Applications of type-2 fuzzy logic systems: handling the uncertainty associated with surveys
Authors:
- J.M. MendelJMJ.M. Mendel
- 1 January 1999
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 3 (10987584) , 1546-1551 vol.3
Abstract
We discuss designs of type-2 fuzzy logic systems (FLSs), where the FLSs are based on rules collected from surveys of multiple experts. Two kinds of uncertainties occur in this problem: 1) the uncertainty due to different experts giving different answers to the same question; and 2) the linguistic uncertainty associated with the labels of the fuzzy sets. The latter causes different experts to assign different values to membership function parameters. We show how both kinds of uncertainties can be approached within the framework of type-2 FLSs. Uncertainty about answers to questions is treated using two different approaches: averaging the responses, or preserving all the responses. The former is computationally more efficient; however, the latter provides more information and can be used when such information is deemed useful. Uncertainty about labels of fuzzy sets is handled directly by using type-2 fuzzy sets. We demonstrate that both kinds of uncertainties can affect the crisp output of a type-2 FLS.Keywords
This publication has 2 references indexed in Scilit:
- Structure identification of fuzzy modelFuzzy Sets and Systems, 1988
- Fuzzy identification of systems and its applications to modeling and controlIEEE Transactions on Systems, Man, and Cybernetics, 1985