Identification of quasi-ARX neurofuzzy model by using SVR-based approach with input selection

Yu Cheng, Lan Wang, Jing Zeng, Takayuki Furuzuki

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

2 Citations (Scopus)

Abstract

Quasi-ARX neurofuzzy (Q-ARX-NF) models have shown great approximation ability and usefulness in nonlinear system identification and control. However, the incorporated neurofuzzy networks suffer from the curse-of-dimensionality problem, which may result in high computational complexity and over-fitting. In this paper, support vector regressor (SVR) based identification approach is used to reduce computational complexity with the help of transforming the original problem into Lagrange space, which is only sensitive to the number of data samples. Furthermore, to improve the generalization capability, a parsimonious model structure is obtained by eliminating insignificant input variables for the incorporated neurofuzzy network, which is implemented by genetic algorithm (GA) based input selection method with a novel fitness evaluation function. Two numerical simulations are tested to show the effectiveness of the proposed method.

Original languageEnglish
Title of host publicationConference Proceedings - IEEE International Conference on Systems, Man and Cybernetics
Pages1585-1590
Number of pages6
DOIs
Publication statusPublished - 2011
Event2011 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2011 - Anchorage, AK
Duration: 2011 Oct 92011 Oct 12

Other

Other2011 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2011
CityAnchorage, AK
Period11/10/911/10/12

Fingerprint

Computational complexity
Identification (control systems)
Function evaluation
Model structures
Nonlinear systems
Genetic algorithms
Computer simulation

Keywords

  • identification
  • input selection
  • Quasi-ARX neurofuzzy network
  • SVR

ASJC Scopus subject areas

  • Electrical and Electronic Engineering
  • Control and Systems Engineering
  • Human-Computer Interaction

Cite this

Cheng, Y., Wang, L., Zeng, J., & Furuzuki, T. (2011). Identification of quasi-ARX neurofuzzy model by using SVR-based approach with input selection. In Conference Proceedings - IEEE International Conference on Systems, Man and Cybernetics (pp. 1585-1590). [6083897] https://doi.org/10.1109/ICSMC.2011.6083897

Identification of quasi-ARX neurofuzzy model by using SVR-based approach with input selection. / Cheng, Yu; Wang, Lan; Zeng, Jing; Furuzuki, Takayuki.

Conference Proceedings - IEEE International Conference on Systems, Man and Cybernetics. 2011. p. 1585-1590 6083897.

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

Cheng, Y, Wang, L, Zeng, J & Furuzuki, T 2011, Identification of quasi-ARX neurofuzzy model by using SVR-based approach with input selection. in Conference Proceedings - IEEE International Conference on Systems, Man and Cybernetics., 6083897, pp. 1585-1590, 2011 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2011, Anchorage, AK, 11/10/9. https://doi.org/10.1109/ICSMC.2011.6083897
Cheng Y, Wang L, Zeng J, Furuzuki T. Identification of quasi-ARX neurofuzzy model by using SVR-based approach with input selection. In Conference Proceedings - IEEE International Conference on Systems, Man and Cybernetics. 2011. p. 1585-1590. 6083897 https://doi.org/10.1109/ICSMC.2011.6083897
Cheng, Yu ; Wang, Lan ; Zeng, Jing ; Furuzuki, Takayuki. / Identification of quasi-ARX neurofuzzy model by using SVR-based approach with input selection. Conference Proceedings - IEEE International Conference on Systems, Man and Cybernetics. 2011. pp. 1585-1590
@inproceedings{00228add7b0a44ce97ea3f7e1043d1f3,
title = "Identification of quasi-ARX neurofuzzy model by using SVR-based approach with input selection",
abstract = "Quasi-ARX neurofuzzy (Q-ARX-NF) models have shown great approximation ability and usefulness in nonlinear system identification and control. However, the incorporated neurofuzzy networks suffer from the curse-of-dimensionality problem, which may result in high computational complexity and over-fitting. In this paper, support vector regressor (SVR) based identification approach is used to reduce computational complexity with the help of transforming the original problem into Lagrange space, which is only sensitive to the number of data samples. Furthermore, to improve the generalization capability, a parsimonious model structure is obtained by eliminating insignificant input variables for the incorporated neurofuzzy network, which is implemented by genetic algorithm (GA) based input selection method with a novel fitness evaluation function. Two numerical simulations are tested to show the effectiveness of the proposed method.",
keywords = "identification, input selection, Quasi-ARX neurofuzzy network, SVR",
author = "Yu Cheng and Lan Wang and Jing Zeng and Takayuki Furuzuki",
year = "2011",
doi = "10.1109/ICSMC.2011.6083897",
language = "English",
isbn = "9781457706523",
pages = "1585--1590",
booktitle = "Conference Proceedings - IEEE International Conference on Systems, Man and Cybernetics",

}

TY - GEN

T1 - Identification of quasi-ARX neurofuzzy model by using SVR-based approach with input selection

AU - Cheng, Yu

AU - Wang, Lan

AU - Zeng, Jing

AU - Furuzuki, Takayuki

PY - 2011

Y1 - 2011

N2 - Quasi-ARX neurofuzzy (Q-ARX-NF) models have shown great approximation ability and usefulness in nonlinear system identification and control. However, the incorporated neurofuzzy networks suffer from the curse-of-dimensionality problem, which may result in high computational complexity and over-fitting. In this paper, support vector regressor (SVR) based identification approach is used to reduce computational complexity with the help of transforming the original problem into Lagrange space, which is only sensitive to the number of data samples. Furthermore, to improve the generalization capability, a parsimonious model structure is obtained by eliminating insignificant input variables for the incorporated neurofuzzy network, which is implemented by genetic algorithm (GA) based input selection method with a novel fitness evaluation function. Two numerical simulations are tested to show the effectiveness of the proposed method.

AB - Quasi-ARX neurofuzzy (Q-ARX-NF) models have shown great approximation ability and usefulness in nonlinear system identification and control. However, the incorporated neurofuzzy networks suffer from the curse-of-dimensionality problem, which may result in high computational complexity and over-fitting. In this paper, support vector regressor (SVR) based identification approach is used to reduce computational complexity with the help of transforming the original problem into Lagrange space, which is only sensitive to the number of data samples. Furthermore, to improve the generalization capability, a parsimonious model structure is obtained by eliminating insignificant input variables for the incorporated neurofuzzy network, which is implemented by genetic algorithm (GA) based input selection method with a novel fitness evaluation function. Two numerical simulations are tested to show the effectiveness of the proposed method.

KW - identification

KW - input selection

KW - Quasi-ARX neurofuzzy network

KW - SVR

UR - http://www.scopus.com/inward/record.url?scp=83755195079&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=83755195079&partnerID=8YFLogxK

U2 - 10.1109/ICSMC.2011.6083897

DO - 10.1109/ICSMC.2011.6083897

M3 - Conference contribution

AN - SCOPUS:83755195079

SN - 9781457706523

SP - 1585

EP - 1590

BT - Conference Proceedings - IEEE International Conference on Systems, Man and Cybernetics

ER -