In this chapter, we suggest a hybrid binary algorithm, namely, binary particle swarm optimization (PSO) with flower pollination algorithm (FPA), and call it by BPSOFPA. In BPSOFPA, PSO performs as a global search and flower pollination algorithm (FPA) conducts a fine-tuned search. We introduce the binary version of the hybridization between PSO and FPA, to solve binary problems, in particular, feature selection (FS) problem. In general, the binary algorithm relies on the so-called transfer function In this study two of the transfer functions (namely, S-shaped and V-shaped) are introduced and evaluated. We test the suggested algorithm BPSOFPA on 18 well-known benchmark UCI datasets to check its performance. The performance of our suggested algorithm is more acceptable than other pertinent works including the traditional version of the binary optimization algorithm. The results show that the suggested V-shaped family of transfer functions enhances the performance of the standard binary PSO and FPA.
CITATION STYLE
Tawhid, M. A., & Ibrahim, A. M. (2020). Hybrid binary particle swarm optimization and flower pollination algorithm based on rough set approach for feature selection problem. In Studies in Computational Intelligence (Vol. 855, pp. 249–273). Springer Verlag. https://doi.org/10.1007/978-3-030-28553-1_12
Mendeley helps you to discover research relevant for your work.