Constructing optimal fuzzy models using statistical information criteria

John Yen, Liang Wang

Research output: Contribution to journalArticle

8 Citations (Scopus)

Abstract

Theoretical studies have shown that fuzzy models are capable of approximating any continuous function on a compact domain to any degree of accuracy. However, good performance in approximation does not necessarily assure good performance in prediction or control. A fuzzy model with a large number of fuzzy rules may have a low accuracy of estimation for the unknown parameters. This is especially true when only limited sample data are available in building the model. Further, such a model often encounters the risk of overfitting the data and thus has a poor ability of generalization. A trade-off is thus required in building a fuzzy model: on the one hand, the number of fuzzy rules must be sufficient to provide the discriminating capability required for the given application; on the other hand, the number of fuzzy rules must be "parsimonious" to guarantee a reasonable accuracy of parameter estimation and a good ability of generalizing to unknown patterns. In this paper we apply statistical information criteria for achieving such a trade-off. In particular, we combine these criteria with an SVD (singular value decomposition) based fuzzy rule selection method to choose the optimal number of fuzzy rules and construct the "best" fuzzy model. The role of these criteria in fuzzy modeling is discussed and their practical applicability is illustrated using a nonlinear system modeling example.

Original languageEnglish (US)
Pages (from-to)185-201
Number of pages17
JournalJournal of Intelligent and Fuzzy Systems
Volume7
Issue number2
StatePublished - Dec 1 1999

Fingerprint

Information Criterion
Fuzzy Model
Fuzzy Rules
Fuzzy rules
Trade-offs
Nonlinear Modeling
Fuzzy Modeling
Overfitting
Singular value decomposition
System Modeling
Unknown Parameters
Parameter Estimation
Continuous Function
Parameter estimation
Nonlinear Systems
Choose
Statistical Models
Nonlinear systems
Sufficient
Unknown

All Science Journal Classification (ASJC) codes

  • Statistics and Probability
  • Engineering(all)
  • Artificial Intelligence

Cite this

@article{e636c62c45464ab38fa96556d91b2345,
title = "Constructing optimal fuzzy models using statistical information criteria",
abstract = "Theoretical studies have shown that fuzzy models are capable of approximating any continuous function on a compact domain to any degree of accuracy. However, good performance in approximation does not necessarily assure good performance in prediction or control. A fuzzy model with a large number of fuzzy rules may have a low accuracy of estimation for the unknown parameters. This is especially true when only limited sample data are available in building the model. Further, such a model often encounters the risk of overfitting the data and thus has a poor ability of generalization. A trade-off is thus required in building a fuzzy model: on the one hand, the number of fuzzy rules must be sufficient to provide the discriminating capability required for the given application; on the other hand, the number of fuzzy rules must be {"}parsimonious{"} to guarantee a reasonable accuracy of parameter estimation and a good ability of generalizing to unknown patterns. In this paper we apply statistical information criteria for achieving such a trade-off. In particular, we combine these criteria with an SVD (singular value decomposition) based fuzzy rule selection method to choose the optimal number of fuzzy rules and construct the {"}best{"} fuzzy model. The role of these criteria in fuzzy modeling is discussed and their practical applicability is illustrated using a nonlinear system modeling example.",
author = "John Yen and Liang Wang",
year = "1999",
month = "12",
day = "1",
language = "English (US)",
volume = "7",
pages = "185--201",
journal = "Journal of Intelligent and Fuzzy Systems",
issn = "1064-1246",
publisher = "IOS Press",
number = "2",

}

Constructing optimal fuzzy models using statistical information criteria. / Yen, John; Wang, Liang.

In: Journal of Intelligent and Fuzzy Systems, Vol. 7, No. 2, 01.12.1999, p. 185-201.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Constructing optimal fuzzy models using statistical information criteria

AU - Yen, John

AU - Wang, Liang

PY - 1999/12/1

Y1 - 1999/12/1

N2 - Theoretical studies have shown that fuzzy models are capable of approximating any continuous function on a compact domain to any degree of accuracy. However, good performance in approximation does not necessarily assure good performance in prediction or control. A fuzzy model with a large number of fuzzy rules may have a low accuracy of estimation for the unknown parameters. This is especially true when only limited sample data are available in building the model. Further, such a model often encounters the risk of overfitting the data and thus has a poor ability of generalization. A trade-off is thus required in building a fuzzy model: on the one hand, the number of fuzzy rules must be sufficient to provide the discriminating capability required for the given application; on the other hand, the number of fuzzy rules must be "parsimonious" to guarantee a reasonable accuracy of parameter estimation and a good ability of generalizing to unknown patterns. In this paper we apply statistical information criteria for achieving such a trade-off. In particular, we combine these criteria with an SVD (singular value decomposition) based fuzzy rule selection method to choose the optimal number of fuzzy rules and construct the "best" fuzzy model. The role of these criteria in fuzzy modeling is discussed and their practical applicability is illustrated using a nonlinear system modeling example.

AB - Theoretical studies have shown that fuzzy models are capable of approximating any continuous function on a compact domain to any degree of accuracy. However, good performance in approximation does not necessarily assure good performance in prediction or control. A fuzzy model with a large number of fuzzy rules may have a low accuracy of estimation for the unknown parameters. This is especially true when only limited sample data are available in building the model. Further, such a model often encounters the risk of overfitting the data and thus has a poor ability of generalization. A trade-off is thus required in building a fuzzy model: on the one hand, the number of fuzzy rules must be sufficient to provide the discriminating capability required for the given application; on the other hand, the number of fuzzy rules must be "parsimonious" to guarantee a reasonable accuracy of parameter estimation and a good ability of generalizing to unknown patterns. In this paper we apply statistical information criteria for achieving such a trade-off. In particular, we combine these criteria with an SVD (singular value decomposition) based fuzzy rule selection method to choose the optimal number of fuzzy rules and construct the "best" fuzzy model. The role of these criteria in fuzzy modeling is discussed and their practical applicability is illustrated using a nonlinear system modeling example.

UR - http://www.scopus.com/inward/record.url?scp=0033419688&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0033419688&partnerID=8YFLogxK

M3 - Article

AN - SCOPUS:0033419688

VL - 7

SP - 185

EP - 201

JO - Journal of Intelligent and Fuzzy Systems

JF - Journal of Intelligent and Fuzzy Systems

SN - 1064-1246

IS - 2

ER -