A type-2 fuzzy c-regression clustering algorithm for Takagi–Sugeno system identification and its application in the steel industry

This paper proposes a new type-2 fuzzy c-regression clustering algorithm for the structure identification phase of Takagi–Sugeno (T–S) systems. We present uncertainties with fuzzifier parameter “ m”. In order to identify the parameters of interval type-2 fuzzy sets, two fuzzifiers “ m 1 ” and “ m 2...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Information sciences Jg. 187; S. 179 - 203
Hauptverfasser: Fazel Zarandi, M.H., Gamasaee, R., Turksen, I.B.
Format: Journal Article
Sprache:Englisch
Veröffentlicht: Elsevier Inc 15.03.2012
Schlagworte:
Algorithms
Clustering
Fuzzy
Fuzzy logic
Fuzzy set theory
Gaussian mixture
Intervals
Iron and steel industry
IT2F c-regression clustering
Mathematical models
Multiple-regression
Steel industry
Structure identification
Weighted least square
Structure identification
Gaussian mixture
Multiple-regression
IT2F c-regression clustering
Steel industry
Weighted least square
ISSN:0020-0255, 1872-6291
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:This paper proposes a new type-2 fuzzy c-regression clustering algorithm for the structure identification phase of Takagi–Sugeno (T–S) systems. We present uncertainties with fuzzifier parameter “ m”. In order to identify the parameters of interval type-2 fuzzy sets, two fuzzifiers “ m 1 ” and “ m 2 ” are used. Then, by utilizing these two fuzzifiers in a fuzzy c-regression clustering algorithm, the interval type-2 fuzzy membership functions are generated. The proposed model in this paper is an extended version of a type-1 FCRM algorithm [25], which is extended to an interval type-2 fuzzy model. The Gaussian Mixture model is used to create the partition matrix of the fuzzy c-regression clustering algorithm. Finally, in order to validate the proposed model, several numerical examples are presented. The model is tested on a real data set from a steel company in Canada. Our computational results show that our model is more effective for robustness and error reduction than type-1 NFCRM and the multiple-regression.
Bibliographie:ObjectType-Article-2
SourceType-Scholarly Journals-1
ObjectType-Feature-1
content type line 23
ISSN:0020-0255
1872-6291
DOI:10.1016/j.ins.2011.10.015