<![CDATA[Heart disease remains one of the leading causes of the highest mortality rates worldwide, requiring a fast and accurate early detection system to minimize the risk of fatality. This study aims to test, compare, and analyze the performance of two popular machine learning methods, Support Vector Machine (SVM) and Naive Bayes, in classifying the risk of heart attacks. The research method applied is quantitative experimental, utilizing structured secondary data from the Kaggle repository, which includes 79,583 patient medical records. The data preprocessing stages involve handling missing values, feature normalization using the MinMax Scaler technique, and dataset splitting with a proportion of 80% training data and 20% testing data. The research findings indicate that the SVM architecture significantly dominates global performance, achieving an accuracy rate of 0.9847, a precision of 0.9594, and an F1-score of 0.8745. On the other hand, the Naive Bayes algorithm records the highest sensitivity (recall) value of 0.9017, compared to SVM, which only reaches 0.8034. The implications of this study confirm that although SVM is highly superior in the aggregate and accurate in suppressing false-positive rates, Naive Bayes demonstrates better characteristics for initial screening scenarios due to its high sensitivity in minimizing the risk of undetected critical patients.]]>
Copyrights © 2026
