Separated antecedent and consequent learning for Takagi-Sugeno fuzzy systems

János Botzheim, Edwin Lughofer, Erich Peter Klement, L. Kóczy, Tamás D. Gedeon

Research output: Chapter in Book/Report/Conference proceedingConference contribution

10 Citations (Scopus)

Abstract

In this paper a new algorithm for the learning of Takagi-Sugeno fuzzy systems is introduced. In the algorithm different learning techniques are applied for the antecedent and the consequent parameters of the fuzzy system. We propose a hybrid method for the antecedent parameters learning based on the combination of the Bacterial Evolutionary Algorithm (BEA) and the Levenberg-Marquardt (LM) method. For the linear parameters in fuzzy systems appearing in the rule consequents the Least Squares (LS) and the Recursive Least Squares (RLS) techniques are applied, which will lead to a global optimal solution of linear parameter vectors in the least squares sense. Therefore a better performance can be guaranteed than with a complete learning by BEA and LM. The paper is concluded by evaluation results based on high-dimensional test data. These evaluation results compare the new method with some conventional fuzzy training methods with respect to approximation accuracy and model complexity.

Original languageEnglish
Title of host publicationIEEE International Conference on Fuzzy Systems
Pages2263-2269
Number of pages7
DOIs
Publication statusPublished - 2006
Event2006 IEEE International Conference on Fuzzy Systems - Vancouver, BC, Canada
Duration: Jul 16 2006Jul 21 2006

Other

Other2006 IEEE International Conference on Fuzzy Systems
CountryCanada
CityVancouver, BC
Period7/16/067/21/06

Fingerprint

Fuzzy systems
Evolutionary algorithms
Learning algorithms

Keywords

  • (recursive) least squares
  • Bacterial evolutionary algorithm
  • Levenberg-Marquardt method
  • Takagi-Sugeno fuzzy systems

ASJC Scopus subject areas

  • Software
  • Safety, Risk, Reliability and Quality
  • Chemical Health and Safety

Cite this

Botzheim, J., Lughofer, E., Klement, E. P., Kóczy, L., & Gedeon, T. D. (2006). Separated antecedent and consequent learning for Takagi-Sugeno fuzzy systems. In IEEE International Conference on Fuzzy Systems (pp. 2263-2269). [1682014] https://doi.org/10.1109/FUZZY.2006.1682014

Separated antecedent and consequent learning for Takagi-Sugeno fuzzy systems. / Botzheim, János; Lughofer, Edwin; Klement, Erich Peter; Kóczy, L.; Gedeon, Tamás D.

