Use of Model-Segmentation Criteria in Clustering and Segmentation of Time Series and Digital Images.
Stanley L. Sclove · 1983
This paper treats the development and use of criteria for model selection, particularly for the choice of the number of groups ( clusters ) in the analysis of multivariate data and of the number of classes of segments in the segmentation of time series and digital images. Criteria such as those of Akaike, Schwarz and Kashyap are considered. (Author)