diabetic-retinopathy-classification.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "cad9807a-a670-475e-93e3-f94d90f710e7",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import tensorflow as tf\n",
    "from tensorflow.keras.models import Sequential\n",
    "from tensorflow.keras.layers import Dense\n",
    "from tensorflow.keras.optimizers import Adam\n",
    "from tensorflow.keras.callbacks import EarlyStopping\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from imblearn.over_sampling import SMOTE\n",
    "from ucimlrepo import fetch_ucirepo\n",
    "\n",
    "# Set the style for seaborn\n",
    "sns.set(style=\"whitegrid\", palette=\"muted\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "5f8dbd6e-a2a4-4d7f-8a29-fc35291d4745",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "   quality  pre_screening  ma1  ma2  ma3  ma4  ma5  ma6   exudate1   exudate2  \\\n",
      "0        1              1   22   22   22   19   18   14  49.895756  17.775994   \n",
      "1        1              1   24   24   22   18   16   13  57.709936  23.799994   \n",
      "2        1              1   62   60   59   54   47   33  55.831441  27.993933   \n",
      "3        1              1   55   53   53   50   43   31  40.467228  18.445954   \n",
      "4        1              1   44   44   44   41   39   27  18.026254   8.570709   \n",
      "\n",
      "    exudate3  exudate5  exudate6  exudate7  exudate8  \\\n",
      "0   5.270920  0.018632  0.006864  0.003923  0.003923   \n",
      "1   3.325423  0.003903  0.003903  0.003903  0.003903   \n",
      "2  12.687485  1.393889  0.373252  0.041817  0.007744   \n",
      "3   9.118901  0.840261  0.272434  0.007653  0.001531   \n",
      "4   0.410381  0.000000  0.000000  0.000000  0.000000   \n",
      "\n",
      "   macula_opticdisc_distance  opticdisc_diameter  am_fm_classification  \n",
      "0                   0.486903            0.100025                     1  \n",
      "1                   0.520908            0.144414                     0  \n",
      "2                   0.530904            0.128548                     0  \n",
      "3                   0.483284            0.114790                     0  \n",
      "4                   0.475935            0.123572                     0  \n",
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 1151 entries, 0 to 1150\n",
      "Data columns (total 18 columns):\n",
      " #   Column                     Non-Null Count  Dtype  \n",
      "---  ------                     --------------  -----  \n",
      " 0   quality                    1151 non-null   int64  \n",
      " 1   pre_screening              1151 non-null   int64  \n",
      " 2   ma1                        1151 non-null   int64  \n",
      " 3   ma2                        1151 non-null   int64  \n",
      " 4   ma3                        1151 non-null   int64  \n",
      " 5   ma4                        1151 non-null   int64  \n",
      " 6   ma5                        1151 non-null   int64  \n",
      " 7   ma6                        1151 non-null   int64  \n",
      " 8   exudate1                   1151 non-null   float64\n",
      " 9   exudate2                   1151 non-null   float64\n",
      " 10  exudate3                   1151 non-null   float64\n",
      " 11  exudate5                   1151 non-null   float64\n",
      " 12  exudate6                   1151 non-null   float64\n",
      " 13  exudate7                   1151 non-null   float64\n",
      " 14  exudate8                   1151 non-null   float64\n",
      " 15  macula_opticdisc_distance  1151 non-null   float64\n",
      " 16  opticdisc_diameter         1151 non-null   float64\n",
      " 17  am_fm_classification       1151 non-null   int64  \n",
      "dtypes: float64(9), int64(9)\n",
      "memory usage: 162.0 KB\n",
      "None\n",
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 1151 entries, 0 to 1150\n",
      "Data columns (total 1 columns):\n",
      " #   Column  Non-Null Count  Dtype\n",
      "---  ------  --------------  -----\n",
      " 0   Class   1151 non-null   int64\n",
      "dtypes: int64(1)\n",
