A global-local learning algorithm for identifying Takagi-Sugeno-Kang fuzzy models

John Yen, Liang Wang, Wayne Gillespie

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

3 Citations (Scopus)

Abstract

The fuzzy inference system proposed by Takagi, Sugeno and Kang, which is known as the TSK model in fuzzy system literature, provides a powerful tool for modeling complex nonlinear systems. Unlike conventional modeling where a single model is used to describe the global behavior of a system, TSK modeling is essentially a multi-model approach in which simple submodels combined to describe the global behavior of the system. Most existing learning algorithms for identifying the TSK model are based on minimizing the square of the residual between the overall outputs of the real system and the identified model. Although these algorithms can generate a TSK model with good global performance, they cannot guarantee the resulting model to have a good local performance. Often, the submodels in the TSK model may exhibit an erratic local behavior which is difficult to interpret. Since one of the important motivations of using the TSK model is to gain insights into the model, it is important to investigate the interpretability issue of the TSK model. We propose a learning algorithm that integrates global learning and local learning in a single algorithmic framework. This algorithm uses the idea of local weighted regression and local approximation in nonparametric statistics, but remains the component of global fitting in the existing learning algorithms. The algorithm is capable of adjusting its parameters based on the user's preference, generating models with good trade-off in terms of global fitting and local interpretation. We illustrate the performance of the proposed algorithm using a motorcycle crash modeling example.

Original languageEnglish (US)
Title of host publication1998 IEEE International Conference on Fuzzy Systems Proceedings - IEEE World Congress on Computational Intelligence
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages967-972
Number of pages6
ISBN (Print)078034863X, 9780780348639
DOIs
StatePublished - Jan 1 1998
Event1998 IEEE International Conference on Fuzzy Systems, FUZZY 1998 - Anchorage, United States
Duration: May 4 1998May 9 1998

Publication series

Name1998 IEEE International Conference on Fuzzy Systems Proceedings - IEEE World Congress on Computational Intelligence
Volume2

Other

Other1998 IEEE International Conference on Fuzzy Systems, FUZZY 1998
CountryUnited States
CityAnchorage
Period5/4/985/9/98

Fingerprint

Local Algorithms
Takagi-Sugeno Fuzzy Model
Learning algorithms
Learning Algorithm
Model
Modeling
Nonparametric Statistics
Multi-model
Local Approximation
Fuzzy Inference System
Interpretability
Motorcycles
User Preferences
Crash
Fuzzy inference
System Modeling
Fuzzy systems
Fuzzy Systems
Complex Systems
Nonlinear systems

All Science Journal Classification (ASJC) codes

  • Logic
  • Control and Optimization
  • Modeling and Simulation
  • Chemical Health and Safety
  • Software
  • Safety, Risk, Reliability and Quality

Cite this

Yen, J., Wang, L., & Gillespie, W. (1998). A global-local learning algorithm for identifying Takagi-Sugeno-Kang fuzzy models. In 1998 IEEE International Conference on Fuzzy Systems Proceedings - IEEE World Congress on Computational Intelligence (pp. 967-972). [686249] (1998 IEEE International Conference on Fuzzy Systems Proceedings - IEEE World Congress on Computational Intelligence; Vol. 2). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/FUZZY.1998.686249
Yen, John ; Wang, Liang ; Gillespie, Wayne. / A global-local learning algorithm for identifying Takagi-Sugeno-Kang fuzzy models. 1998 IEEE International Conference on Fuzzy Systems Proceedings - IEEE World Congress on Computational Intelligence. Institute of Electrical and Electronics Engineers Inc., 1998. pp. 967-972 (1998 IEEE International Conference on Fuzzy Systems Proceedings - IEEE World Congress on Computational Intelligence).
@inproceedings{a6dec687ddbb40ad9a5364605d12433e,
title = "A global-local learning algorithm for identifying Takagi-Sugeno-Kang fuzzy models",
abstract = "The fuzzy inference system proposed by Takagi, Sugeno and Kang, which is known as the TSK model in fuzzy system literature, provides a powerful tool for modeling complex nonlinear systems. Unlike conventional modeling where a single model is used to describe the global behavior of a system, TSK modeling is essentially a multi-model approach in which simple submodels combined to describe the global behavior of the system. Most existing learning algorithms for identifying the TSK model are based on minimizing the square of the residual between the overall outputs of the real system and the identified model. Although these algorithms can generate a TSK model with good global performance, they cannot guarantee the resulting model to have a good local performance. Often, the submodels in the TSK model may exhibit an erratic local behavior which is difficult to interpret. Since one of the important motivations of using the TSK model is to gain insights into the model, it is important to investigate the interpretability issue of the TSK model. We propose a learning algorithm that integrates global learning and local learning in a single algorithmic framework. This algorithm uses the idea of local weighted regression and local approximation in nonparametric statistics, but remains the component of global fitting in the existing learning algorithms. The algorithm is capable of adjusting its parameters based on the user's preference, generating models with good trade-off in terms of global fitting and local interpretation. We illustrate the performance of the proposed algorithm using a motorcycle crash modeling example.",
author = "John Yen and Liang Wang and Wayne Gillespie",
year = "1998",
month = "1",
day = "1",
doi = "10.1109/FUZZY.1998.686249",
language = "English (US)",
isbn = "078034863X",
series = "1998 IEEE International Conference on Fuzzy Systems Proceedings - IEEE World Congress on Computational Intelligence",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
pages = "967--972",
booktitle = "1998 IEEE International Conference on Fuzzy Systems Proceedings - IEEE World Congress on Computational Intelligence",
address = "United States",

