An Efficient Hybrid Volleyball Premier League and Antlion Metaheuristic Algorithm for Feature Selection

This study presents a new hybrid metaheuristic optimization algorithm that combines two different metaheuristics-the antlion metaheuristics and the binary volleyball premier league metaheuristics-to solve the problem of feature selection. The aim of this integration is to improve the learning phase of the binary Volleyball Premier League method and decrease its execution costs by incorporating an occurrence list that stores the previously obtained fitness results. This approach is compared with several metaheuristic algorithms, namely the binary volleyball premier league, antlion optimizer, ant colony optimization, and sine-cosine metaheuristics, on ten Parkinson's disease datasets. The experimental findings suggest that the hybrid metaheuristic showed improved performance in comparison to the alternative metaheuristics across 9 out of 10 datasets related to numerous measures, including fitness, accuracy, and the number of selected features. Statistical tests are also presented, along with the convergence curves that confirms the results. The application of the binary volleyball Premier League method indicates significant potential for addressing the issue of feature selection and forecasting Parkinson's disease.