Binary grey wolf optimizer with a novel population adaptation strategy for feature selection

Feature selection is a fundamental pre-processing step in machine learning that aims to reduce the dimensionality of dataset by selecting most effective features from original features. This process regarded as combinatorial optimization problem, and grey wolf optimizer (GWO), novel meta-heuristic algorithm, has gained popularity feature due its fast convergence speed easy implementation. In this paper, an improved binary GWO algorithm incorporating Population Adaptation strategy called PA-BGWO proposed. The takes into account characteristics problem designs three strategies. proposed includes adaptive individual update procedure enhance exploitation ability accelerate speed, head fine-tuned mechanism exert impact on each independent objective function, filter-based method ReliefF for calculating weights with dynamically adjusted mutation probabilities based ranking effectively escape local optima. Experimental comparisons several state-of-the-art methods 15 classification problems demonstrate approach can select small subset higher accuracy cases.