On the associations between variables in expert systems including default relations (Q908716)

A central function of the knowledge base of an expert system is to represent the associations between variables relevant to the domain of the system. We suggest that an expert's lack of complete knowledge of the underlying domain requires more sophisticated methods to represent these associations than used in the past. We discuss the appropriateness of the theory of approximate reasoning as a tool for handling knowledge which is nonspecific. We look at a methodology, based upon the theory of approximate reasoning, for representing and reasoning with default knowledge.