부산대학교 도서관

Cardiovascular diseases pose a substantial global health burden. Therefore, it is crucial to have a timely and accurate prediction of such diseases for an effective healthcare intervention. This research comprehensively analyze the effectiveness of classical machine learning and ensemble learning techniques in predicting heart disease. On a diverse dataset with clinical, demographic, and lifestyle factors, the performance of these models is empirically assessed using a variety of performance metrics, such as cross-validation mean accuracy, precision, recall, F1 score, and ROC - AUC Score. Through this comprehensive analysis, this research concludes that ensemble learning techniques outperform conventional machine learning classifiers in terms of accuracy and robustness. Gradient Boost has the best overall cross-validation accuracy score of 88.69% and recall of 89.71%. In summary, this study aims to enhance cardiovascular disease management and reducing associated mortality rates.