.3030 - val_accuracy: 0.9274 - lr: 1.0000e-04\n",
      "Epoch 11/100\n",
      "5800/5802 [============================>.] - ETA: 0s - loss: 0.0356 - accuracy: 0.9888\n",
      "Epoch 11: val_accuracy did not improve from 0.94099\n",
      "5802/5802 [==============================] - 48s 8ms/step - loss: 0.0356 - accuracy: 0.9888 - val_loss: 0.3273 - val_accuracy: 0.9266 - lr: 1.0000e-04\n",
      "Epoch 12/100\n",
      "5797/5802 [============================>.] - ETA: 0s - loss: 0.0333 - accuracy: 0.9896\n",
      "Epoch 12: val_accuracy did not improve from 0.94099\n",
      "5802/5802 [==============================] - 49s 8ms/step - loss: 0.0333 - accuracy: 0.9896 - val_loss: 0.3406 - val_accuracy: 0.9243 - lr: 1.0000e-04\n",
      "Epoch 13/100\n",
      "5796/5802 [============================>.] - ETA: 0s - loss: 0.0314 - accuracy: 0.9902\n",
      "Epoch 13: val_accuracy did not improve from 0.94099\n",
      "Restoring model weights from the end of the best epoch: 3.\n",
      "\n",
      "Epoch 13: ReduceLROnPlateau reducing learning rate to 1.0000000474974514e-05.\n",
      "5802/5802 [==============================] - 47s 8ms/step - loss: 0.0314 - accuracy: 0.9902 - val_loss: 0.3590 - val_accuracy: 0.9246 - lr: 1.0000e-04\n",
      "Epoch 13: early stopping\n"
     ]
    }
   ],
   "source": [
    "# Define the number of epochs\n",
    "num_epochs = 100\n",
    "\n",
    "# Train the model\n",
    "history = model.fit(training_padded, training_labels, epochs=num_epochs, validation_data=(testing_padded, testing_labels), callbacks=callbacks)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1451/1451 [==============================] - 4s 2ms/step - loss: 0.1588 - accuracy: 0.9410\n",
      "Test results - Loss: 0.15879160165786743 - Accuracy: 94.09888982772827%\n"
     ]
    }
   ],
   "source": [
    "# Evaluate the model\n",
    "results_base = model.evaluate(testing_padded, testing_labels)\n",
    "\n",
    "# print results\n",
    "print(f'Test results - Loss: {results_base[0]} - Accuracy: {100*results_base[1]}%')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "id": "tscfgTUxEk4K",
    "outputId": "3c42022d-dc3a-49a8-d2e8-762226ac2682",
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot utility\n",
    "def plot_graphs(history, string):\n",
    "  plt.plot(history.history[string])\n",
    "  plt.plot(history.history['val_'+string])\n",
    "  plt.xlabel(\"Epochs\")\n",
    "  plt.ylabel(string)\n",
    "  plt.legend([string, 'val_'+string])\n",
    "  plt.show()\n",
    "  \n",
    "# Plot the accuracy and loss\n",
    "plot_graphs(history, \"accuracy\")\n",
    "plot_graphs(history, \"loss\")"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {
    "id": "B0Msocs1Ek4K"
   },
   "source": [
    "# Model Basic + Global Average Pooling"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "id": "smU8ZKgyEk4K",
    "outputId": "4eaa90c3-6807-4370-f2f5-25139bbb2192",
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_1\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " embedding_1 (Embedding)     (None, 10000, 16)         160000    \n",
      "                                                                 \n",
      " global_average_pooling1d (G  (None, 16)               0         \n",
      " lobalAveragePooling1D)                                          \n",
      "                                                                 \n",
      " dense_4 (Dense)             (None, 128)               2176      \n",
      "                                                                 \n",
      " dense_5 (Dense)             (None, 128)               16512     \n",
      "                                                                 \n",
      " dense_6 (Dense)             (None, 128)               16512     \n",
