Knowledge acquisition in fuzzy reasoning

Summary form only given. Fuzzy reasoning is the fuzzy extension of inference operations used in AI or expert systems. Among the various fuzzy reasoning methods, rule-based fuzzy reasoning is the most important from a viewpoint of industrial applications. In Japan, industrial fuzzy expert systems have been developed since the early 1980s, especially in the area of control. In such applications the control knowledge of human experts is represented by a set of so-called fuzzy production rules. In such a fuzzy knowledge base, the size is generally rather small, since the fuzzy (or ambiguous) descriptions are allowed, so the knowledge-acquisition process is not so difficult compared with ordinary (or crisp) expert systems, where the number of rules is more than a few thousand. A question-and-answer method is used in the first step, then the parameter tuning is done, and finally a complete set of fuzzy rules is obtained.