The Centroid Initialization for K-Means Clustering Algorithm Based on T-Score Ranking Method

In this paper, we propose an algorithm to compute initial cluster centers for K-means clustering based on T-Score ranking. This scoring technique is a statistical method of ranking numerical and nominal attributes based on distance measure. The data are sorted based on the score values. Then divide the ranked data into k subsets. Calculate the mean values of each k subsets. Pick the nearby value of data to the mean as the initial centroid. The experimental results suggest that the proposed algorithm is effective, converge to better clustering results than those of the random initialization method. The research also indicated the proposed algorithm would greatly improve the likelihood of every cluster containing some data in it.