Optimization of Feature Selection Using the Rival Genetic Algorithm for Cardiovascular Disease Prediction

Cardiovascular Diseases (CVD) increasingly affect the global population, often going undetected until severe symptoms arise, leading to high mortality rates. Therefore, early diagnosis and prevention are crucial. Recently, innovative machine learning (ML) approaches have been developed to predict CVD onset before symptoms appear, with feature selection (FS) being a vital preprocessing step to enhance classification accuracy. This study employs the Rival Genetic Algorithm (RGA) for FS, which improves upon traditional Genetic Algorithms (GA) by incorporating a competitive strategy and new selection and crossover techniques, as well as a dynamic mutation rate to refine search behavior. The results showed that the RGA significantly enhanced ML prediction models, achieving a classification accuracy of 94.92% on the Cleveland dataset with a Multi-Layer Perceptron (MLP) model and 91.21% on the Heart Failure dataset with a K-Nearest Neighbors (KNN) model.