      "                                                                 \n",
      " dense_7 (Dense)             (None, 1)                 129       \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 195,329\n",
      "Trainable params: 195,329\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "embedding_dim = 16\n",
    "\n",
    "# Initialize the model\n",
    "model_GlobalAveragePooling1D = tf.keras.Sequential([\n",
    "    tf.keras.layers.Embedding(vocab_size, embedding_dim, input_length=max_length),\n",
    "    tf.keras.layers.GlobalAveragePooling1D(),\n",
    "    tf.keras.layers.Dense(128, activation='relu'),\n",
    "    tf.keras.layers.Dense(128, activation='relu'),\n",
    "    tf.keras.layers.Dense(128, activation='relu'),\n",
    "    tf.keras.layers.Dense(1, activation='sigmoid')\n",
    "])\n",
    "\n",
    "# Compile the model\n",
    "model_GlobalAveragePooling1D.compile(loss='binary_crossentropy',optimizer='adam',metrics=['accuracy'])\n",
    "\n",
    "# Print the model summary\n",
    "model_GlobalAveragePooling1D.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Difine the callbacks\n",
    "# Save the best model\n",
    "checkpoint_callback = ModelCheckpoint(filepath=\"Model Basic + Global Average Pooling.h5\", save_best_only=True, monitor=\"val_accuracy\", mode=\"max\", verbose=1)\n",
    "\n",
    "# Early stopping\n",
    "early_stopping_callback = EarlyStopping(monitor=\"val_accuracy\", mode=\"max\", patience=10, verbose=1, restore_best_weights=True)\n",
    "\n",
    "# Reduce learning rate\n",
    "reduce_lr_callback = ReduceLROnPlateau(monitor=\"val_loss\", factor=0.1, patience=5, verbose=1, mode=\"min\", min_delta=0.0001, cooldown=0, min_lr=0)\n",
    "\n",
    "# Callbacks list\n",
    "callbacks=[checkpoint_callback, early_stopping_callback, reduce_lr_callback]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "id": "kgUSXmCqEk4L",
    "outputId": "ba0c3d0c-5aa0-4624-b5f8-ffef531433e5",
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/100\n",
      "5801/5802 [============================>.] - ETA: 0s - loss: 0.4517 - accuracy: 0.7460\n",
      "Epoch 1: val_accuracy improved from -inf to 0.91623, saving model to Model Basic + Global Average Pooling.h5\n",
      "5802/5802 [==============================] - 32s 5ms/step - loss: 0.4517 - accuracy: 0.7461 - val_loss: 0.2283 - val_accuracy: 0.9162 - lr: 0.0010\n",
      "Epoch 2/100\n",
      "5793/5802 [============================>.] - ETA: 0s - loss: 0.2382 - accuracy: 0.9082\n",
      "Epoch 2: val_accuracy improved from 0.91623 to 0.92147, saving model to Model Basic + Global Average Pooling.h5\n",
      "5802/5802 [==============================] - 31s 5ms/step - loss: 0.2382 - accuracy: 0.9082 - val_loss: 0.2023 - val_accuracy: 0.9215 - lr: 0.0010\n",
      "Epoch 3/100\n",
      "5800/5802 [============================>.] - ETA: 0s - loss: 0.2209 - accuracy: 0.9155\n",
      "Epoch 3: val_accuracy did not improve from 0.92147\n",
      "5802/5802 [==============================] - 31s 5ms/step - loss: 0.2209 - accuracy: 0.9155 - val_loss: 0.2722 - val_accuracy: 0.8907 - lr: 0.0010\n",
      "Epoch 4/100\n",
      "5799/5802 [============================>.] - ETA: 0s - loss: 0.2085 - accuracy: 0.9206\n",
      "Epoch 4: val_accuracy improved from 0.92147 to 0.92729, saving model to Model Basic + Global Average Pooling.h5\n",
      "5802/5802 [==============================] - 31s 5ms/step - loss: 0.2085 - accuracy: 0.9206 - val_loss: 0.1905 - val_accuracy: 0.9273 - lr: 0.0010\n",
      "Epoch 5/100\n",
      "5794/5802 [============================>.] - ETA: 0s - loss: 0.2028 - accuracy: 0.9231\n",
      "Epoch 5: val_accuracy did not improve from 0.92729\n",
      "5802/5802 [==============================] - 31s 5ms/step - loss: 0.2027 - accuracy: 0.9231 - val_loss: 0.1942 - val_accuracy: 0.9244 - lr: 0.0010\n",
