Information semantics of possibilistic uncertainty

Abstract

All major theories of uncertainty permit defining numerical uncertainty measures. Formulae can originate by axiomatic approach or by modifying either Shannon or related entropy model. The latter has a very natural semantics as the complexity of transmission of specific information about the uncertainty assignment. Such interpretation has been lacking in the possibility model, based on fuzzy set theory. This paper presents such information semantics for this theory. It is accomplished by demonstrating that possibilistic uncertainty can be viewed as the expected value of information identifying the acceptability of choices from the fuzzy domain. © 1995.