      "memory usage: 9.1 KB\n",
      "None\n"
     ]
    }
   ],
   "source": [
    "# Fetch dataset\n",
    "diabetic_retinopathy_debrecen = fetch_ucirepo(id=329)\n",
    "\n",
    "# Data (as pandas dataframes)\n",
    "X = diabetic_retinopathy_debrecen.data.features\n",
    "y = diabetic_retinopathy_debrecen.data.targets\n",
    "\n",
    "# Drop duplicate columns\n",
    "X = X.loc[:, ~X.columns.duplicated()]\n",
    "\n",
    "# Display first few columns of the features\n",
    "print(X.head())\n",
    "\n",
    "# Display info about the dataset\n",
    "print(X.info())\n",
    "print(y.info())\n",
    "\n",
    "# Standardize features\n",
    "scaler = StandardScaler()\n",
    "X_standardized = scaler.fit_transform(X)\n",
    "\n",
    "# Handle class imbalance using SMOTE\n",
    "smote = SMOTE(random_state=42)\n",
    "X_resampled, y_resampled = smote.fit_resample(X_standardized, y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "25b85595-525c-4a36-9774-3a3f31a3891c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "X_train shape: (977, 18)\n",
      "X_val shape: (122, 18)\n",
      "X_test shape: (123, 18)\n",
      "y_train shape: (977, 1)\n",
      "y_val shape: (122, 1)\n",
      "y_test shape: (123, 1)\n",
      "\n",
      "Class distribution in training set:\n",
      "Class\n",
      "1    489\n",
      "0    488\n",
      "Name: count, dtype: int64\n",
      "\n",
      "Class distribution in validation set:\n",
      "Class\n",
      "0    61\n",
      "1    61\n",
      "Name: count, dtype: int64\n",
      "\n",
      "Class distribution in test set:\n",
      "Class\n",
      "0    62\n",
      "1    61\n",
      "Name: count, dtype: int64\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1000x600 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1000x600 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1000x600 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Initial train/test/validation split\n",
    "X_train, X_temp, y_train, y_temp = train_test_split(\n",
    "    X_resampled, y_resampled, \n",
    "    test_size=0.2, \n",
    "    random_state=42, \n",
    "    stratify=y_resampled\n",
    ")\n",
    "X_val, X_test, y_val, y_test = train_test_split(\n",
    "    X_temp, y_temp, \n",
    "    test_size=0.5,  # 50% of the temporary set\n",
    "    random_state=42, \n",
    "    stratify=y_temp\n",
    ")\n",
    "\n",
    "# Create dataframes for each set with the target variable\n",
    "df_train = pd.DataFrame(X_train, columns=X.columns).copy()\n",
    "df_train['Class'] = y_train.values\n",
    "\n",
    "df_val = pd.DataFrame(X_val, columns=X.columns).copy()\n",
    "df_val['Class'] = y_val.values\n",
    "\n",
    "df_test = pd.DataFrame(X_test, columns=X.columns).copy()\n",
    "df_test['Class'] = y_test.values\n",
    "\n",
    "# Display the shapes of the splits\n",
    "print(f'X_train shape: {X_train.shape}')\n",
    "print(f'X_val shape: {X_val.shape}')\n",
    "print(f'X_test shape: {X_test.shape}')\n",
    "print(f'y_train shape: {y_train.shape}')\n",
    "print(f'y_val shape: {y_val.shape}')\n",
    "print(f'y_test shape: {y_test.shape}')\n",
    "\n",
    "# Display class distributions\n",
    "print(\"\\nClass distribution in training set:\")\n",
    "print(df_train['Class'].value_counts())\n",
    "\n",
    "print(\"\\nClass distribution in validation set:\")\n",
    "print(df_val['Class'].value_counts())\n",
    "\n",
    "print(\"\\nClass distribution in test set:\")\n",
    "print(df_test['Class'].value_counts())\n",
    "\n",
    "# Function to plot class distribution\n",
    "def plot_class_distribution(df, title):\n",
    "    plt.figure(figsize=(10, 6))\n",
    "    sns.countplot(x='Class', data=df)\n",
    "    plt.title(title)\n",
    "    plt.xlabel('Class')\n",
    "    plt.ylabel('Count')\n",
    "    plt.show()\n",
    "\n",
    "# Plot class distributions\n",
    "plot_class_distribution(df_train, 'Class Distribution in Training Set')\n",
    "plot_class_distribution(df_val, 'Class Distribution in Validation Set')\n",
