Improving whale optimization algorithm for feature selection with a time-varying transfer function

<p style='text-indent:20px;'>Feature selection is a valuable tool in supervised machine learning research fields, such as pattern recognition or classification problems. Feature used to eliminate irrelevant and noise features that adversely affect results. Swarm algorithms are usually feature problem; these need transfer functions change search space from continuous the discrete. However, backbone of all binary swarm algorithms. Transfer current formula cannot provide with fit balance between exploration exploitation stages. In this work, approach based on whale optimization algorithm different kinds updating techniques for time-varying proposed. To evaluate performance proposed method, three each chemical biological datasets used. The results proved BWOA-TV2 has consistency it gives rise high accuracy more congruent convergence. It worth mentioning method advance over competitor algorithms, particle (PSO) firefly (FO) commonly field.</p>