Heart Disease Analysis and Prediction Using Advanced Deep Learning Approaches: A Comparative Study with Traditional Deep Learning Models

Abstract

Heart disease remains one of the leading causes of mortality worldwide, creating a significant burden on healthcare systems and emphasizing the need for early and accurate diagnosis. Conventional diagnostic approaches often rely on clinical expertise and traditional machine learning or basic deep learning techniques, which may encounter limitations in capturing complex and nonlinear relationships among medical attributes. Recent developments in advanced deep learning architectures have demonstrated promising capabilities for improving disease prediction through enhanced feature extraction, representation learning, and predictive accuracy. This study presents a comprehensive analysis and prediction framework for heart disease using advanced deep learning approaches and compares their performance with traditional deep learning models. The proposed research employs standardized heart disease datasets, followed by data preprocessing, feature engineering, normalization, and model training using Artificial Neural Networks (ANN), Deep Neural Networks (DNN), Convolutional Neural Networks (CNN), Long Short-Term Memory (LSTM), Bidirectional LSTM (Bi-LSTM), Gated Recurrent Unit (GRU), and Transformer-based architectures. Model performance is evaluated using multiple validation techniques and performance metrics, including accuracy, precision, recall, F1-score, specificity, sensitivity, Matthews Correlation Coefficient (MCC), Cohen's Kappa, and ROC-AUC. Comparative experimental analysis is conducted to identify the strengths and limitations of each model with respect to predictive performance, computational efficiency, and generalization capability. The findings are expected to demonstrate that advanced deep learning approaches provide superior predictive accuracy and robustness over traditional deep learning models, thereby supporting intelligent clinical decision-making and facilitating early diagnosis of cardiovascular diseases.