}

Yen, J, Wang, L & Gillespie, W 1998, A global-local learning algorithm for identifying Takagi-Sugeno-Kang fuzzy models. in 1998 IEEE International Conference on Fuzzy Systems Proceedings - IEEE World Congress on Computational Intelligence., 686249, 1998 IEEE International Conference on Fuzzy Systems Proceedings - IEEE World Congress on Computational Intelligence, vol. 2, Institute of Electrical and Electronics Engineers Inc., pp. 967-972, 1998 IEEE International Conference on Fuzzy Systems, FUZZY 1998, Anchorage, United States, 5/4/98. https://doi.org/10.1109/FUZZY.1998.686249

A global-local learning algorithm for identifying Takagi-Sugeno-Kang fuzzy models. / Yen, John; Wang, Liang; Gillespie, Wayne.

1998 IEEE International Conference on Fuzzy Systems Proceedings - IEEE World Congress on Computational Intelligence. Institute of Electrical and Electronics Engineers Inc., 1998. p. 967-972 686249 (1998 IEEE International Conference on Fuzzy Systems Proceedings - IEEE World Congress on Computational Intelligence; Vol. 2).

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

TY - GEN

T1 - A global-local learning algorithm for identifying Takagi-Sugeno-Kang fuzzy models

AU - Yen, John

AU - Wang, Liang

AU - Gillespie, Wayne

PY - 1998/1/1

Y1 - 1998/1/1

N2 - The fuzzy inference system proposed by Takagi, Sugeno and Kang, which is known as the TSK model in fuzzy system literature, provides a powerful tool for modeling complex nonlinear systems. Unlike conventional modeling where a single model is used to describe the global behavior of a system, TSK modeling is essentially a multi-model approach in which simple submodels combined to describe the global behavior of the system. Most existing learning algorithms for identifying the TSK model are based on minimizing the square of the residual between the overall outputs of the real system and the identified model. Although these algorithms can generate a TSK model with good global performance, they cannot guarantee the resulting model to have a good local performance. Often, the submodels in the TSK model may exhibit an erratic local behavior which is difficult to interpret. Since one of the important motivations of using the TSK model is to gain insights into the model, it is important to investigate the interpretability issue of the TSK model. We propose a learning algorithm that integrates global learning and local learning in a single algorithmic framework. This algorithm uses the idea of local weighted regression and local approximation in nonparametric statistics, but remains the component of global fitting in the existing learning algorithms. The algorithm is capable of adjusting its parameters based on the user's preference, generating models with good trade-off in terms of global fitting and local interpretation. We illustrate the performance of the proposed algorithm using a motorcycle crash modeling example.

AB - The fuzzy inference system proposed by Takagi, Sugeno and Kang, which is known as the TSK model in fuzzy system literature, provides a powerful tool for modeling complex nonlinear systems. Unlike conventional modeling where a single model is used to describe the global behavior of a system, TSK modeling is essentially a multi-model approach in which simple submodels combined to describe the global behavior of the system. Most existing learning algorithms for identifying the TSK model are based on minimizing the square of the residual between the overall outputs of the real system and the identified model. Although these algorithms can generate a TSK model with good global performance, they cannot guarantee the resulting model to have a good local performance. Often, the submodels in the TSK model may exhibit an erratic local behavior which is difficult to interpret. Since one of the important motivations of using the TSK model is to gain insights into the model, it is important to investigate the interpretability issue of the TSK model. We propose a learning algorithm that integrates global learning and local learning in a single algorithmic framework. This algorithm uses the idea of local weighted regression and local approximation in nonparametric statistics, but remains the component of global fitting in the existing learning algorithms. The algorithm is capable of adjusting its parameters based on the user's preference, generating models with good trade-off in terms of global fitting and local interpretation. We illustrate the performance of the proposed algorithm using a motorcycle crash modeling example.

UR - http://www.scopus.com/inward/record.url?scp=0031630894&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0031630894&partnerID=8YFLogxK

U2 - 10.1109/FUZZY.1998.686249

DO - 10.1109/FUZZY.1998.686249

M3 - Conference contribution

AN - SCOPUS:0031630894

SN - 078034863X

SN - 9780780348639

T3 - 1998 IEEE International Conference on Fuzzy Systems Proceedings - IEEE World Congress on Computational Intelligence

SP - 967

EP - 972

BT - 1998 IEEE International Conference on Fuzzy Systems Proceedings - IEEE World Congress on Computational Intelligence

PB - Institute of Electrical and Electronics Engineers Inc.

ER -

Yen J, Wang L, Gillespie W. A global-local learning algorithm for identifying Takagi-Sugeno-Kang fuzzy models. In 1998 IEEE International Conference on Fuzzy Systems Proceedings - IEEE World Congress on Computational Intelligence. Institute of Electrical and Electronics Engineers Inc. 1998. p. 967-972. 686249. (1998 IEEE International Conference on Fuzzy Systems Proceedings - IEEE World Congress on Computational Intelligence). https://doi.org/10.1109/FUZZY.1998.686249