A consistent procedure for determining the number of clusters in regression clustering

In this paper, an information-based criterion for determining the number of clusters in the problem of regression clustering is proposed. It is shown that, under a probabilistically structured population, the proposed criterion selects the true number of regression hyperplanes with probability one among all class-growing sequences of classifications, when the number of observations n from the population increases to infinity. Results from a simulation study are also presented. (c) 2004 Elsevier B.V. All rights reserved.