Fuzzy expert system technology

Abstract

Fuzzy expert systems are the second generation expert systems. In real world activities, system states have various attributes that at best could be assessed as a matter of degree. If we express our knowledge of system states in terms of an all or nothing paradigm then much valuable information does not appear in system's models. However, if we express our knowledge of systems with fuzzy sets, much valuable information stays embedded in the rules of behaviour in system's models. Thus, the fuzzy set paradigm gives our knowledge of systems an effective representation.

For this purpose, four levels of knowledge representation are identified as: (i) linguistic, (ii) meta-linguistic, (iii) propositional, and (iv) computational. In particular, four classes of fuzzy rule base schema are identified within a unified framework of implication statements. That is we identify at least four classes of propositional expressions that represent linguistic expression of rules. Interval valued fuzzy sets are defined as particular Type II fuzzy sets in order to expose second order semantic uncertainty embedded in linguistic expressions.