<![CDATA[Heart failure is one of the leading causes of hospitalization and mortality, particularly among older adults. Early detection is essential to support effective clinical decision-making. This study aims to develop a heart failure prediction model using the Random Tree classification algorithm optimized with Particle Swarm Optimization (PSO) for feature selection. Random Tree was chosen for its simplicity and interpretability, while PSO was employed to identify the most relevant features and remove less important ones. The dataset was obtained from the UCI Machine Learning Repository and consists of 299 patient records with 12 clinical attributes. Model performance was evaluated using accuracy, precision, recall, and Area Under the Curve (AUC). The baseline Random Tree model achieved an accuracy of 75.58% and an AUC of 0.632. After applying PSO-based feature selection, the optimized model achieved an accuracy of 82.27% and an AUC of 0.740. These findings indicate that integrating PSO with Random Tree effectively improves heart failure prediction performance and has potential as a clinical decision-support tool]]>
Copyrights © 2026
