A MULTISEED NON-HIERARCHICAL CLUSTERING TECHNIQUE FOR DATA-ANALYSIS

Clustering methods such as K-means and its variations, such as Forgy, as well as their improved version ISODATA, do not work well if the sha pe of the cluster is elongated. It is pointed out that a single seed p oint cannot correctly reflect the nature of the data of an elongated c luster. A multiseed clustering algorithm is proposed, where one cluste r may contain more than one seed point. A density-based algorithm is u sed to choose the initial seed points. To assign several seed points t o one cluster, a minimal spanning tree guided novel merging technique is proposed. The merging technique is quite general and may be applied to other clustering approaches as well. Experimental results are pres ented to demonstrate the efficiency of this clustering procedure.