<![CDATA[Heart failure disease shows an alarming increase in global prevalence with significant clinical impact complexity. This study implements the Naive Bayes algorithm to predict heart failure risk, presenting a solution that is more computationally and interpretationally efficient than the high computationally time-consuming Random Forest or SVM with 92% accuracy. The methodological approach includes structured data preprocessing, including missing value handling, feature development, scale normalization, and dataset balancing. The application of K-Fold Cross Validation with K variations (2, 4, 5, 10) achieved optimal performance at K=4 with an accuracy of 85.1%, which enabled a reduction in the misdiagnosis rate to 14.9%. Achieving a precision of 81.1%, recall of 86.1%, and AUC-ROC of 0.914 contributed to savings in treatment costs through early identification accuracy. The system can be integrated in automated screening for efficient allocation of medical resources, resulting in significant operational savings through prioritization of high-risk patients and timely preventive interventions. Performance stability with consistent AUC-ROC (0.91-0.92) makes it a reliable foundation for clinical decision support systems that improve overall patient outcomes.]]>
Copyrights © 2025
