<![CDATA[Heart disease is one of the leading causes of death worldwide, with approximately 17.8 million deaths reported in 2021. In Indonesia, the number of cases reached an estimated 15.5 million in 2022, highlighting the need for accurate early detection. This study aims to improve heart disease classification by integrating Correlation Feature Selection (CFS), Synthetic Minority Oversampling Technique (SMOTE), Particle Swarm Optimization (PSO), and Random Forest. CFS was applied to remove less relevant features before classification, SMOTE addressed class imbalance, and PSO optimized the Random Forest hyperparameters. The dataset consisted of 1,025 records from four heart disease datasets, which were cleaned into 302 unique instances. CFS eliminated the fasting blood sugar (fbs) attribute due to its very weak correlation with the target variable. Evaluation using 10-fold cross-validation showed that the baseline Random Forest achieved an accuracy of 82.80%, precision of 83.62%, recall of 86.07%, and f1-score of 84.46%, while Random Forest–PSO with SMOTE produced the best performance, achieving 85.11% accuracy, 84.62% precision, 89.74% recall, and 86.77% f1-score. These findings indicate that PSO provides the greatest performance improvement, while CFS simplifies the feature set before classification. The proposed hybrid model contributes by integrating feature selection, data balancing, and hyperparameter optimization into a single classification framework for more effective heart disease prediction.]]>
