A binary cuckoo search and its application for feature selection

In classification problems, it is common to find datasets with a large amount of features, some of theses features may be considered as noisy. In this context, one of the most used strategies to deal with this problem is to perform a feature selection process in order to build a subset of features that can better represents the dataset. As feature selection can be modeled as an optimization problem, several studies have to attempted to use nature-inspired optimization techniques due to their large generalization capabilities. In this chapter, we use the Cuckoo Search (CS) algorithm in the context of feature selection tasks. For this purpose, we present a binary version of the Cuckoo Search, namely BCS, as well as we evaluate it with different transfer functions that map continuous solutions to binary ones. Additionally, the Optimum-Path Forest classifier accuracy is used as the fitness function. We conducted simulations comparing BCS with binary versions of the Bat Algorithm, Firefly Algorithm and Particle Swarm Optimization. BCS has obtained reasonable results when we consider the compared techniques for feature selection purposes. © 2014 Springer International Publishing Switzerland.