A Hybrid K-harmonic Means with Abcclustering Algorithm Using an Optimal K Value for High Performance Clustering

Large quantities of data are emerging every year and an accurate clustering algorithm is needed to derive information from these data. K-means clustering algorithm is popular and simple, but has many limitations like its sensitivity to initialization, provides local optimum solutions. K-harmonic means clustering is an improved variant of K-means which is insensitive to the initialization of centroids, but still in some cases it ends up with local optimum solutions. Clustering using Artificial Bee Colony (ABC) algorithm always gives global optimum solutions. In this paper a new hybrid clustering algorithm (KHM-ABC) is presented by combining both K-harmonic means and ABC algorithm to perform accurate clustering. Experimental results indicate that the performance of the proposed algorithm is superior to the available algorithms in terms of the quality of clusters.