Author : Wilson Joseph, W. Jeberson and Klinsega Jeberson

The k-means clustering algorithm is a premier algorithm for data clustering. However, one of its limitations is the need to specify the number of clusters, K, before the algorithm is implemented. In this paper, we present a novel technique which assists us in determining K while performing data clustering on one dimensional data using enhanced k-means algorithm. The technique is based upon a Count test, which is performed on the dataset. The elements which pass the test become the initial cluster heads for k-means clustering. The experimental results suggest that the proposed technique is efficient, produces better results than rule of thumb technique used for determining K. The technique also helps in addressing the problem of empty clusters.

Keywords: Clustering algorithm, K-means algorithm, Initial cluster heads, Count Test, Rule of thumb, Cluster Analysis.

Article published in International Journal of Current Engineering and Technology, Vol.5, No.2 (April-2015)