Intercategory Relations and Taxonomic Analysis

Abstract

Some basic concepts for intercategory relations have already been introduced (e.g., inclusion), but a framework for them has not yet been presented. Section 5.1 does so, beginning with the concepts of fuzzy equivalence and similarity relations and drawing comparisons with the traditional behavioral science measures of correlation and association. Section 5.1.2 introduces a fuzzier relation than similarity, namely overlap. The remainder of this chapter develops several techniques for data reduction and taxonomic analysis based on fuzzy inclusion, similarity, and overlap. Two of these are entirely based on fuzzy set theory (fuzzy clustering, and fuzzy overlap analysis), while others augment or extend conventional techniques (e.g., nonfuzzy clustering and discriminant analysis).