A. Appendix Available Programs
2001
The purpose of this appendix is to describe briefly the operation of the nineteen programs mentioned in the endnotes throughout the monograph. A summary list is first given below that groups these nineteen programs into five general task areas: (1) object partitioning, (2) hierarchical clustering, (3) object sequencing/seriation, (4) object sequencing by constructing optimal paths, and (5) constructing ordered partitions. The various program acronyms are given, along with an indication of how the acronyms were constructed. Also, a brief statement is provided as to what each program is intended to do. These short summaries are elaborated upon in the five main sections of this appendix proper, which discuss the operation of each program under its general task area. The various options/alternatives offered are noted, and sample input/output provided. Summary List of Available Programs (1) Object partitioning— DPCL1U (Dynamic Programming CLustering 1-mode Unrestricted): the primary partitioning program for a symmetric proximity matrix. There are twelve options for defining subset heterogeneity, and two optimization criteria of minimizing either the sum or the maximum heterogeneity measure over the classes of a partition. DPCL1R (Dynamic Programming CLustering 1-mode Restricted): all the same options as DPCL1U for a symmetric proximity matrix, but with the additional requirement that an object order be given; all subsets considered contain objects consecutive with respect to this ordering. DPCL2U (Dynamic Programming CLustering 2-mode Unrestricted): the analogue of DPCL1U but now for two-mode data that are provided in the form of a rectangular proximity matrix between two disjoint sets.