Search

In Machine Learning (ML) tasks, Feature Selection (FS) methods are typically used by learning machines to filter data prior to training to lessen training complexity, thus alleviating the challenge of dimensional catastrophe. FS is a combinatorial optimization problem that can be used with meta-heuristics to better select the optimum combination of relevant features to help ML tasks provide adequate classification accuracy. White Shark Optimizer (WSO), as a swarm intelligence algorithm, has been extensively applied to address optimization problems with diverse levels of complexity. Although WSO is a promising meta-heuristic, it occasionally exhibits low solution accuracy and slow pace of convergence in addressing optimization tasks. These critical issues could be significantly curtailed by carefully balancing exploratory and exploitative mechanisms in WSO. In this, three binary variants of WSO were proposed, namely: (1) Binary Adaptive WSO (BAWSO) which assesses population distribution and estimates evolutionary status, (2) Binary Comprehensive Learning WSO (BCLWSO) which uses all the search agents' best solutions and which may act as a model for velocity updating, and (3) Binary Heterogeneous CLWSO (BHCLWSO) which splits the population into two subpopulations in order to undertake exploration and exploitation separately and without interfering with each other. Further, a Binary variant of the classical WSO (BWSO) was created alongside these variants. A thorough evaluation of these algorithms was conducted using 24 well-known datasets. The findings showed that the proposed algorithms significantly boosted the performance level of BWSO and showed that BHCLWSO outperformed other binary algorithms in classification accuracy and number of selected features. [ABSTRACT FROM AUTHOR]

The selection process is the most attractive operator in the optimization algorithms. It normally mimics the natural selection of survival of the fittest principle. When the selection is too greedy, the selection pressure will be high and therefore the search becomes biased toward exploitation. In contrast, when the selection has a tendency to be random, the selection pressure will be low and thus the exploration is observed more. The selection process in Grey Wolf Optimizer (GWO) tends to be too greedy since the search is driven by the three best solutions. In this paper, different selection methods extracted from other evolutionary algorithms (EAs) are investigated for GWO. Along with the original selection method of GWO called greedy-based selection method, five others selection methods which are proportional, tournament, universal sampling, linear rank, and random selection methods are investigated. Accordingly, six versions of GWO are proposed which are Greedy-based GWO (GGWO), Proportional-based GWO (PGWO), Tournament-based GWO (TGWO), Universal sampling-based GWO (UGWO), Linear rank-based GWO (LGWO), Random-based GWO (RGWO). The six versions are evaluated using 23 test functions circulated in the literature with different characteristics and complexity. The sensitivity analysis of the control parameters of some new versions are discussed. Interestingly, TGWO achieved the best results for almost all benchmark functions. This proves that the selection methods have a high impact on the performance of GWO. [ABSTRACT FROM AUTHOR]