Using a distance between fuzzy number-valued fuzzy measures to determine attribute importance: An outcome-oriented perspective

The purpose of this paper is to introduce a new method for measuring attribute salience in decision analysis. Fuzzy measures have been widely used to determine the degrees of subjective importance of evaluation items. We consider the value of fuzzy measures as a linguistic value and then convert linguistic terms to fuzzy numbers. The grades of attribute importance are correspondingly expressed by fuzzy number-valued fuzzy measures. However, the leniency error may exist when most attributes are assigned unduly high ratings. To reduce positive leniency in fuzzy measure ratings, we develop an importance-assessing method by comparison of fuzzy number-valued fuzzy measures using a fuzzy distance measure. An outcome-oriented approach is adopted to validate the proposed method. Several multiattribute evaluation cases in consumer decision-making reality are addressed to examine the feasibility and applicability of the current method. © 2006 TFSA.