A semantics-driven, fuzzy logic-based approach to knowledge representation and inference

Abstract

The use of Fuzzy Logic allows for a natural expression of concepts employed by experts and users. Although the use of Fuzzy Logic partially reduces the software systems maintenance problems associated to vagueness modelling, it introduces other problems in tuning the numbers used by the system to model such vagueness. On the other hand, Ripple Down Rules (RDR) provide for easy maintenance when dealing with Knowledge-based systems but it does not allow to model terminological vagueness. The work presented in this paper shows the fundamentals of a theoretical approach that combines both methodologies RDR and Fuzzy Logic into a semantics-driven framework. We believe that the advantages of using/implementing this framework are clear, although a number research issues remain open.