      "Epoch 6/100\n",
      "5801/5802 [============================>.] - ETA: 0s - loss: 0.1984 - accuracy: 0.9246\n",
      "Epoch 6: val_accuracy did not improve from 0.92729\n",
      "5802/5802 [==============================] - 32s 5ms/step - loss: 0.1984 - accuracy: 0.9246 - val_loss: 0.2045 - val_accuracy: 0.9226 - lr: 0.0010\n",
      "Epoch 7/100\n",
      "5797/5802 [============================>.] - ETA: 0s - loss: 0.1940 - accuracy: 0.9260\n",
      "Epoch 7: val_accuracy did not improve from 0.92729\n",
      "5802/5802 [==============================] - 35s 6ms/step - loss: 0.1940 - accuracy: 0.9260 - val_loss: 0.2019 - val_accuracy: 0.9241 - lr: 0.0010\n",
      "Epoch 8/100\n",
      "5799/5802 [============================>.] - ETA: 0s - loss: 0.1880 - accuracy: 0.9294\n",
      "Epoch 8: val_accuracy did not improve from 0.92729\n",
      "5802/5802 [==============================] - 35s 6ms/step - loss: 0.1881 - accuracy: 0.9294 - val_loss: 0.2111 - val_accuracy: 0.9241 - lr: 0.0010\n",
      "Epoch 9/100\n",
      "5794/5802 [============================>.] - ETA: 0s - loss: 0.1873 - accuracy: 0.9279\n",
      "Epoch 9: val_accuracy improved from 0.92729 to 0.93250, saving model to Model Basic + Global Average Pooling.h5\n",
      "5802/5802 [==============================] - 52s 9ms/step - loss: 0.1873 - accuracy: 0.9279 - val_loss: 0.1859 - val_accuracy: 0.9325 - lr: 0.0010\n",
      "Epoch 10/100\n",
      "5800/5802 [============================>.] - ETA: 0s - loss: 0.1850 - accuracy: 0.9287\n",
      "Epoch 10: val_accuracy improved from 0.93250 to 0.93349, saving model to Model Basic + Global Average Pooling.h5\n",
      "5802/5802 [==============================] - 56s 10ms/step - loss: 0.1850 - accuracy: 0.9288 - val_loss: 0.1757 - val_accuracy: 0.9335 - lr: 0.0010\n",
      "Epoch 11/100\n",
      "5795/5802 [============================>.] - ETA: 0s - loss: 0.1818 - accuracy: 0.9303\n",
      "Epoch 11: val_accuracy did not improve from 0.93349\n",
      "5802/5802 [==============================] - 34s 6ms/step - loss: 0.1817 - accuracy: 0.9303 - val_loss: 0.1756 - val_accuracy: 0.9322 - lr: 0.0010\n",
      "Epoch 12/100\n",
      "5797/5802 [============================>.] - ETA: 0s - loss: 0.1795 - accuracy: 0.9311\n",
      "Epoch 12: val_accuracy did not improve from 0.93349\n",
      "5802/5802 [==============================] - 34s 6ms/step - loss: 0.1795 - accuracy: 0.9311 - val_loss: 0.1871 - val_accuracy: 0.9290 - lr: 0.0010\n",
      "Epoch 13/100\n",
      "5794/5802 [============================>.] - ETA: 0s - loss: 0.1781 - accuracy: 0.9314\n",
      "Epoch 13: val_accuracy did not improve from 0.93349\n",
      "5802/5802 [==============================] - 33s 6ms/step - loss: 0.1780 - accuracy: 0.9314 - val_loss: 0.1764 - val_accuracy: 0.9327 - lr: 0.0010\n",
      "Epoch 14/100\n",
      "5798/5802 [============================>.] - ETA: 0s - loss: 0.1778 - accuracy: 0.9314\n",
      "Epoch 14: val_accuracy did not improve from 0.93349\n",
      "5802/5802 [==============================] - 33s 6ms/step - loss: 0.1778 - accuracy: 0.9314 - val_loss: 0.1901 - val_accuracy: 0.9332 - lr: 0.0010\n",
      "Epoch 15/100\n",
      "5799/5802 [============================>.] - ETA: 0s - loss: 0.1750 - accuracy: 0.9326\n",
      "Epoch 15: val_accuracy improved from 0.93349 to 0.93379, saving model to Model Basic + Global Average Pooling.h5\n",
      "\n",
      "Epoch 15: ReduceLROnPlateau reducing learning rate to 0.00010000000474974513.\n",
      "5802/5802 [==============================] - 33s 6ms/step - loss: 0.1750 - accuracy: 0.9326 - val_loss: 0.1828 - val_accuracy: 0.9338 - lr: 0.0010\n",
      "Epoch 16/100\n",
      "5799/5802 [============================>.] - ETA: 0s - loss: 0.1520 - accuracy: 0.9412\n",
