Detecting Non-Spherical Clusters Using Modified CURE Algorithm

Clustering using representatives (CURE) algorithm is a robust hierarchical clustering algorithm which is dealing with noise and outliers. CURE algorithm merges and divides the clusters in some datasets which are not separate enough or have density difference between them. The obtained results show that CURE clustering is sensitive to input parameters. In this paper, the advantages of density-based cluster center detection are represented, and a modified CURE clustering algorithm is developed. The new algorithm determines the cluster centers and doesn't allow merging the clusters which contain cluster centers in data points. Experimental results show that the proposed algorithm has the capability to extract clusters more efficiently than the traditional CURE algorithm.