    "plot_class_distribution(df_test, 'Class Distribution in Test Set')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "8863829b-7979-4711-ae71-a7992227f20f",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Build the updated neural network model with increased complexity\n",
    "model = Sequential([\n",
    "    Dense(128, activation='relu', input_shape=(X_train.shape[1],)),  # Increased units\n",
    "    Dense(64, activation='relu'),  # Increased units\n",
    "    Dense(32, activation='relu'),  # Additional hidden layer\n",
    "    Dense(1, activation='sigmoid')  # Use 'softmax' if more than two classes\n",
    "])\n",
    "\n",
    "# Compile the model with a slower learning rate\n",
    "optimizer = Adam(learning_rate=0.00003)  # Slower learning rate\n",
    "model.compile(optimizer=optimizer,\n",
    "              loss='binary_crossentropy',  # Use 'categorical_crossentropy' for multiple classes\n",
    "              metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "03340714-5cc9-4c9c-a895-991306e17139",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/200\n",
      "31/31 [==============================] - 1s 9ms/step - loss: 0.7408 - accuracy: 0.5005 - val_loss: 0.7473 - val_accuracy: 0.4836\n",
      "Epoch 2/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.7262 - accuracy: 0.4985 - val_loss: 0.7352 - val_accuracy: 0.4590\n",
      "Epoch 3/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.7134 - accuracy: 0.4964 - val_loss: 0.7249 - val_accuracy: 0.4590\n",
      "Epoch 4/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.7030 - accuracy: 0.4954 - val_loss: 0.7157 - val_accuracy: 0.4754\n",
      "Epoch 5/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.6938 - accuracy: 0.5261 - val_loss: 0.7094 - val_accuracy: 0.5000\n",
      "Epoch 6/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.6865 - accuracy: 0.5445 - val_loss: 0.7033 - val_accuracy: 0.4918\n",
      "Epoch 7/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.6799 - accuracy: 0.5640 - val_loss: 0.6991 - val_accuracy: 0.5164\n",
      "Epoch 8/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.6747 - accuracy: 0.5834 - val_loss: 0.6950 - val_accuracy: 0.5246\n",
      "Epoch 9/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.6699 - accuracy: 0.5906 - val_loss: 0.6913 - val_accuracy: 0.5246\n",
      "Epoch 10/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.6656 - accuracy: 0.5977 - val_loss: 0.6880 - val_accuracy: 0.5246\n",
      "Epoch 11/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.6616 - accuracy: 0.6039 - val_loss: 0.6855 - val_accuracy: 0.5328\n",
      "Epoch 12/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.6582 - accuracy: 0.6111 - val_loss: 0.6828 - val_accuracy: 0.5410\n",
      "Epoch 13/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.6549 - accuracy: 0.6172 - val_loss: 0.6807 - val_accuracy: 0.5410\n",
      "Epoch 14/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.6519 - accuracy: 0.6264 - val_loss: 0.6787 - val_accuracy: 0.5410\n",
      "Epoch 15/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.6492 - accuracy: 0.6264 - val_loss: 0.6773 - val_accuracy: 0.5656\n",
      "Epoch 16/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.6467 - accuracy: 0.6233 - val_loss: 0.6757 - val_accuracy: 0.5738\n",
      "Epoch 17/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.6443 - accuracy: 0.6254 - val_loss: 0.6740 - val_accuracy: 0.5820\n",
      "Epoch 18/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.6420 - accuracy: 0.6233 - val_loss: 0.6733 - val_accuracy: 0.5820\n",
      "Epoch 19/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.6399 - accuracy: 0.6244 - val_loss: 0.6715 - val_accuracy: 0.5820\n",
      "Epoch 20/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.6378 - accuracy: 0.6264 - val_loss: 0.6703 - val_accuracy: 0.5820\n",
      "Epoch 21/200\n",
      "31/31 [==============================] - 0s 5ms/step - loss: 0.6356 - accuracy: 0.6274 - val_loss: 0.6690 - val_accuracy: 0.5820\n",
      "Epoch 22/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.6337 - accuracy: 0.6264 - val_loss: 0.6672 - val_accuracy: 0.5738\n",