      "Epoch 16: val_accuracy improved from 0.93379 to 0.93478, saving model to Model Basic + Global Average Pooling.h5\n",
      "5802/5802 [==============================] - 33s 6ms/step - loss: 0.1520 - accuracy: 0.9412 - val_loss: 0.1705 - val_accuracy: 0.9348 - lr: 1.0000e-04\n",
      "Epoch 17/100\n",
      "5798/5802 [============================>.] - ETA: 0s - loss: 0.1505 - accuracy: 0.9418\n",
      "Epoch 17: val_accuracy did not improve from 0.93478\n",
      "5802/5802 [==============================] - 33s 6ms/step - loss: 0.1505 - accuracy: 0.9418 - val_loss: 0.1766 - val_accuracy: 0.9342 - lr: 1.0000e-04\n",
      "Epoch 18/100\n",
      "5799/5802 [============================>.] - ETA: 0s - loss: 0.1501 - accuracy: 0.9415\n",
      "Epoch 18: val_accuracy improved from 0.93478 to 0.93500, saving model to Model Basic + Global Average Pooling.h5\n",
      "5802/5802 [==============================] - 34s 6ms/step - loss: 0.1501 - accuracy: 0.9415 - val_loss: 0.1701 - val_accuracy: 0.9350 - lr: 1.0000e-04\n",
      "Epoch 19/100\n",
      "5799/5802 [============================>.] - ETA: 0s - loss: 0.1496 - accuracy: 0.9419\n",
      "Epoch 19: val_accuracy improved from 0.93500 to 0.93539, saving model to Model Basic + Global Average Pooling.h5\n",
      "5802/5802 [==============================] - 34s 6ms/step - loss: 0.1496 - accuracy: 0.9419 - val_loss: 0.1699 - val_accuracy: 0.9354 - lr: 1.0000e-04\n",
      "Epoch 20/100\n",
      "5799/5802 [============================>.] - ETA: 0s - loss: 0.1489 - accuracy: 0.9419\n",
      "Epoch 20: val_accuracy did not improve from 0.93539\n",
      "5802/5802 [==============================] - 33s 6ms/step - loss: 0.1489 - accuracy: 0.9419 - val_loss: 0.1699 - val_accuracy: 0.9353 - lr: 1.0000e-04\n",
      "Epoch 21/100\n",
      "5801/5802 [============================>.] - ETA: 0s - loss: 0.1485 - accuracy: 0.9421\n",
      "Epoch 21: val_accuracy did not improve from 0.93539\n",
      "5802/5802 [==============================] - 33s 6ms/step - loss: 0.1485 - accuracy: 0.9421 - val_loss: 0.1779 - val_accuracy: 0.9339 - lr: 1.0000e-04\n",
      "Epoch 22/100\n",
      "5802/5802 [==============================] - ETA: 0s - loss: 0.1481 - accuracy: 0.9427\n",
      "Epoch 22: val_accuracy did not improve from 0.93539\n",
      "5802/5802 [==============================] - 33s 6ms/step - loss: 0.1481 - accuracy: 0.9427 - val_loss: 0.1780 - val_accuracy: 0.9319 - lr: 1.0000e-04\n",
      "Epoch 23/100\n",
      "5795/5802 [============================>.] - ETA: 0s - loss: 0.1481 - accuracy: 0.9424\n",
      "Epoch 23: val_accuracy did not improve from 0.93539\n",
      "5802/5802 [==============================] - 33s 6ms/step - loss: 0.1482 - accuracy: 0.9424 - val_loss: 0.1719 - val_accuracy: 0.9345 - lr: 1.0000e-04\n",
      "Epoch 24/100\n",
      "5793/5802 [============================>.] - ETA: 0s - loss: 0.1477 - accuracy: 0.9423\n",
      "Epoch 24: val_accuracy did not improve from 0.93539\n",
      "\n",
      "Epoch 24: ReduceLROnPlateau reducing learning rate to 1.0000000474974514e-05.\n",
      "5802/5802 [==============================] - 33s 6ms/step - loss: 0.1478 - accuracy: 0.9423 - val_loss: 0.1737 - val_accuracy: 0.9342 - lr: 1.0000e-04\n",
      "Epoch 25/100\n",
      "5801/5802 [============================>.] - ETA: 0s - loss: 0.1443 - accuracy: 0.9438\n",
      "Epoch 25: val_accuracy did not improve from 0.93539\n",
      "5802/5802 [==============================] - 38s 6ms/step - loss: 0.1443 - accuracy: 0.9438 - val_loss: 0.1709 - val_accuracy: 0.9353 - lr: 1.0000e-05\n",
      "Epoch 26/100\n",
      "5802/5802 [==============================] - ETA: 0s - loss: 0.1442 - accuracy: 0.9436\n",
      "Epoch 26: val_accuracy did not improve from 0.93539\n",
