Identification of NARX Hammerstein models based on large scale Support Vector Machine regression

Mujahed Al Dhaifallah; David T. Westwick

doi:10.3182/20090706-3-FR-2004.0396

Identification of NARX Hammerstein models based on large scale Support Vector Machine regression

Mujahed Al Dhaifallah^*
, David T. Westwick

^*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

1 Scopus citations

Abstract

Methods for the identification of Hammerstein models consisting of a Support Vector Machine nonlinearity followed by an ARX linear system are developed. The models are identified by minimizing epsilon insensitive cost functions based on either the sum of absolute residuals or the sum of squared residuals. Large scale implementations of these techniques are then derived using subset selection methods, and used to identify a model of the stretch reflex electromyogram from experimental data. The effects of the various cost functions and tuning parameters are demonstrated with the experimental results.

Original language	English
Title of host publication	15th Symposium on System Identification, SYSID 2009 - Preprints
Pages	1656-1661
Number of pages	6
Edition	PART 1
DOIs	https://doi.org/10.3182/20090706-3-FR-2004.0396
State	Published - 2009
Externally published	Yes

Publication series

Name	IFAC Proceedings Volumes (IFAC-PapersOnline)
Number	PART 1
Volume	15
ISSN (Print)	1474-6670

Bibliographical note

Funding Information:
This work is supported by NSERC (Canada).

Keywords

Hammerstein
Identification
Support vector machines

ASJC Scopus subject areas

Control and Systems Engineering

Access to Document

10.3182/20090706-3-FR-2004.0396

Cite this

@inproceedings{2a9cd08049f3428e8ea3dab32dfe8ff1,

title = "Identification of NARX Hammerstein models based on large scale Support Vector Machine regression",

abstract = "Methods for the identification of Hammerstein models consisting of a Support Vector Machine nonlinearity followed by an ARX linear system are developed. The models are identified by minimizing epsilon insensitive cost functions based on either the sum of absolute residuals or the sum of squared residuals. Large scale implementations of these techniques are then derived using subset selection methods, and used to identify a model of the stretch reflex electromyogram from experimental data. The effects of the various cost functions and tuning parameters are demonstrated with the experimental results.",

keywords = "Hammerstein, Identification, Support vector machines",

author = "\{Al Dhaifallah\}, Mujahed and Westwick, \{David T.\}",

note = "Funding Information: This work is supported by NSERC (Canada).",

year = "2009",

doi = "10.3182/20090706-3-FR-2004.0396",

language = "English",

isbn = "9783902661470",

series = "IFAC Proceedings Volumes (IFAC-PapersOnline)",

number = "PART 1",

pages = "1656--1661",

booktitle = "15th Symposium on System Identification, SYSID 2009 - Preprints",

edition = "PART 1",

}

Identification of NARX Hammerstein models based on large scale Support Vector Machine regression. / Al Dhaifallah, Mujahed; Westwick, David T.
15th Symposium on System Identification, SYSID 2009 - Preprints. PART 1. ed. 2009. p. 1656-1661 (IFAC Proceedings Volumes (IFAC-PapersOnline); Vol. 15, No. PART 1).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Identification of NARX Hammerstein models based on large scale Support Vector Machine regression

AU - Al Dhaifallah, Mujahed

AU - Westwick, David T.

N1 - Funding Information: This work is supported by NSERC (Canada).

PY - 2009

Y1 - 2009

N2 - Methods for the identification of Hammerstein models consisting of a Support Vector Machine nonlinearity followed by an ARX linear system are developed. The models are identified by minimizing epsilon insensitive cost functions based on either the sum of absolute residuals or the sum of squared residuals. Large scale implementations of these techniques are then derived using subset selection methods, and used to identify a model of the stretch reflex electromyogram from experimental data. The effects of the various cost functions and tuning parameters are demonstrated with the experimental results.

AB - Methods for the identification of Hammerstein models consisting of a Support Vector Machine nonlinearity followed by an ARX linear system are developed. The models are identified by minimizing epsilon insensitive cost functions based on either the sum of absolute residuals or the sum of squared residuals. Large scale implementations of these techniques are then derived using subset selection methods, and used to identify a model of the stretch reflex electromyogram from experimental data. The effects of the various cost functions and tuning parameters are demonstrated with the experimental results.

KW - Hammerstein

KW - Identification

KW - Support vector machines

UR - https://www.scopus.com/pages/publications/80051651600

U2 - 10.3182/20090706-3-FR-2004.0396

DO - 10.3182/20090706-3-FR-2004.0396

M3 - Conference contribution

AN - SCOPUS:80051651600

SN - 9783902661470

T3 - IFAC Proceedings Volumes (IFAC-PapersOnline)

SP - 1656

EP - 1661

BT - 15th Symposium on System Identification, SYSID 2009 - Preprints

ER -

Identification of NARX Hammerstein models based on large scale Support Vector Machine regression

Abstract

Publication series

Bibliographical note

Keywords

ASJC Scopus subject areas

Access to Document

Fingerprint

Cite this