Heart Attack Analysis Prediction using ANN

Heart Attack Analysis & Prediction using Deep Learning (Artificial Neural Network)

---

Table of Contents

1. Introduction
2. Dataset
3. Installation
4. Usage
5. Model Architecture
6. Results
7. Contributors
8. License

Introduction

This project involves the analysis and prediction of heart attacks using Artificial Neural Networks (ANN). The goal is to develop a predictive model that can accurately identify the likelihood of a heart attack based on various health metrics and patient data.

Dataset

The dataset used for this project is sourced from the Heart Attack Analysis & Prediction Dataset. It includes features such as age, sex, chest pain type, resting blood pressure, serum cholesterol, fasting blood sugar, resting electrocardiographic results, maximum heart rate achieved, exercise-induced angina, ST depression induced by exercise, the slope of the peak exercise ST segment, number of major vessels, and thalassemia.

Installation

To run this project locally, follow these steps:

1. Clone the repository:

   git clone https://github.com/Islam-hady9/Heart-Attack-Analysis-Prediction-using-ANN.git

2. Navigate to the project directory:

   cd Heart-Attack-Analysis-Prediction-using-ANN

3. Install the required dependencies:

   pip install -r requirements.txt

Usage

To run the entire Jupyter Notebook for analysis and prediction, follow these steps:

1. Ensure the dataset is in the correct format and available in the project directory.
2. Start Jupyter Notebook:

   jupyter notebook

3. Open the Heart Attack Analysis & Prediction using ANN.ipynb file from the Jupyter Notebook dashboard.
4. Run all cells in the notebook to perform the data analysis, model training, and prediction steps.

Model Architecture

The ANN model is constructed using the following layers:

1. Input Layer: Corresponding to the number of features in the dataset.
2. Hidden Layers: Multiple hidden layers with ReLU activation functions.
3. Output Layer: A single neuron with a sigmoid activation function to output the probability of a heart attack.

The model is trained using the Adam optimizer and binary cross-entropy loss function.

Here is a simplified code snippet of the model architecture:

# Set the random seed for reproducibility
tf.random.set_seed(42)
# Define the number of folds for KFold cross-validation
n_splits = 5
kfold = KFold(n_splits=n_splits, shuffle=True, random_state=42)
# Prepare to collect scores and histories
accuracies = []
all_histories = []

# KFold Cross Validation
for train_index, val_index in kfold.split(X_train):
    # Split data
    X_train_kfold, X_val_kfold = X_train[train_index], X_train[val_index]
    y_train_kfold, y_val_kfold = y_train[train_index], y_train[val_index]
    
    # Create a new instance of the model (to reinitialize weights)
    ANN_model = tf.keras.Sequential([
        tf.keras.layers.Dense(32, activation="relu", input_shape=(X_train.shape[1],)),
        tf.keras.layers.Dense(16, activation="relu"),
        tf.keras.layers.Dense(1, activation="sigmoid")                                  
    ])
    
    # Compile the model
    ANN_model.compile(loss="binary_crossentropy",
                  optimizer=tf.keras.optimizers.Adam(),
                  metrics=["accuracy"])
    
    # Early stopping callback
    early_stopping = tf.keras.callbacks.EarlyStopping(
        monitor='val_loss',  
        patience=10,  
        restore_best_weights=True  
    )
    
    # Fit the model
    history = ANN_model.fit(X_train_kfold, y_train_kfold, 
                        epochs=100, 
                        validation_data=(X_val_kfold, y_val_kfold), 
                        callbacks=[early_stopping], 
                        verbose=0)  # Set verbose to 0 to reduce output
    
    # Collect the history from each fold
    all_histories.append(history)
    
    # Evaluate the model on the validation set
    scores = ANN_model.evaluate(X_val_kfold, y_val_kfold, verbose=0)
    accuracies.append(scores[1])  # Assume that the accuracy is the second metric
# Print the accuracy for each fold
print("Accuracy for each fold:", accuracies)
# Print the average accuracy
print("Average accuracy:", np.mean(accuracies))

Model Architecture Diagram

Results

The model achieves an accuracy of 88.5% on the test set, with a precision of 87% and a recall of 87%.

Confusion Matrix Visualization

Plots Training and Validation Accuracy and Loss for Each Fold on Shared Plots

Contributors

• Islam Abd_Elhady Hassanein (Project Lead)
• Enas Ragab Abdel_Latif
• Mariam Tarek Saad

License

This project is licensed under the MIT License. See the LICENSE file for more details.