DocumentCode
3624666
Title
On Possible Constraints in Applications of Basic Defuzzification Techniques
Author
Dragan Z. Saletic;Uros Popovic
Author_Institution
Member, IEEE, Military-Technical Institute, Belgrade, Serbia, School of Computing, University "Union", Belgrade. phone +381-11-3539713
fYear
2006
Firstpage
225
Lastpage
230
Abstract
The purpose of this paper is to give the results of research analysis of basic defuzzification techniques and to point out to possible constraints in application of these techniques. The defuzzification process is present in a fuzzy system when an output fuzzy set should be mapped to a crisped value. Features are given which is the base for a defuzzification technique comparison. Techniques are classified into several groups. The analysis of defuzzification techniques is given, as well as their comparison on the base of given features. The results from computer experiments suggest constrained domain of applicability of defuzzification techniques. Suggestions about applicability of techniques in specified applications are given
Keywords
"Fuzzy systems","Fuzzy sets","Hybrid intelligent systems","Neural networks","Expert systems","Fuzzy logic","Artificial intelligence","Uncertainty","Military computing","Automatic control"
Publisher
ieee
Conference_Titel
Neural Network Applications in Electrical Engineering, 2006. NEUREL 2006. 8th Seminar on
Print_ISBN
1-4244-0432-0
Type
conf
DOI
10.1109/NEUREL.2006.341218
Filename
4147206
Link To Document