      "5802/5802 [==============================] - 34s 6ms/step - loss: 0.1442 - accuracy: 0.9436 - val_loss: 0.1712 - val_accuracy: 0.9348 - lr: 1.0000e-05\n",
      "Epoch 27/100\n",
      "5801/5802 [============================>.] - ETA: 0s - loss: 0.1441 - accuracy: 0.9437\n",
      "Epoch 27: val_accuracy did not improve from 0.93539\n",
      "5802/5802 [==============================] - 34s 6ms/step - loss: 0.1441 - accuracy: 0.9437 - val_loss: 0.1704 - val_accuracy: 0.9352 - lr: 1.0000e-05\n",
      "Epoch 28/100\n",
      "5797/5802 [============================>.] - ETA: 0s - loss: 0.1440 - accuracy: 0.9437\n",
      "Epoch 28: val_accuracy did not improve from 0.93539\n",
      "5802/5802 [==============================] - 34s 6ms/step - loss: 0.1441 - accuracy: 0.9436 - val_loss: 0.1710 - val_accuracy: 0.9351 - lr: 1.0000e-05\n",
      "Epoch 29/100\n",
      "5799/5802 [============================>.] - ETA: 0s - loss: 0.1441 - accuracy: 0.9437\n",
      "Epoch 29: val_accuracy did not improve from 0.93539\n",
      "Restoring model weights from the end of the best epoch: 19.\n",
      "\n",
      "Epoch 29: ReduceLROnPlateau reducing learning rate to 1.0000000656873453e-06.\n",
      "5802/5802 [==============================] - 33s 6ms/step - loss: 0.1441 - accuracy: 0.9437 - val_loss: 0.1720 - val_accuracy: 0.9354 - lr: 1.0000e-05\n",
      "Epoch 29: early stopping\n"
     ]
    }
   ],
   "source": [
    "num_epochs = 100\n",
    "\n",
    "# Train the model\n",
    "history_GlobalAveragePooling1D = model_GlobalAveragePooling1D.fit(training_padded, training_labels, epochs=num_epochs, validation_data=(testing_padded, testing_labels), callbacks=callbacks)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1451/1451 [==============================] - 3s 2ms/step - loss: 0.1699 - accuracy: 0.9354\n",
      "Test results - Loss: 0.16993598639965057 - Accuracy: 93.53872537612915%\n"
     ]
    }
   ],
   "source": [
    "# Evaluate the model\n",
    "results_GlobalAveragePooling1D = model_GlobalAveragePooling1D.evaluate(testing_padded, testing_labels)\n",
    "\n",
    "# print results\n",
    "print(f'Test results - Loss: {results_GlobalAveragePooling1D[0]} - Accuracy: {100*results_GlobalAveragePooling1D[1]}%')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "id": "7lb9p_i4Ek4L",
    "outputId": "d9ee52e8-a47b-492e-a4f7-80afa24eed24",
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot the accuracy and loss\n",
    "plot_graphs(history_GlobalAveragePooling1D, \"accuracy\")\n",
    "plot_graphs(history_GlobalAveragePooling1D, \"loss\")"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {
    "id": "sffDRy1uEk4N"
   },
   "source": [
    "# Model Embedding + Conv1D"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "id": "vbpy4arMEk4N"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_2\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " embedding_2 (Embedding)     (None, 10000, 16)         160000    \n",
      "                                                                 \n",
      " conv1d (Conv1D)             (None, 9996, 128)         10368     \n",
      "                                                                 \n",
      " global_max_pooling1d (Globa  (None, 128)              0         \n",
      " lMaxPooling1D)                                                  \n",
      "                                                                 \n",
      " dense_8 (Dense)             (None, 128)               16512     \n",
      "                                                                 \n",
      " dense_9 (Dense)             (None, 128)               16512     \n",
      "                                                                 \n",
      " dense_10 (Dense)            (None, 128)               16512     \n",
