Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
397710 | International Journal of Approximate Reasoning | 2013 | 17 Pages |
There is no established formal framework for expert systems based on weighted IF–THEN rules. We discuss three mathematical models that have been recently proposed by the authors for CADIAG-2—a well-known system of this kind. The three frameworks are based on fuzzy logics, probability theory and possibilistic logic, respectively. CADIAG-2 is used here as a case study to evaluate these frameworks. We point out their use, advantages and disadvantages. In addition, the described models provide insight into various aspects of CADIAG-2.
► We discuss three different formal frameworks for the expert system CADIAG-2. ► The three frameworks are based on t-norm based logics, probability theory and possibilistic logic. ► We point out their advantages and disadvantages. ► We show how to use the frameworks based on t-norm based logics and probability theory for checking the system’s rules. ► Our frameworks also provide insights into various aspects of CADIAG-2.