Enhancing heart disease prediction accuracy with hybrid machine learning

In this study, cardiovascular disease prediction was performed using adaptive boosting (ADA) and histogram gradient boosting (HGB) machine learning models. To improve their predictive accuracy, metaheuristic optimization algorithms, the Sea-Horse Optimizer (SHO) and the Chaos Game Optimizer (CGO), were integrated with the models. This led to the development of hybrid models: ADSH (ADA + SHO), ADCG (ADA + CGO), HGSH (HGB + SHO), and HGCG (HGB + CGO). Among them, the HGSH model achieved the highest accuracy of 0.912, outperforming the others. HGCG followed with 0.902, while the base ADA model showed lower performance with a precision of 0.840.