<![CDATA[Cardiovascular health is vital, with heart disease, particularly Coronary Heart Disease (CHD), being a significant health concern in Indonesia. The 2023 Indonesian Health Survey reported 877,531 cases of heart disease. Traditional CHD diagnosis is often costly and invasive. Therefore, machine learning-based classification has emerged as a promising alternative for enhancing the accuracy and efficiency of detection. This study aims to predict CHD using a hybrid approach combining the Naïve Bayes algorithm with the Bagging ensemble method. Naïve Bayes was selected for its computational efficiency and effectiveness with high-dimensional data, while Bagging was employed to mitigate its inherent weaknesses by reducing variance and increasing prediction stability. The CRISP-DM methodology was applied to a secondary dataset of 462 rows from Kaggle. The research process included data preprocessing, method implementation, and evaluation using a confusion matrix. Results show the Bagging method with n=2 estimators achieved optimal performance, with 76.34% accuracy, 65.00% precision, and an f1-score of 70.27%. This study demonstrates that ensemble techniques can effectively improve the accuracy and stability of CHD prediction models, offering a reliable and low-cost solution for initial screening.]]>
Copyrights © 2025
