No label defined (Q4945172)

Fuzzy set theory appears to be an adequate formal theory for handling the difficult task of computing with words, i.e. with verbal quantitative data. The main purpose of this paper is to extend, discuss and interpret a fuzzy set-theoretical model of verbal quantities, formulated by the authors in some previous papers. Trying to obtain a model for verbal quantities, one can show that each verbal expression corresponds to a relatively wide set of fuzzy numbers. If the fuzzy numbers (and operations with them) represent verbal expressions then the construction of their membership functions has to be derived from the quantitative and qualitative (semantic) content of the related words, including the quality of their sources. Within this paper, the formal representation of the qualitative verbal expression is realized by means of a triple consisting of: (a) the crisp core (determining the numerical value referred to by the words); (b) the scale (characterizing the qualities of the source of verbal data), and (c) the shape generator (describing a normalized form of the membership functions and representing the qualitative component of the expression). The proposed model of fuzzy quantities based on generator triples offers a wider and better approach to the fuzziness than the classical concepts of fuzzy quantities (e.g., triangular or trapezoidal fuzzy numbers), and an improved processing efficiency in terms of the extension principle (or elementary manipulation with triangular and trapezoidal fuzzy numbers). For the entire collection see [Zbl 0931.00022].