A PROC MATRIX program for preference-dissimilarity multidimensional scaling

A computer program can be a means of communicating the structure of an algorithm as well as a tool for data analysis. From this perspective high-level matrix-oriented languages like PROC MATRIX in the SAS system are especially useful because of their readability and compactness. An algorithm for the joint analysis of dissimilarity and preference data using maximum likelihood estimation is presented in PROC MATRIX code.This research was supported by Grant APA 3020 from the Natural Sciences and Engineering Research Council of Canada.