A clustering method for symbolic interval-type data using adaptive chebyshev distances

Abstract

This work presents a partitioning method for clustering symbolic interval-type data using a dynamic cluster algorithm with adaptive Chebyshev distances. This method furnishes a partition and a proto-type for each cluster by optimizing an adequacy criterion that measures the fitting between the clusters and their representatives. To compare interval-type data, the method uses an adaptive Chebyshev distance that changes for each cluster according to its intra-class structure at each iteration of the algorithm. Experiments with real and artificial interval-type data sets demonstrate the usefulness of the proposed method. © Springer-Verlag 2004.