A4 Conference proceedings
Linguistic approximation under different distances/similarity measures for fuzzy numbers
Open Access publication
Publication Details Authors: Talášek Tomáš, Stoklasa Jan
Publication year: 2016
Language: English
Related Journal or Series Information: LUT Scientific and Expertise Publications : Tutkimusraportit - Research reports Title of parent publication: Proceedings of the NSAIS’16 Workshop on Adaptive and Intelligent Systems 2016
Start page: 49
End page: 52
ISBN: 978-952-265-985-9
eISBN: 978-952-265-986-6
ISSN: 2243-3376
JUFO-Level of this publication: 0
Open Access: Open Access publication
|
Abstract
The paper explores the performance of two selecteddistance measures of fuzzy numbers (the Bhattacharyya distanceand the dissemblance index) and two different fuzzy similaritymeasures in the context of linguistic approximation. Symmetricaltriangular fuzzy numbers are considered as the approximatedentities. A numerical experiment is employed to map the behaviorof the four distance/similarity measures under uniform linguisticscales used for the linguistic approximation. We identify aclear difference of the performance of the selected methodswhen the approximation of low-uncertain fuzzy numbers areapproximated. One of the selected fuzzy similarity measurestends to emphasize extreme evaluations and disregard intermediatelinguistic values. For highly uncertain fuzzy numbers themethods seem to be close to equivalent in terms of the linguisticapproximation.