Fuzzy association rules: Semantic issues and quality measures

Abstract

Fuzzy association rules provide a data mining tool which is especially interesting from a knowledge-representational point of view since fuzzy attribute values allow for expressing rules in terms of natural language. So far, however, association rules of this type have not been investigated thoroughly from a semantical point of view. Particularly, the quality measures which have been proposed for assessing such rules are mostly "ad-hoc" generalizations of measures for classical rules. In this paper, we show that fuzzy association rules can be interpreted in different ways and that the interpretation has a strong influence on their assessment and, hence, on the process of rule mining. We motivate the use of multiple-valued implication operators in order to model fuzzy association rules and propose quality measures suitable for this type of rule. © Springer-Verlag 2001.