Feature Selection with Chaotic Hybrid Artificial Bee Colony Algorithm based on Fuzzy (CHABCF)

Feature selection plays an important role in data mining and pattern recognition, especially in the case of large scale data. Feature selection is done due to large amount of noise and irrelevant features in the original data set. Hence, the efficiency of learning algorithms will increase incredibly if these irrelevant data are removed by this procedure. A novel approach for feature selection is introduced in this paper using CHABCF, (Chaotic Artificial Bee Colony based on Fuzzy), algorithm which is a combination of three paradigms: (1) Chaos theory (2) Artificial Bee Colony optimization and (3) Fuzzy logic. The fuzzy logic is used for ambiguity removal while chaos is used for generating better diversity in the initial population of our bee colony optimization algorithm. To demonstrate the efficiency of our algorithm, we have tested it on some well-known benchmarks such as wine, diabet and iris.