On the choices of the parameters in general constrained learning algorithms

De-Shuang Huang; Horace H. S. Ip

doi:10.1007/978-3-540-45080-1_137

On the choices of the parameters in general constrained learning algorithms

De-Shuang Huang
, Horace H. S. Ip

Department of Computer Science

Research output: Chapters, Conference Papers, Creative and Literary Works › RGC 32 - Refereed conference paper (with host publication) › peer-review

3 Citations (Scopus)

Abstract

This paper addresses the constrained learning algorithm (CLA) proposed by Perantonis et al, which is an efficient and fast back propagation (BP) algorithm formed by imposing the constraint condition, referred to as the a priori information, implicit in the issues into the conventional BP algorithm. It is found, through analyzing the CLA, that the choice of the values of the three learning parameters {δP, θ_p, η} in the algorithm is critical to successful application of the technique. Otherwise, the algorithm will not be able to converge within a limited time, or even diverge. This paper will discuss how to choose the three learning parameters based on an exhaustive understanding on the CLA. Finally, several computer simulation results show that our analyses and conclusions are completely correct. © Springer-Verlag 2003.

Original language	English
Title of host publication	Intelligent Data Engineering and Automated Learning
Subtitle of host publication	4th International Conference, IDEAL 2003 Hong Kong, China, March 21–23, 2003 Revised Papers
Editors	Jiming Liu, Yiu-ming Cheung, Hujun Yin
Place of Publication	Berlin, Heidelberg
Publisher	Springer
Pages	967-974
ISBN (Electronic)	978-3-540-45080-1
ISBN (Print)	978-3-540-40550-4
DOIs	https://doi.org/10.1007/978-3-540-45080-1_137
Publication status	Published - 2004
Event	4th International Conference on Intelligent Data Engineering and Automated Learning (IDEAL 2003) - The Hong Kong Convention and Exhibition Centre, Hong Kong, China Duration: 21 Mar 2003 → 23 Mar 2003

Publication series

Name	Lecture Notes in Computer Science
Volume	2690
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	4th International Conference on Intelligent Data Engineering and Automated Learning (IDEAL 2003)
Place	China
City	Hong Kong
Period	21/03/03 → 23/03/03

Access Output

10.1007/978-3-540-45080-1_137

Other Access Options

Check CityUHK Library

Permanent Link

Right click to copy link

Cite this

Huang, D.-S., & Ip, H. H. S. (2004). On the choices of the parameters in general constrained learning algorithms. In J. Liu, Y. Cheung, & H. Yin (Eds.), Intelligent Data Engineering and Automated Learning: 4th International Conference, IDEAL 2003 Hong Kong, China, March 21–23, 2003 Revised Papers (pp. 967-974). (Lecture Notes in Computer Science; Vol. 2690). Springer . https://doi.org/10.1007/978-3-540-45080-1_137

Huang, De-Shuang ; Ip, Horace H. S. / On the choices of the parameters in general constrained learning algorithms. Intelligent Data Engineering and Automated Learning: 4th International Conference, IDEAL 2003 Hong Kong, China, March 21–23, 2003 Revised Papers. editor / Jiming Liu ; Yiu-ming Cheung ; Hujun Yin. Berlin, Heidelberg : Springer , 2004. pp. 967-974 (Lecture Notes in Computer Science).

@inproceedings{3863da72926842b48162a8e7191d9ce0,

title = "On the choices of the parameters in general constrained learning algorithms",

abstract = "This paper addresses the constrained learning algorithm (CLA) proposed by Perantonis et al, which is an efficient and fast back propagation (BP) algorithm formed by imposing the constraint condition, referred to as the a priori information, implicit in the issues into the conventional BP algorithm. It is found, through analyzing the CLA, that the choice of the values of the three learning parameters \{δP, θp, η\} in the algorithm is critical to successful application of the technique. Otherwise, the algorithm will not be able to converge within a limited time, or even diverge. This paper will discuss how to choose the three learning parameters based on an exhaustive understanding on the CLA. Finally, several computer simulation results show that our analyses and conclusions are completely correct. {\textcopyright} Springer-Verlag 2003.",

author = "De-Shuang Huang and Ip, \{Horace H. S.\}",

year = "2004",

doi = "10.1007/978-3-540-45080-1\_137",

language = "English",

isbn = "978-3-540-40550-4",

series = "Lecture Notes in Computer Science",

publisher = "Springer ",

pages = "967--974",

editor = "Liu, \{Jiming \} and Cheung, \{Yiu-ming \} and Hujun Yin",

booktitle = "Intelligent Data Engineering and Automated Learning",

note = "4th International Conference on Intelligent Data Engineering and Automated Learning (IDEAL 2003) ; Conference date: 21-03-2003 Through 23-03-2003",

