Data Clustering Using the Cooperative Search Based Artificial Bee Colony Algorithm

Data clustering is a significant and strong data analysis technique which has a broad range of applications in many domains. In this paper, the artificial bee colony algorithm (ABC) is adopted to partition data sets into K clusters. To trade off the global and local searching ability of ABC algorithm, two kinds of cooperative search based ABC algorithms are proposed, that is N2ABC and WCABC. Then, the proposed algorithms are combined with K-means to deal with data clustering. For the purpose of demonstrating the efficiency of two hybrid clustering algorithms (N2ABCC and WCABCC), one artificial data set and six benchmark data sets are selected to test clustering results. Meanwhile, five algorithms, namely K-means, PSOC, ABCC, GABCC and CABCC, are chosen for comparison. The clustering results indicate that the proposed algorithms have better clustering validity than other algorithms.