Novel binary walrus optimization algorithms BWaOA and BWaOA-C with crossover operator for feature selection in high-dimensional data
Year: 2025
Type: article
Abstract: Ensuring high accuracy of data analysis and efficient exploration is now crucial due to the prevalence of redundant and irrelevant features in high-dimensional data. Feature Selection (FS) is a promising approach in machine learning to enhance classification accuracy by identifying relevant and discriminative features. In this study, we propose a novel wrapper FS approach leveraging the Walrus Optimization Algorithm (WaOA), a recent sophisticated metaheuristic algorithm known for its effective exploration capabilities and intelligence. To adapt WaOA for binary search spaces, we introduce two binary variants, BWaOA and BWaOA-C. The BWaOA variant uses S-shaped and V-shaped transfer functions, while BWaOA-C employs a crossover operator to enhance exploration capabilities and diversity in the search process. The proposed methods were tested on 30 high-dimensional datasets from diverse domains and benchmarked against seven popular FS methods including Chimpanzee Optimization Algorithm (ChOA), Binary Differential Evolution (BDE), Genetic Algorithm (GA), Grey Wolf Optimization (GWO), Harris Hawks Optimization (HHO), Particle Swarm Optimization (PSO), and Whale Optimization Algorithm (WOA). Experimental results highlight the superiority of the proposed methods in reducing irrelevant features and improving classification performance. On average, they achieved 86.33% feature reduction and 92.56% classification accuracy, outperforming the comparative FS methods in most cases.
Source: Discover Computing
Authors: Farid Ayeche, Adel Alti
Institutions: University Ferhat Abbas of Setif, Qassim University
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: gold
Estimated APC paid: $1,520
OpenAlex ID: W4415564103