Fuzzy CBR based on Pattern Recognition and its Application

Abstract

Case-based reasoning (CBR) has been widely applied in expert systems. This method can find out the solution of the problem to be solved in terms of the former experience by analyzing the similarity of information between the problem to be solved and the existing cases. However, among the information from the problem, some data are fuzzy. In order to find out the similar cases successfully, a method of fuzzy CBR based on pattern recognition is presented in this paper. According to this method, all the relative existing cases in the case database are fuzzed and each can be regarded as typical pattern. And then a new general closeness degree algorithm was constructed to deal with general and fuzzy data by combining two kinds of closeness degree algorithm and is used as the similarity degree function between the cases and the problem to be solved so as to realize case matching under the fuzzy strategy by using the method of pattern recognition. The validity of this method has been confirmed in electrical apparatus product design