      "Epoch 23/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.6317 - accuracy: 0.6315 - val_loss: 0.6659 - val_accuracy: 0.5738\n",
      "Epoch 24/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.6299 - accuracy: 0.6325 - val_loss: 0.6646 - val_accuracy: 0.5738\n",
      "Epoch 25/200\n",
      "31/31 [==============================] - 0s 5ms/step - loss: 0.6280 - accuracy: 0.6325 - val_loss: 0.6633 - val_accuracy: 0.5738\n",
      "Epoch 26/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.6261 - accuracy: 0.6346 - val_loss: 0.6620 - val_accuracy: 0.5820\n",
      "Epoch 27/200\n",
      "31/31 [==============================] - 0s 5ms/step - loss: 0.6244 - accuracy: 0.6366 - val_loss: 0.6606 - val_accuracy: 0.5902\n",
      "Epoch 28/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.6225 - accuracy: 0.6366 - val_loss: 0.6591 - val_accuracy: 0.5902\n",
      "Epoch 29/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.6208 - accuracy: 0.6397 - val_loss: 0.6578 - val_accuracy: 0.5984\n",
      "Epoch 30/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.6192 - accuracy: 0.6407 - val_loss: 0.6560 - val_accuracy: 0.5984\n",
      "Epoch 31/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.6173 - accuracy: 0.6418 - val_loss: 0.6550 - val_accuracy: 0.5984\n",
      "Epoch 32/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.6157 - accuracy: 0.6459 - val_loss: 0.6533 - val_accuracy: 0.5984\n",
      "Epoch 33/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.6139 - accuracy: 0.6479 - val_loss: 0.6520 - val_accuracy: 0.5984\n",
      "Epoch 34/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.6123 - accuracy: 0.6510 - val_loss: 0.6502 - val_accuracy: 0.6148\n",
      "Epoch 35/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.6106 - accuracy: 0.6592 - val_loss: 0.6489 - val_accuracy: 0.6148\n",
      "Epoch 36/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.6090 - accuracy: 0.6571 - val_loss: 0.6475 - val_accuracy: 0.6148\n",
      "Epoch 37/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.6074 - accuracy: 0.6581 - val_loss: 0.6458 - val_accuracy: 0.6230\n",
      "Epoch 38/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.6058 - accuracy: 0.6602 - val_loss: 0.6441 - val_accuracy: 0.6311\n",
      "Epoch 39/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.6044 - accuracy: 0.6633 - val_loss: 0.6423 - val_accuracy: 0.6393\n",
      "Epoch 40/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.6027 - accuracy: 0.6653 - val_loss: 0.6413 - val_accuracy: 0.6311\n",
      "Epoch 41/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.6012 - accuracy: 0.6653 - val_loss: 0.6396 - val_accuracy: 0.6393\n",
      "Epoch 42/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5996 - accuracy: 0.6694 - val_loss: 0.6379 - val_accuracy: 0.6393\n",
      "Epoch 43/200\n",
      "31/31 [==============================] - 0s 5ms/step - loss: 0.5981 - accuracy: 0.6704 - val_loss: 0.6364 - val_accuracy: 0.6230\n",
      "Epoch 44/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5965 - accuracy: 0.6766 - val_loss: 0.6352 - val_accuracy: 0.6393\n",
      "Epoch 45/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5950 - accuracy: 0.6827 - val_loss: 0.6337 - val_accuracy: 0.6393\n",
      "Epoch 46/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5935 - accuracy: 0.6807 - val_loss: 0.6322 - val_accuracy: 0.6475\n",
      "Epoch 47/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5922 - accuracy: 0.6817 - val_loss: 0.6308 - val_accuracy: 0.6475\n",
      "Epoch 48/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5907 - accuracy: 0.6807 - val_loss: 0.6295 - val_accuracy: 0.6557\n",
      "Epoch 49/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5892 - accuracy: 0.6817 - val_loss: 0.6287 - val_accuracy: 0.6639\n",
      "Epoch 50/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5879 - accuracy: 0.6807 - val_loss: 0.6267 - val_accuracy: 0.6639\n",
      "Epoch 51/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5866 - accuracy: 0.6817 - val_loss: 0.6263 - val_accuracy: 0.6639\n",
