Abstract
In this paper, a framework for representing vagueness and uncertainty based on the notion of context model introduced by Gebhardt and Kruse (1993) is reviewed. From a concept analysis point of view, the context model can be semantically considered as a data
model for fuzzy concept analysis. Under such a consideration, the context model provides a practical method for constructing membership functions of fuzzy concepts in connection with a likelihood view on the interpretation of membership grades. On the other hand, from a decision analysis point of view, the Dempster-Shafer theory of evidence will be re-interpreted within the framework of context model, and in order to deal with the problem of synthesis of vague evidence linguistically provided by the experts in some situations of decision analysis, the notions of context-dependent vague characteristics and fuzzy context model will be
introduced. It is shown that each context-dependent vague characteristic within fuzzy context model directly induces a uncertainty measure of type 2 interpreted as "vague" belief function, which is inferred from vague evidence expressed linguistically.
