Heart Attack Prediction Using Neural Networks

Abstract

Cardiovascular diseases, particularly heart attacks, remain one of the leading causes of mortality worldwide. Timely prediction and intervention are critical for reducing the morbidity and mortality associated with these conditions. Machine learning techniques, particularly neural networks, have shown promise in predicting cardiovascular events based on various risk factors and medical data. This study aims to develop an accurate predictive model for identifying individuals at risk of heart attacks using neural networks.The dataset used in this study comprises a comprehensive collection of demographic information, clinical variables, and medical history of patients, obtained from various sources including hospitals, clinics, and research databases. Feature selection techniques were employed to identify the most relevant predictors contributing to the occurrence of heart attacks. The dataset was pre- processed to handle missing values, normalize features, and mitigate class imbalance issues.A feedforward neural network architecture was chosen for its ability to handle complex nonlinear relationships within the data. The neural network model was trained using a subset of the data and validated using k-fold cross-validation to assess its generalization performance. Various hyperparameters such as the number of hidden layers, neurons per layer, and activation functions were fine-tuned using grid search and randomized search techniques to optimize model performance.The results demonstrate that the neural network model achieved high accuracy and AUC-ROC score in predicting heart attacks, outperforming traditional risk prediction models. The model exhibited robustness across different evaluation metrics and demonstrated good calibration and clinical utility.