A Fuzzy Extension of SWRL

Experience in using ontologies and rules in applications has shown that in many cases we would like to extend their representational and reasoning capabilities to deal with vague or imprecise knowledge. For example, multimedia applications have highlighted the need to extend representation languages with capabilities which allow for the treatment of the inherent imprecision in multimedia object representation, matching, detection and retrieval. Unfortunately, neither OWL nor SWRL provides such capabilities. In order to capture imprecision in rules, we propose fuzzy extensions of SWRL (SWRL is proposed by the Joint US/EU ad hoc Agent Markup Language Committee., the reader is referred to [HPSB04] for full details of the model-theoretic semantics and abstract syntax of SWRL.). In this extension OWL facts (axioms relating to individuals) can include a specification of the “degree” (a truth value between 0 and 1) of confidence with which we assert that an individual (resp. pair of individuals) is an instance of a given class (resp. property); and SWRL atoms can include a “weight” (a truth value between 0 and 1) that represents the “importance” of the atom in a rule.The syntax extends the existing abstract syntax for SWRL, while the semantics is based on the theory of fuzzy sets [Zad65]. For example, the following fuzzy rule asserts that if one has a parent that is happy, then one is likely (with the weight 0.8) to be happy: