Improving the Binary Fish School Search Algorithm for feature selection

Abstract

The Fish School Search (FSS) algorithm has a very useful engaging mechanism to avoid the simple reactive agents of being trapped into local minima. In 2014, a binary version of the FSS was proposed and applied for feature selection. In this paper we propose some improvements in the Binary Fish School Search algorithm (BFSS). We show that the BFSS with these modifications outperformed the original BFSS. The IBFSS also outperformed other well-known swarm intelligence algorithms for feature selection purposes.