      "Epoch 52/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5849 - accuracy: 0.6899 - val_loss: 0.6242 - val_accuracy: 0.6721\n",
      "Epoch 53/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5835 - accuracy: 0.6847 - val_loss: 0.6228 - val_accuracy: 0.6721\n",
      "Epoch 54/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5821 - accuracy: 0.6888 - val_loss: 0.6215 - val_accuracy: 0.6557\n",
      "Epoch 55/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5809 - accuracy: 0.6868 - val_loss: 0.6209 - val_accuracy: 0.6639\n",
      "Epoch 56/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5794 - accuracy: 0.6929 - val_loss: 0.6193 - val_accuracy: 0.6557\n",
      "Epoch 57/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5782 - accuracy: 0.6940 - val_loss: 0.6180 - val_accuracy: 0.6557\n",
      "Epoch 58/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5769 - accuracy: 0.7073 - val_loss: 0.6158 - val_accuracy: 0.6557\n",
      "Epoch 59/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5757 - accuracy: 0.7021 - val_loss: 0.6151 - val_accuracy: 0.6557\n",
      "Epoch 60/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5743 - accuracy: 0.7052 - val_loss: 0.6135 - val_accuracy: 0.6557\n",
      "Epoch 61/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5731 - accuracy: 0.7052 - val_loss: 0.6124 - val_accuracy: 0.6557\n",
      "Epoch 62/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5718 - accuracy: 0.7062 - val_loss: 0.6118 - val_accuracy: 0.6557\n",
      "Epoch 63/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5706 - accuracy: 0.7093 - val_loss: 0.6103 - val_accuracy: 0.6639\n",
      "Epoch 64/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5695 - accuracy: 0.7103 - val_loss: 0.6089 - val_accuracy: 0.6639\n",
      "Epoch 65/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5682 - accuracy: 0.7103 - val_loss: 0.6085 - val_accuracy: 0.6639\n",
      "Epoch 66/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5672 - accuracy: 0.7114 - val_loss: 0.6070 - val_accuracy: 0.6639\n",
      "Epoch 67/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5659 - accuracy: 0.7114 - val_loss: 0.6056 - val_accuracy: 0.6639\n",
      "Epoch 68/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5645 - accuracy: 0.7165 - val_loss: 0.6049 - val_accuracy: 0.6639\n",
      "Epoch 69/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5634 - accuracy: 0.7144 - val_loss: 0.6042 - val_accuracy: 0.6639\n",
      "Epoch 70/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5623 - accuracy: 0.7155 - val_loss: 0.6033 - val_accuracy: 0.6721\n",
      "Epoch 71/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5612 - accuracy: 0.7175 - val_loss: 0.6018 - val_accuracy: 0.6639\n",
      "Epoch 72/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5603 - accuracy: 0.7185 - val_loss: 0.6011 - val_accuracy: 0.6639\n",
      "Epoch 73/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5589 - accuracy: 0.7206 - val_loss: 0.6001 - val_accuracy: 0.6721\n",
      "Epoch 74/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5586 - accuracy: 0.7216 - val_loss: 0.5983 - val_accuracy: 0.6639\n",
      "Epoch 75/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5566 - accuracy: 0.7206 - val_loss: 0.5979 - val_accuracy: 0.6557\n",
      "Epoch 76/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5556 - accuracy: 0.7226 - val_loss: 0.5973 - val_accuracy: 0.6557\n",
      "Epoch 77/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5552 - accuracy: 0.7216 - val_loss: 0.5959 - val_accuracy: 0.6557\n",
      "Epoch 78/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5534 - accuracy: 0.7226 - val_loss: 0.5956 - val_accuracy: 0.6557\n",
      "Epoch 79/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5524 - accuracy: 0.7257 - val_loss: 0.5946 - val_accuracy: 0.6557\n",
      "Epoch 80/200\n",
      "31/31 [==============================] - 0s 5ms/step - loss: 0.5513 - accuracy: 0.7236 - val_loss: 0.5941 - val_accuracy: 0.6639\n",
      "Epoch 81/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5504 - accuracy: 0.7267 - val_loss: 0.5934 - val_accuracy: 0.6639\n",
