(28) Stanley L. Sclove. Statistical models and methods for cluster analysis and segmentation. Proceedings of the 31st Conference on the Design of Experiments. U.S. Army Research Office, Research Triangle Park, NC, 1986, pp. 29-38.

Clustering of individuals, segmentation of time series and segmentation of numerical images can all be considered as labeling problems, for each can be described in terms of pairs (x(t),g(t)), t = 1,2,...,n, where x(t) is the observation at instance t and g(t) is the unobservable "label" of instance t. The labels are to be estimated, along with any unspecified distributional parameters. In cluster analysis the values of t are the individuals (cases) observed and the x's are independent. In time series the values of t are time instants and there is temporal correlation. In numerical image segmentation the values of t denote picture elements (pixels) and spatial correlation between neighboring pixels can be utilized.

Signals and time-series often are not homogeneous but rather are generated by mechanisms or processes with various phases. Similarly, images are not homogeneous but contain various objects. "Segmentation" is a process of attempting to recover automatically the phases or objects. A model for representing such signals, time series, and images was discussed in a paper by the present author in the Proceedings of the 30th Conference; some approaches to estimation and segmentation in this model were presented. The present paper summarizes the work on all these types of labeling problems, clustering as well as time series- and image-segmentation.

Key words and phrases: statistical pattern recognition, classification; temporal correlation, spatial correlation; optimization by relaxation method