Selecting ridge parameters in infinite dimensional hypothesis spaces

Masashi Sugiyama; Klaus Robert Müller

doi:10.1007/3-540-46084-5_86

Selecting ridge parameters in infinite dimensional hypothesis spaces

Masashi Sugiyama, Klaus Robert Müller

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

Abstract

Previously, an unbiased estimator of the generalization error called the subspace information criterion (SIC) was proposed for a finite dimensional reproducing kernel Hilbert space (RKHS). In this paper, we extend SIC so that it can be applied to any RKHSs including infinite dimensional ones. Computer simulations show that the extended SIC works well in ridge parameter selection.

Original language	English
Title of host publication	Artificial Neural Networks, ICANN 2002 - International Conference, Proceedings
Editors	Jose R. Dorronsoro, Jose R. Dorronsoro
Publisher	Springer Verlag
Pages	528-534
Number of pages	7
ISBN (Print)	9783540440741
DOIs	https://doi.org/10.1007/3-540-46084-5_86
Publication status	Published - 2002
Externally published	Yes
Event	2002 International Conference on Artificial Neural Networks, ICANN 2002 - Madrid, Spain Duration: 2002 Aug 28 → 2002 Aug 30

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	2415 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Other

Other	2002 International Conference on Artificial Neural Networks, ICANN 2002
Country/Territory	Spain
City	Madrid
Period	02/8/28 → 02/8/30

ASJC Scopus subject areas

Theoretical Computer Science
Computer Science(all)

Access to Document

10.1007/3-540-46084-5_86

Cite this

Sugiyama, M., & Müller, K. R. (2002). Selecting ridge parameters in infinite dimensional hypothesis spaces. In J. R. Dorronsoro, & J. R. Dorronsoro (Eds.), Artificial Neural Networks, ICANN 2002 - International Conference, Proceedings (pp. 528-534). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 2415 LNCS). Springer Verlag. https://doi.org/10.1007/3-540-46084-5_86

Selecting ridge parameters in infinite dimensional hypothesis spaces. / Sugiyama, Masashi; Müller, Klaus Robert.
Artificial Neural Networks, ICANN 2002 - International Conference, Proceedings. ed. / Jose R. Dorronsoro; Jose R. Dorronsoro. Springer Verlag, 2002. p. 528-534 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 2415 LNCS).

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

Sugiyama, M & Müller, KR 2002, Selecting ridge parameters in infinite dimensional hypothesis spaces. in JR Dorronsoro & JR Dorronsoro (eds), Artificial Neural Networks, ICANN 2002 - International Conference, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 2415 LNCS, Springer Verlag, pp. 528-534, 2002 International Conference on Artificial Neural Networks, ICANN 2002, Madrid, Spain, 02/8/28. https://doi.org/10.1007/3-540-46084-5_86

Sugiyama M, Müller KR. Selecting ridge parameters in infinite dimensional hypothesis spaces. In Dorronsoro JR, Dorronsoro JR, editors, Artificial Neural Networks, ICANN 2002 - International Conference, Proceedings. Springer Verlag. 2002. p. 528-534. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/3-540-46084-5_86

Sugiyama, Masashi ; Müller, Klaus Robert. / Selecting ridge parameters in infinite dimensional hypothesis spaces. Artificial Neural Networks, ICANN 2002 - International Conference, Proceedings. editor / Jose R. Dorronsoro ; Jose R. Dorronsoro. Springer Verlag, 2002. pp. 528-534 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{f97911b3b13b4bcfb1b72cdefda522db,

title = "Selecting ridge parameters in infinite dimensional hypothesis spaces",

abstract = "Previously, an unbiased estimator of the generalization error called the subspace information criterion (SIC) was proposed for a finite dimensional reproducing kernel Hilbert space (RKHS). In this paper, we extend SIC so that it can be applied to any RKHSs including infinite dimensional ones. Computer simulations show that the extended SIC works well in ridge parameter selection.",

author = "Masashi Sugiyama and M{\"u}ller, {Klaus Robert}",

year = "2002",

doi = "10.1007/3-540-46084-5_86",

language = "English",

isbn = "9783540440741",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Verlag",

pages = "528--534",

editor = "Dorronsoro, {Jose R.} and Dorronsoro, {Jose R.}",

booktitle = "Artificial Neural Networks, ICANN 2002 - International Conference, Proceedings",

note = "2002 International Conference on Artificial Neural Networks, ICANN 2002 ; Conference date: 28-08-2002 Through 30-08-2002",

}

TY - GEN

T1 - Selecting ridge parameters in infinite dimensional hypothesis spaces

AU - Sugiyama, Masashi

AU - Müller, Klaus Robert

PY - 2002

Y1 - 2002

N2 - Previously, an unbiased estimator of the generalization error called the subspace information criterion (SIC) was proposed for a finite dimensional reproducing kernel Hilbert space (RKHS). In this paper, we extend SIC so that it can be applied to any RKHSs including infinite dimensional ones. Computer simulations show that the extended SIC works well in ridge parameter selection.

AB - Previously, an unbiased estimator of the generalization error called the subspace information criterion (SIC) was proposed for a finite dimensional reproducing kernel Hilbert space (RKHS). In this paper, we extend SIC so that it can be applied to any RKHSs including infinite dimensional ones. Computer simulations show that the extended SIC works well in ridge parameter selection.

UR - http://www.scopus.com/inward/record.url?scp=84902208989&partnerID=8YFLogxK

U2 - 10.1007/3-540-46084-5_86

DO - 10.1007/3-540-46084-5_86

M3 - Conference contribution

AN - SCOPUS:84902208989

SN - 9783540440741

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 528

EP - 534

BT - Artificial Neural Networks, ICANN 2002 - International Conference, Proceedings

A2 - Dorronsoro, Jose R.

PB - Springer Verlag

T2 - 2002 International Conference on Artificial Neural Networks, ICANN 2002

Y2 - 28 August 2002 through 30 August 2002

ER -

Selecting ridge parameters in infinite dimensional hypothesis spaces

Abstract

Publication series

Other

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this