      "Epoch 82/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5494 - accuracy: 0.7257 - val_loss: 0.5917 - val_accuracy: 0.6639\n",
      "Epoch 83/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5484 - accuracy: 0.7247 - val_loss: 0.5910 - val_accuracy: 0.6639\n",
      "Epoch 84/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5472 - accuracy: 0.7257 - val_loss: 0.5904 - val_accuracy: 0.6639\n",
      "Epoch 85/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5465 - accuracy: 0.7277 - val_loss: 0.5892 - val_accuracy: 0.6639\n",
      "Epoch 86/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5456 - accuracy: 0.7288 - val_loss: 0.5895 - val_accuracy: 0.6639\n",
      "Epoch 87/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5442 - accuracy: 0.7298 - val_loss: 0.5885 - val_accuracy: 0.6639\n",
      "Epoch 88/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5434 - accuracy: 0.7277 - val_loss: 0.5878 - val_accuracy: 0.6639\n",
      "Epoch 89/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5426 - accuracy: 0.7308 - val_loss: 0.5857 - val_accuracy: 0.6721\n",
      "Epoch 90/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5413 - accuracy: 0.7318 - val_loss: 0.5862 - val_accuracy: 0.6721\n",
      "Epoch 91/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5404 - accuracy: 0.7308 - val_loss: 0.5846 - val_accuracy: 0.6721\n",
      "Epoch 92/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5396 - accuracy: 0.7308 - val_loss: 0.5844 - val_accuracy: 0.6721\n",
      "Epoch 93/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5387 - accuracy: 0.7329 - val_loss: 0.5845 - val_accuracy: 0.6803\n",
      "Epoch 94/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5375 - accuracy: 0.7349 - val_loss: 0.5836 - val_accuracy: 0.6721\n",
      "Epoch 95/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5368 - accuracy: 0.7247 - val_loss: 0.5829 - val_accuracy: 0.6639\n",
      "Epoch 96/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5363 - accuracy: 0.7267 - val_loss: 0.5819 - val_accuracy: 0.6803\n",
      "Epoch 97/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5350 - accuracy: 0.7318 - val_loss: 0.5811 - val_accuracy: 0.6803\n",
      "Epoch 98/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5340 - accuracy: 0.7288 - val_loss: 0.5812 - val_accuracy: 0.6721\n",
      "Epoch 99/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5333 - accuracy: 0.7339 - val_loss: 0.5799 - val_accuracy: 0.6803\n",
      "Epoch 100/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5324 - accuracy: 0.7349 - val_loss: 0.5798 - val_accuracy: 0.6721\n",
      "Epoch 101/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5315 - accuracy: 0.7359 - val_loss: 0.5794 - val_accuracy: 0.6803\n",
      "Epoch 102/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5305 - accuracy: 0.7359 - val_loss: 0.5790 - val_accuracy: 0.6721\n",
      "Epoch 103/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5300 - accuracy: 0.7349 - val_loss: 0.5789 - val_accuracy: 0.6721\n",
      "Epoch 104/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5292 - accuracy: 0.7380 - val_loss: 0.5777 - val_accuracy: 0.6721\n",
      "Epoch 105/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5285 - accuracy: 0.7339 - val_loss: 0.5771 - val_accuracy: 0.6967\n",
      "Epoch 106/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5271 - accuracy: 0.7390 - val_loss: 0.5764 - val_accuracy: 0.6721\n",
      "Epoch 107/200\n",
      "31/31 [==============================] - 0s 5ms/step - loss: 0.5264 - accuracy: 0.7390 - val_loss: 0.5764 - val_accuracy: 0.6721\n",
      "Epoch 108/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5256 - accuracy: 0.7390 - val_loss: 0.5755 - val_accuracy: 0.6721\n",
      "Epoch 109/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5247 - accuracy: 0.7400 - val_loss: 0.5758 - val_accuracy: 0.6721\n",
      "Epoch 110/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5242 - accuracy: 0.7390 - val_loss: 0.5745 - val_accuracy: 0.6885\n",
      "Epoch 111/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5235 - accuracy: 0.7380 - val_loss: 0.5748 - val_accuracy: 0.6721\n",
