A FUZZY SCHEME WITH APPLICATION TO EXPERT CLASSIFICATION SYSTEMS FOR COMPUTING THE DEGREE OF MATCH BETWEEN A RULE AND AN ASSERTION
Seongwon Cho, Mark R. Lehto · Cybernetics & Systems · 1992
We propose a fuzzy scheme for the computation of the degree of match between the antecedent part of a production rule in a rule base and an assertion stored in working memory when imprecise and precise propositions are mixed freely. In order to estimate to what extent the antecedent part X is A is matched against an assertion X is A*. a value between -1 and + 1 is computed from the truth relationship between A and A*. The proposed scheme is computationally efficient and can be applied consistently for all combinations of precise and imprecise propositions. Furthermore, the scheme allows estimates of imprecision to be combined with certainty factors. The scheme consequently seems especially useful for application in expert classification systems.