Multi-Objective Gray Wolf Optimizer with Cost-Sensitive Feature Selection for Predicting Students’ Academic Performance in College English

Feature selection is a widely utilized technique in educational data mining that aims to simplify and reduce the computational burden associated with analysis. However, previous studies have overlooked high costs involved acquiring certain types of data. In this study, we investigate application multi-objective gray wolf optimizer (GWO) cost-sensitive feature predict students’ academic performance college English, while minimizing both prediction error cost. To improve binary GWO, novel position update method mechanism for a, b, d are proposed. Additionally, adaptive mutation Pareto optimal solutions improves convergence avoids falling into local traps. The repairing duplicate expands population diversity reduces Experiments using UCI datasets demonstrate proposed algorithm outperforms existing state-of-the-art algorithms hypervolume (HV), inverted generational distance (IGD), solutions. Finally, when predicting students superiority again confirmed, as well its acquisition key features impact selection.