IEEE International Conference on Fuzzy Systems. 2006. p. 2263-2269 1682014.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Botzheim, J, Lughofer, E, Klement, EP, Kóczy, L & Gedeon, TD 2006, Separated antecedent and consequent learning for Takagi-Sugeno fuzzy systems. in IEEE International Conference on Fuzzy Systems., 1682014, pp. 2263-2269, 2006 IEEE International Conference on Fuzzy Systems, Vancouver, BC, Canada, 7/16/06. https://doi.org/10.1109/FUZZY.2006.1682014
Botzheim J, Lughofer E, Klement EP, Kóczy L, Gedeon TD. Separated antecedent and consequent learning for Takagi-Sugeno fuzzy systems. In IEEE International Conference on Fuzzy Systems. 2006. p. 2263-2269. 1682014 https://doi.org/10.1109/FUZZY.2006.1682014
Botzheim, János ; Lughofer, Edwin ; Klement, Erich Peter ; Kóczy, L. ; Gedeon, Tamás D. / Separated antecedent and consequent learning for Takagi-Sugeno fuzzy systems. IEEE International Conference on Fuzzy Systems. 2006. pp. 2263-2269
@inproceedings{3fa454b0bf774216b0a89690e441ca5c,
title = "Separated antecedent and consequent learning for Takagi-Sugeno fuzzy systems",
abstract = "In this paper a new algorithm for the learning of Takagi-Sugeno fuzzy systems is introduced. In the algorithm different learning techniques are applied for the antecedent and the consequent parameters of the fuzzy system. We propose a hybrid method for the antecedent parameters learning based on the combination of the Bacterial Evolutionary Algorithm (BEA) and the Levenberg-Marquardt (LM) method. For the linear parameters in fuzzy systems appearing in the rule consequents the Least Squares (LS) and the Recursive Least Squares (RLS) techniques are applied, which will lead to a global optimal solution of linear parameter vectors in the least squares sense. Therefore a better performance can be guaranteed than with a complete learning by BEA and LM. The paper is concluded by evaluation results based on high-dimensional test data. These evaluation results compare the new method with some conventional fuzzy training methods with respect to approximation accuracy and model complexity.",
keywords = "(recursive) least squares, Bacterial evolutionary algorithm, Levenberg-Marquardt method, Takagi-Sugeno fuzzy systems",
author = "J{\'a}nos Botzheim and Edwin Lughofer and Klement, {Erich Peter} and L. K{\'o}czy and Gedeon, {Tam{\'a}s D.}",
year = "2006",
doi = "10.1109/FUZZY.2006.1682014",
language = "English",
isbn = "0780394887",
pages = "2263--2269",
booktitle = "IEEE International Conference on Fuzzy Systems",

}

TY - GEN

T1 - Separated antecedent and consequent learning for Takagi-Sugeno fuzzy systems

AU - Botzheim, János

AU - Lughofer, Edwin

AU - Klement, Erich Peter

AU - Kóczy, L.

AU - Gedeon, Tamás D.

PY - 2006

Y1 - 2006

N2 - In this paper a new algorithm for the learning of Takagi-Sugeno fuzzy systems is introduced. In the algorithm different learning techniques are applied for the antecedent and the consequent parameters of the fuzzy system. We propose a hybrid method for the antecedent parameters learning based on the combination of the Bacterial Evolutionary Algorithm (BEA) and the Levenberg-Marquardt (LM) method. For the linear parameters in fuzzy systems appearing in the rule consequents the Least Squares (LS) and the Recursive Least Squares (RLS) techniques are applied, which will lead to a global optimal solution of linear parameter vectors in the least squares sense. Therefore a better performance can be guaranteed than with a complete learning by BEA and LM. The paper is concluded by evaluation results based on high-dimensional test data. These evaluation results compare the new method with some conventional fuzzy training methods with respect to approximation accuracy and model complexity.

AB - In this paper a new algorithm for the learning of Takagi-Sugeno fuzzy systems is introduced. In the algorithm different learning techniques are applied for the antecedent and the consequent parameters of the fuzzy system. We propose a hybrid method for the antecedent parameters learning based on the combination of the Bacterial Evolutionary Algorithm (BEA) and the Levenberg-Marquardt (LM) method. For the linear parameters in fuzzy systems appearing in the rule consequents the Least Squares (LS) and the Recursive Least Squares (RLS) techniques are applied, which will lead to a global optimal solution of linear parameter vectors in the least squares sense. Therefore a better performance can be guaranteed than with a complete learning by BEA and LM. The paper is concluded by evaluation results based on high-dimensional test data. These evaluation results compare the new method with some conventional fuzzy training methods with respect to approximation accuracy and model complexity.

KW - (recursive) least squares

KW - Bacterial evolutionary algorithm

KW - Levenberg-Marquardt method

KW - Takagi-Sugeno fuzzy systems

UR - http://www.scopus.com/inward/record.url?scp=34250725862&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=34250725862&partnerID=8YFLogxK

U2 - 10.1109/FUZZY.2006.1682014

DO - 10.1109/FUZZY.2006.1682014

M3 - Conference contribution

SN - 0780394887

SN - 9780780394889

SP - 2263

EP - 2269

BT - IEEE International Conference on Fuzzy Systems

ER -