}

Huang, D-S & Ip, HHS 2004, On the choices of the parameters in general constrained learning algorithms. in J Liu, Y Cheung & H Yin (eds), Intelligent Data Engineering and Automated Learning: 4th International Conference, IDEAL 2003 Hong Kong, China, March 21–23, 2003 Revised Papers. Lecture Notes in Computer Science, vol. 2690, Springer , Berlin, Heidelberg, pp. 967-974, 4th International Conference on Intelligent Data Engineering and Automated Learning (IDEAL 2003), Hong Kong, China, 21/03/03. https://doi.org/10.1007/978-3-540-45080-1_137

On the choices of the parameters in general constrained learning algorithms. / Huang, De-Shuang; Ip, Horace H. S.
Intelligent Data Engineering and Automated Learning: 4th International Conference, IDEAL 2003 Hong Kong, China, March 21–23, 2003 Revised Papers. ed. / Jiming Liu; Yiu-ming Cheung; Hujun Yin. Berlin, Heidelberg: Springer , 2004. p. 967-974 (Lecture Notes in Computer Science; Vol. 2690).

Research output: Chapters, Conference Papers, Creative and Literary Works › RGC 32 - Refereed conference paper (with host publication) › peer-review

TY - GEN

T1 - On the choices of the parameters in general constrained learning algorithms

AU - Huang, De-Shuang

AU - Ip, Horace H. S.

PY - 2004

Y1 - 2004

N2 - This paper addresses the constrained learning algorithm (CLA) proposed by Perantonis et al, which is an efficient and fast back propagation (BP) algorithm formed by imposing the constraint condition, referred to as the a priori information, implicit in the issues into the conventional BP algorithm. It is found, through analyzing the CLA, that the choice of the values of the three learning parameters {δP, θp, η} in the algorithm is critical to successful application of the technique. Otherwise, the algorithm will not be able to converge within a limited time, or even diverge. This paper will discuss how to choose the three learning parameters based on an exhaustive understanding on the CLA. Finally, several computer simulation results show that our analyses and conclusions are completely correct. © Springer-Verlag 2003.

AB - This paper addresses the constrained learning algorithm (CLA) proposed by Perantonis et al, which is an efficient and fast back propagation (BP) algorithm formed by imposing the constraint condition, referred to as the a priori information, implicit in the issues into the conventional BP algorithm. It is found, through analyzing the CLA, that the choice of the values of the three learning parameters {δP, θp, η} in the algorithm is critical to successful application of the technique. Otherwise, the algorithm will not be able to converge within a limited time, or even diverge. This paper will discuss how to choose the three learning parameters based on an exhaustive understanding on the CLA. Finally, several computer simulation results show that our analyses and conclusions are completely correct. © Springer-Verlag 2003.

UR - https://www.scopus.com/pages/publications/35048881503

UR - https://www.scopus.com/record/pubmetrics.uri?eid=2-s2.0-35048881503&origin=recordpage

U2 - 10.1007/978-3-540-45080-1_137

DO - 10.1007/978-3-540-45080-1_137

M3 - RGC 32 - Refereed conference paper (with host publication)

SN - 978-3-540-40550-4

T3 - Lecture Notes in Computer Science

SP - 967

EP - 974

BT - Intelligent Data Engineering and Automated Learning

A2 - Liu, Jiming

A2 - Cheung, Yiu-ming

A2 - Yin, Hujun

PB - Springer

CY - Berlin, Heidelberg

T2 - 4th International Conference on Intelligent Data Engineering and Automated Learning (IDEAL 2003)

Y2 - 21 March 2003 through 23 March 2003

ER -

Huang DS, Ip HHS. On the choices of the parameters in general constrained learning algorithms. In Liu J, Cheung Y, Yin H, editors, Intelligent Data Engineering and Automated Learning: 4th International Conference, IDEAL 2003 Hong Kong, China, March 21–23, 2003 Revised Papers. Berlin, Heidelberg: Springer . 2004. p. 967-974. (Lecture Notes in Computer Science). doi: 10.1007/978-3-540-45080-1_137

On the choices of the parameters in general constrained learning algorithms

Abstract

Publication series

Conference

Access Output

Other Access Options

Permanent Link

Other Files and Links

Fingerprint

Cite this