SPECIFICATION OF A FUZZY OBJECT INFERENCE: TOWARDS AN ADVANCED KNOWLEDGE-BASED PROGRAMMING

This paper provide a formal fuzzy object inference model to solve the following four significant drawbacks identified in extant fuzzy rule-based languages. First, they have difficulty in handling composite objects as a unit of inference. Second, they don't support fuzzy reasoning which is semantically easy to understand and conceptually simple to use. Third, their knowledge representation and reasoning style have a great semantic gap with those of current database models in syntax and semantics. Finally, they do not provide a comprehensive framework in treating uncertainties. In this paper, we demonstrate that the proposed model naturally models a target application environment in terms of composite objects possibly containing uncertain information, and successfully performs a fuzzy inference between them. To practically model the environment, we use the constructs of ICOT (Integrated C-Object Tool) extended for well implementing the structural semantics of the proposed model.