      "Epoch 112/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5224 - accuracy: 0.7421 - val_loss: 0.5744 - val_accuracy: 0.6803\n",
      "Epoch 113/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5218 - accuracy: 0.7431 - val_loss: 0.5741 - val_accuracy: 0.6803\n",
      "Epoch 114/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5208 - accuracy: 0.7410 - val_loss: 0.5723 - val_accuracy: 0.6803\n",
      "Epoch 115/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5200 - accuracy: 0.7431 - val_loss: 0.5724 - val_accuracy: 0.6803\n",
      "Epoch 116/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5192 - accuracy: 0.7410 - val_loss: 0.5720 - val_accuracy: 0.6803\n",
      "Epoch 117/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5185 - accuracy: 0.7410 - val_loss: 0.5718 - val_accuracy: 0.6803\n",
      "Epoch 118/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5177 - accuracy: 0.7441 - val_loss: 0.5722 - val_accuracy: 0.6721\n",
      "Epoch 119/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5169 - accuracy: 0.7451 - val_loss: 0.5710 - val_accuracy: 0.6803\n",
      "Epoch 120/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5162 - accuracy: 0.7421 - val_loss: 0.5712 - val_accuracy: 0.6721\n",
      "Epoch 121/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5153 - accuracy: 0.7441 - val_loss: 0.5711 - val_accuracy: 0.6721\n",
      "Epoch 122/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5149 - accuracy: 0.7441 - val_loss: 0.5697 - val_accuracy: 0.6721\n",
      "Epoch 123/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5140 - accuracy: 0.7441 - val_loss: 0.5699 - val_accuracy: 0.6803\n",
      "Epoch 124/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5132 - accuracy: 0.7431 - val_loss: 0.5699 - val_accuracy: 0.6803\n",
      "Epoch 125/200\n",
      "31/31 [==============================] - 0s 4ms/step - loss: 0.5127 - accuracy: 0.7431 - val_loss: 0.5697 - val_accuracy: 0.6803\n"
     ]
    }
   ],
   "source": [
    "# Define the early stopping callback with patience of 3\n",
    "early_stopping = EarlyStopping(\n",
    "    monitor='val_loss',  # or 'val_accuracy'\n",
    "    patience=3,  # Number of epochs to wait for improvement\n",
    "    restore_best_weights=True\n",
    ")\n",
    "\n",
    "# Train the model with early stopping\n",
    "history = model.fit(\n",
    "    X_train, y_train,\n",
    "    epochs=200,  # Increased number of epochs\n",
    "    batch_size=32,  # Adjust the batch size as needed\n",
    "    validation_data=(X_val, y_val),\n",
    "    callbacks=[early_stopping]  # Include the early stopping callback\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "62c95602-bc1c-4a69-b7af-d84ce3e88bc4",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "4/4 [==============================] - 0s 3ms/step - loss: 0.5084 - accuracy: 0.7886\n",
      "Test Loss: 0.5084152817726135\n",
      "Test Accuracy: 0.7886179089546204\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1200x600 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Evaluate the model on the test set\n",
    "test_loss, test_accuracy = model.evaluate(X_test, y_test)\n",
    "print(f'Test Loss: {test_loss}')\n",
    "print(f'Test Accuracy: {test_accuracy}')\n",
    "\n",
    "# Plot training history\n",
    "plt.figure(figsize=(12, 6))\n",
    "plt.subplot(1, 2, 1)\n",
    "plt.plot(history.history['accuracy'])\n",
    "plt.plot(history.history['val_accuracy'])\n",
    "plt.title('Model Accuracy')\n",
    "plt.xlabel('Epoch')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.legend(['Train', 'Val'], loc='upper left')\n",
    "\n",
    "plt.subplot(1, 2, 2)\n",
    "plt.plot(history.history['loss'])\n",
    "plt.plot(history.history['val_loss'])\n",
    "plt.title('Model Loss')\n",
    "plt.xlabel('Epoch')\n",
    "plt.ylabel('Loss')\n",
    "plt.legend(['Train', 'Val'], loc='upper left')\n",
    "\n",
    "plt.show()"
   ]
  }
 ],
 "metadata": {
  "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": "3.10.14"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}