      "                                                                 \n",
      " dense_11 (Dense)            (None, 1)                 129       \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 220,033\n",
      "Trainable params: 220,033\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "kernel_size = 5\n",
    "filters = 128\n",
    "\n",
    "# Initialize the model\n",
    "model_Conv1D = tf.keras.Sequential([\n",
    "    tf.keras.layers.Embedding(vocab_size, embedding_dim, input_length=max_length),\n",
    "    tf.keras.layers.Conv1D(filters=filters, kernel_size=kernel_size, activation='relu'),\n",
    "    tf.keras.layers.GlobalMaxPooling1D(),\n",
    "    tf.keras.layers.Dense(128, activation='relu'),\n",
    "    tf.keras.layers.Dense(128, activation='relu'),\n",
    "    tf.keras.layers.Dense(128, activation='relu'),\n",
    "    tf.keras.layers.Dense(1, activation='sigmoid')\n",
    "])\n",
    "\n",
    "# Compile the model\n",
    "model_Conv1D.compile(loss='binary_crossentropy',optimizer='adam',metrics=['accuracy'])\n",
    "\n",
    "# Print the model summary\n",
    "model_Conv1D.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Difine the callbacks\n",
    "# Save the best model\n",
    "checkpoint_callback = ModelCheckpoint(filepath=\"Model Embedding + Conv1D.h5\", save_best_only=True, monitor=\"val_accuracy\", mode=\"max\", verbose=1)\n",
    "\n",
    "# Early stopping\n",
    "early_stopping_callback = EarlyStopping(monitor=\"val_accuracy\", mode=\"max\", patience=10, verbose=1, restore_best_weights=True)\n",
    "\n",
    "# Reduce learning rate\n",
    "reduce_lr_callback = ReduceLROnPlateau(monitor=\"val_loss\", factor=0.1, patience=5, verbose=1, mode=\"min\", min_delta=0.0001, cooldown=0, min_lr=0)\n",
    "\n",
    "# Callbacks list\n",
    "callbacks=[checkpoint_callback, early_stopping_callback, reduce_lr_callback]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "id": "WNv2Rk3bEk4N"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/100\n",
      "5802/5802 [==============================] - ETA: 0s - loss: 0.1804 - accuracy: 0.9309\n",
      "Epoch 1: val_accuracy improved from -inf to 0.93530, saving model to Model Embedding + Conv1D.h5\n",
      "5802/5802 [==============================] - 174s 29ms/step - loss: 0.1804 - accuracy: 0.9309 - val_loss: 0.1678 - val_accuracy: 0.9353 - lr: 0.0010\n",
      "Epoch 2/100\n",
      "5801/5802 [============================>.] - ETA: 0s - loss: 0.1314 - accuracy: 0.9509\n",
      "Epoch 2: val_accuracy improved from 0.93530 to 0.94523, saving model to Model Embedding + Conv1D.h5\n",
      "5802/5802 [==============================] - 170s 29ms/step - loss: 0.1314 - accuracy: 0.9509 - val_loss: 0.1463 - val_accuracy: 0.9452 - lr: 0.0010\n",
      "Epoch 3/100\n",
      "5801/5802 [============================>.] - ETA: 0s - loss: 0.1075 - accuracy: 0.9601\n",
      "Epoch 3: val_accuracy did not improve from 0.94523\n",
      "5802/5802 [==============================] - 170s 29ms/step - loss: 0.1075 - accuracy: 0.9601 - val_loss: 0.1657 - val_accuracy: 0.9402 - lr: 0.0010\n",
      "Epoch 4/100\n",
      "5801/5802 [============================>.] - ETA: 0s - loss: 0.0856 - accuracy: 0.9685\n",
      "Epoch 4: val_accuracy did not improve from 0.94523\n",
      "5802/5802 [==============================] - 170s 29ms/step - loss: 0.0856 - accuracy: 0.9685 - val_loss: 0.1732 - val_accuracy: 0.9369 - lr: 0.0010\n",
      "Epoch 5/100\n",
      "5801/5802 [============================>.] - ETA: 0s - loss: 0.0689 - accuracy: 0.9742\n",
      "Epoch 5: val_accuracy did not improve from 0.94523\n",
      "5802/5802 [==============================] - 171s 29ms/step - loss: 0.0689 - accuracy: 0.9742 - val_loss: 0.1986 - val_accuracy: 0.9384 - lr: 0.0010\n",
      "Epoch 6/100\n",
