Abstract
Many approaches have been tried out for feature selection, which is aimed at finding a minimal subset of the original features with predetermined targets. However, a complete search isn’t feasible for even medium-sized datasets and it has been proved that finding a minimal subset of the features is a NP-hard problem. Rough set theory is one of the effective methods to feature selection, and gravitational search algorithm (GSA), which has a flexible and well-balanced mechanism to enhance exploration and exploitation, has been successfully applied in many difficult problems. In this paper, a novel approach, called FSRG, for feature selection based on rough set and GSA is proposed, and 5 UCI datasets are used as an illustrated example. The results demonstrate that FSRG is an efficient method for feature selection.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have