Authors: Fujiki Morii
A clustering criterion based on distortion ratios and its algorithm are proposed without the knowledge of the number of clusters. Computing distortion ratios on splitting and distortion ratios on merging for clusters of a data set, the criterion function is defined as the mean of the Euclidian distances between points of those distortion ratios and a reference point. To realize the criterion, an algorithm using split operations and merge operations for clusters is executed to decrease the criterion function. Through several classification experiments, the effectiveness of the criterion and its algorithm is demonstrated.
Keywords: clustering; criterion; distortion ratio; algorithm