This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Feature selection is a critical preprocessing step in machine learning, aimed at identifying relevant features from high-dimensional datasets to improve model performance and reduce computational cost. Due to its Nondeterministic Polynomial (NP)-hard nature, metaheuristic algorithms have gained prominence for efficiently navigating the vast search space. This review examines approximately 150 metaheuristic algorithms developed or refined between 2020 and 2025, categorized into evolutionary, physics-based, human-social, and swarm intelligence approaches. Swarm intelligence algorithms dominate recent advances, comprising 55% of the surveyed methods, reflecting their scalability and effectiveness in complex domains such as healthcare and cybersecurity. The review highlights algorithmic trends, including hybridization, chaos-based diversity enhancement, and multi-objective optimization, and proposes future directions focused on adaptive, interpretable, and Artificial Intelligence (AI)-integrated frameworks.
Null