Integrating Deep Learning and Machine Learning for Enhanced Heart Disease Detection

Abstract

Heart disease remains a leading cause of global mortality, emphasizing the need for early and accurate diagnostic tools. This study presents an integrative framework combining machine learning (ML) and deep learning (DL) techniques for effective heart disease detection using both structured clinical data and unstructured ECG signals. Key methodologies include Chi-square feature selection, principal component analysis (PCA), and a combination of classifiers such as Support Vector Machines (SVM), Naïve Bayes, Convolutional Neural Networks (CNN), and ensemble voting models. Additionally, active learning strategies were employed to reduce labeling costs and enhance generalizability. Experimental results demonstrate that the ensemble model achieved the highest diagnostic accuracy (94.6%) and AUC (0.95), outperforming individual classifiers. The integration of interpretability tools further bridges the gap between AI predictions and clinical decision-making. Ethical considerations, data privacy, and regulatory compliance were also addressed to ensure responsible AI deployment. This comprehensive approach highlights the transformative potential of AI in improving heart disease diagnostics and lays the groundwork for future clinical integration.