Dimensionality reduction approach based on modified hunger games search: case study on Parkinson’s disease phonation

Abstract Hunger Games Search (HGS) is a newly developed swarm-based algorithm inspired by the cooperative behavior of animals and their hunting strategies to find prey. However, HGS has been observed exhibit slow convergence may struggle with unbalanced exploration exploitation phases. To address these issues, this study proposes modified version called mHGS, which incorporates five techniques: (1) production operator, (2) variation control, (3) local escaping (4) transition factor, (5) foraging behavior. validate effectiveness mHGS method, 18 different benchmark datasets for dimensionality reduction are utilized, covering range sizes (small, medium, large). Additionally, two Parkinson’s disease phonation employed as real-world applications demonstrate superior capabilities proposed approach. Experimental statistical results obtained through method indicate its significant performance improvements in terms Recall, selected attribute count, Precision, F-score, accuracy when compared classical seven other well-established methods: Gradient-based optimizer (GBO), Grasshopper Optimization Algorithm (GOA), Gray Wolf Optimizer (GWO), Salp Swarm (SSA), Whale (WOA), Harris Hawks (HHO), Ant Lion (ALO).