A modified Bee Colony Optimization (MBCO) and its hybridization with k-means for an application to data clustering

摘要

Among the nature inspired heuristic or meta-heuristic optimization algorithms, Bee Colony Optimization (BCO) algorithms are widely used to solve clustering problem. In this paper, a modified BCO (MBCO) approach is proposed for data clustering. In the proposed MBCO, the forgiveness characteristics of bees and giving a fair chance to both trustworthy and untrustworthy bees are being taken care of. A probability based selection (Pbselection) approach is introduced in the proposed MBCO for assigning unallocated data points in every iteration. The result shows that, the proposed method gives faster convergence as compared to the existing well known algorithms. In order to improve the performance of MBCO further and to obtain global optimal and diverse solution, the proposed MBCO is hybridized with k-means algorithm. In average, the hybridized MKCLUST and KMCLUST give same or better result than the proposed MBCO. To validate the proposed algorithms, seven standard data sets are considered. From classification error percentages calculation, it is observed that the proposed algorithms perform better compared to some existing algorithms. The simulation results infer that the proposed algorithms can be efficiently used for data clustering. (C) 2018 Elsevier B.V. All rights reserved.