      "5801/5802 [============================>.] - ETA: 0s - loss: 0.0571 - accuracy: 0.9789\n",
      "Epoch 6: val_accuracy did not improve from 0.94523\n",
      "5802/5802 [==============================] - 171s 29ms/step - loss: 0.0571 - accuracy: 0.9789 - val_loss: 0.1971 - val_accuracy: 0.9381 - lr: 0.0010\n",
      "Epoch 7/100\n",
      "5801/5802 [============================>.] - ETA: 0s - loss: 0.0474 - accuracy: 0.9826\n",
      "Epoch 7: val_accuracy did not improve from 0.94523\n",
      "\n",
      "Epoch 7: ReduceLROnPlateau reducing learning rate to 0.00010000000474974513.\n",
      "5802/5802 [==============================] - 171s 29ms/step - loss: 0.0474 - accuracy: 0.9826 - val_loss: 0.2382 - val_accuracy: 0.9371 - lr: 0.0010\n",
      "Epoch 8/100\n",
      "5801/5802 [============================>.] - ETA: 0s - loss: 0.0259 - accuracy: 0.9907\n",
      "Epoch 8: val_accuracy did not improve from 0.94523\n",
      "5802/5802 [==============================] - 170s 29ms/step - loss: 0.0259 - accuracy: 0.9907 - val_loss: 0.3170 - val_accuracy: 0.9351 - lr: 1.0000e-04\n",
      "Epoch 9/100\n",
      "5801/5802 [============================>.] - ETA: 0s - loss: 0.0192 - accuracy: 0.9933\n",
      "Epoch 9: val_accuracy did not improve from 0.94523\n",
      "5802/5802 [==============================] - 171s 29ms/step - loss: 0.0192 - accuracy: 0.9933 - val_loss: 0.3824 - val_accuracy: 0.9334 - lr: 1.0000e-04\n",
      "Epoch 10/100\n",
      "5802/5802 [==============================] - ETA: 0s - loss: 0.0153 - accuracy: 0.9948\n",
      "Epoch 10: val_accuracy did not improve from 0.94523\n",
      "5802/5802 [==============================] - 171s 29ms/step - loss: 0.0153 - accuracy: 0.9948 - val_loss: 0.4618 - val_accuracy: 0.9325 - lr: 1.0000e-04\n",
      "Epoch 11/100\n",
      "5801/5802 [============================>.] - ETA: 0s - loss: 0.0129 - accuracy: 0.9956\n",
      "Epoch 11: val_accuracy did not improve from 0.94523\n",
      "5802/5802 [==============================] - 170s 29ms/step - loss: 0.0129 - accuracy: 0.9956 - val_loss: 0.5350 - val_accuracy: 0.9315 - lr: 1.0000e-04\n",
      "Epoch 12/100\n",
      "5801/5802 [============================>.] - ETA: 0s - loss: 0.0108 - accuracy: 0.9963\n",
      "Epoch 12: val_accuracy did not improve from 0.94523\n",
      "Restoring model weights from the end of the best epoch: 2.\n",
      "\n",
      "Epoch 12: ReduceLROnPlateau reducing learning rate to 1.0000000474974514e-05.\n",
      "5802/5802 [==============================] - 170s 29ms/step - loss: 0.0108 - accuracy: 0.9963 - val_loss: 0.6187 - val_accuracy: 0.9312 - lr: 1.0000e-04\n",
      "Epoch 12: early stopping\n"
     ]
    }
   ],
   "source": [
    "num_epochs = 100\n",
    "\n",
    "# Train the model\n",
    "history_Conv1D = model_Conv1D.fit(training_padded, training_labels, epochs=num_epochs, validation_data=(testing_padded, testing_labels), callbacks=callbacks)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1451/1451 [==============================] - 17s 11ms/step - loss: 0.1463 - accuracy: 0.9452\n",
      "Test results - Loss: 0.14631249010562897 - Accuracy: 94.52332258224487%\n"
     ]
    }
   ],
   "source": [
    "# Evaluate the model\n",
    "results_Conv1D = model_Conv1D.evaluate(testing_padded, testing_labels)\n",
    "\n",
    "# print results\n",
    "print(f'Test results - Loss: {results_Conv1D[0]} - Accuracy: {100*results_Conv1D[1]}%')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "id": "Pu8IIFrnEk4N"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot the accuracy and loss\n",
    "plot_graphs(history_Conv1D, \"accuracy\")\n",
    "plot_graphs(history_Conv1D, \"loss\")"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Testing Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Load all model\n",
