A hybrid bio-inspired algorithm and its application

Clustering is one of the attractive and major tasks in data mining that is used in many applications. It refers to group together data points, which are similar to one another based on some criteria. One of the efficient algorithms which applied on data clustering is particle swarm optimization (PSO) algorithm. However, PSO often leads to premature convergence and its performance is highly depended on parameter tuning and many efforts have been done to improve its performance in different ways. In order to improve the efficiency of the PSO on data clustering, it is hybridized with the big bang-big crunch algorithm (BB-BC) in this paper. In the proposed algorithm, namely PSO-BB-BC, PSO is used to explore the search space for finding the optimal centroids of the clusters. Whenever PSO loses its exploration, to prevent premature convergence, BB-BC algorithm is used to diversify the particles. The performance of the hybrid algorithm is compared with PSO, BB-BC and K-means algorithms using six benchmark datasets taken from the UCI machine learning repository. Experimental results show that the hybrid algorithm is superior to other test algorithms in all test datasets in terms of the quality of the clusters.