Cluster Center Initialization Using Hierarchical Two-Division of a Data Set along Each Dimension

This paper proposes a hierarchical two-division method that divides each mother subset of a data set at the same layer into two subsets along a dimension, and hierarchically divides the data set into a series of leaf subsets when the two-division process passes through each dimension of the data set. Then the initial cluster centers are picked out from the series of leaf subsets according to the rule that optimizes the dissimilarities among the initial cluster centers. Thus a new cluster center initialization method is developed. Experiments on real data sets show that the proposed cluster center initialization method is desirable.