On some logical points of fuzzy conditional decision making (Q1095883)

The paper deals with some properties of PL (Probability Logic) and FL (Fuzzy Logic). FL is derived from PL by means of choosing the connectives such that the logical system created is truth-functional. From many alternatives the usual connectives of FL emerged. As is known, PL is not truth-functional. Conditionality is considered by means of implication and using a fuzzy counterpart of conditional probability. In that way some basic tools are created for Bayesian-like decision making in a fuzzy environment. The analysis shows that the implication and the conditional probability-like counterpart are not necessarily the same thing. A labelling system is introduced in fuzzy-valued FL such that a finite set of non-negative integers forms the set of truth-values, which actually are fuzzy sets of the unit interval [0,1]. In this way new quantitative expressions are obtained for each qualitative linguistic variable. In doing so, the system preserves fuzziness. Some general rules of inference are also considered.