Generally, human knowledge can be vague and/or ambiguous. When the knowledge is represented by facts and rules, the vagueness can be dealt with fuzzy value representations, and the ambiguity can be dealt with certainty factors(CF). In this method, the...
Generally, human knowledge can be vague and/or ambiguous. When the knowledge is represented by facts and rules, the vagueness can be dealt with fuzzy value representations, and the ambiguity can be dealt with certainty factors(CF). In this method, the antecident part of the rule may not be exactly matched with the given fact. So their matching degree should be considered in the calculation of the certainty factor.
In this thesis, the reasons for considering fuzzy matching in the vague and ambiguous knowledge-contained expert systems is discussed, and a new method of processing CF with a fuzzy matching degree in inference networks is proposed.