    "model_base = tf.keras.models.load_model('Basic-Model-Embeding-Layers.h5')\n",
    "model_GlobalAveragePooling1D = tf.keras.models.load_model('Model-Basic-Global-Average-Pooling.h5')\n",
    "model_Conv1D = tf.keras.models.load_model('Model-Embedding-Conv1D.h5')"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Testing Base Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1/1 [==============================] - 0s 15ms/step\n",
      "Potential Suicide Post\n"
     ]
    }
   ],
   "source": [
    "# Define the tweet\n",
    "twt = ['I will kill myself']\n",
    "twt = tokenizer.texts_to_sequences(twt)\n",
    "twt = pad_sequences(twt, maxlen=max_length, padding=padding_type, truncating=trunc_type)\n",
    "\n",
    "# Predict the sentiment\n",
    "prediction_base = model_base.predict(twt)\n",
    "\n",
    "# Print the prediction\n",
    "if(np.argmax(prediction_base) == 0):\n",
    "    print(\"Potential Suicide Post\")\n",
    "elif (np.argmax(prediction_base) == 1):\n",
    "    print(\"Non Suicide Post\")"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Testing Basic + Global Average Pooling"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1/1 [==============================] - 0s 53ms/step\n",
      "Potential Suicide Post\n"
     ]
    }
   ],
   "source": [
    "# Testing Basic + Global Average Pooling\n",
    "prediction_Globalave = model_GlobalAveragePooling1D.predict(twt)\n",
    "\n",
    "# Print the prediction\n",
    "if(np.argmax(prediction_Globalave) == 0):\n",
    "    print(\"Potential Suicide Post\")\n",
    "elif (np.argmax(prediction_Globalave) == 1):\n",
    "    print(\"Non Suicide Post\")"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Testing Embeding + Conv1D"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1/1 [==============================] - 0s 221ms/step\n",
      "Potential Suicide Post\n"
     ]
    }
   ],
   "source": [
    "# Testing Embeding + Conv1D\n",
    "prediction_EmbedCov1D = model_Conv1D.predict(twt)\n",
    "\n",
    "# Print the prediction\n",
    "if(np.argmax(prediction_EmbedCov1D) == 0):\n",
    "    print(\"Potential Suicide Post\")\n",
    "elif (np.argmax(prediction_EmbedCov1D) == 1):\n",
    "    print(\"Non Suicide Post\")"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Comparation Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "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>Accuracy Score</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Model</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Base + CNN</th>\n",
       "      <td>0.945233</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Base Model</th>\n",
       "      <td>0.940989</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Base + Global Average Pooling</th>\n",
       "      <td>0.935387</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                               Accuracy Score\n",
       "Model                                        \n",
       "Base + CNN                           0.945233\n",
       "Base Model                           0.940989\n",
       "Base + Global Average Pooling        0.935387"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Compare model\n",
    "results=pd.DataFrame({'Model':['Base Model','Base + Global Average Pooling','Base + CNN'],\n",
    "                     'Accuracy Score':[results_base[1],results_GlobalAveragePooling1D[1],results_Conv1D[1]]})\n",
    "result_df=results.sort_values(by='Accuracy Score', ascending=False)\n",
    "result_df=result_df.set_index('Model')\n",
    "result_df"
   ]
  }
 ],
 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "provenance": []
  },
  "gpuClass": "standard",
  "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": "3.10.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}