A novel dimension reduction procedure for searching non-Gaussian subspaces

Motoaki Kawanabe; Gilles Blanchard; Masashi Sugiyama; Vladimir Spokoiny; Klaus Robert Müller

doi:10.1007/11679363_19

A novel dimension reduction procedure for searching non-Gaussian subspaces

Motoaki Kawanabe, Gilles Blanchard, Masashi Sugiyama, Vladimir Spokoiny, Klaus Robert Müller

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

Abstract

In this article, we consider high-dimensional data which contains a low-dimensional non-Gaussian structure contaminated with Gaussian noise and propose a new linear method to identify the non-Gaussian subspace. Our method NGCA (Non-Gaussian Component Analysis) is based on a very general semiparametric framework and has a theoretical guarantee that the estimation error of finding the non-Gaussian components tends to zero at a parametric rate. NGCA can be used not only as preprocessing for ICA, but also for extracting and visualizing more general structures like clusters. A numerical study demonstrates the usefulness of our method.

Original language	English
Title of host publication	Independent Component Analysis and Blind Signal Separation - 6th International Conference, ICA 2006, Proceedings
Pages	149-156
Number of pages	8
DOIs	https://doi.org/10.1007/11679363_19
Publication status	Published - 2006
Externally published	Yes
Event	6th International Conference on Independent Component Analysis and Blind Signal Separation, ICA 2006 - Charleston, SC, United States Duration: 2006 Mar 5 → 2006 Mar 8

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	3889 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Other

Other	6th International Conference on Independent Component Analysis and Blind Signal Separation, ICA 2006
Country/Territory	United States
City	Charleston, SC
Period	06/3/5 → 06/3/8

ASJC Scopus subject areas

Theoretical Computer Science
Computer Science(all)

Access to Document

10.1007/11679363_19

Cite this

Kawanabe, M., Blanchard, G., Sugiyama, M., Spokoiny, V., & Müller, K. R. (2006). A novel dimension reduction procedure for searching non-Gaussian subspaces. In Independent Component Analysis and Blind Signal Separation - 6th International Conference, ICA 2006, Proceedings (pp. 149-156). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 3889 LNCS). https://doi.org/10.1007/11679363_19

A novel dimension reduction procedure for searching non-Gaussian subspaces. / Kawanabe, Motoaki; Blanchard, Gilles; Sugiyama, Masashi et al.

Independent Component Analysis and Blind Signal Separation - 6th International Conference, ICA 2006, Proceedings. 2006. p. 149-156 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 3889 LNCS).

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

Kawanabe, M, Blanchard, G, Sugiyama, M, Spokoiny, V & Müller, KR 2006, A novel dimension reduction procedure for searching non-Gaussian subspaces. in Independent Component Analysis and Blind Signal Separation - 6th International Conference, ICA 2006, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 3889 LNCS, pp. 149-156, 6th International Conference on Independent Component Analysis and Blind Signal Separation, ICA 2006, Charleston, SC, United States, 06/3/5. https://doi.org/10.1007/11679363_19

Kawanabe M, Blanchard G, Sugiyama M, Spokoiny V, Müller KR. A novel dimension reduction procedure for searching non-Gaussian subspaces. In Independent Component Analysis and Blind Signal Separation - 6th International Conference, ICA 2006, Proceedings. 2006. p. 149-156. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/11679363_19

Kawanabe, Motoaki ; Blanchard, Gilles ; Sugiyama, Masashi et al. / A novel dimension reduction procedure for searching non-Gaussian subspaces. Independent Component Analysis and Blind Signal Separation - 6th International Conference, ICA 2006, Proceedings. 2006. pp. 149-156 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{edaafca01a854fa58501b43fd3e5763b,

title = "A novel dimension reduction procedure for searching non-Gaussian subspaces",

abstract = "In this article, we consider high-dimensional data which contains a low-dimensional non-Gaussian structure contaminated with Gaussian noise and propose a new linear method to identify the non-Gaussian subspace. Our method NGCA (Non-Gaussian Component Analysis) is based on a very general semiparametric framework and has a theoretical guarantee that the estimation error of finding the non-Gaussian components tends to zero at a parametric rate. NGCA can be used not only as preprocessing for ICA, but also for extracting and visualizing more general structures like clusters. A numerical study demonstrates the usefulness of our method.",

author = "Motoaki Kawanabe and Gilles Blanchard and Masashi Sugiyama and Vladimir Spokoiny and M{\"u}ller, {Klaus Robert}",

year = "2006",

doi = "10.1007/11679363_19",

language = "English",

isbn = "3540326308",

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

pages = "149--156",

booktitle = "Independent Component Analysis and Blind Signal Separation - 6th International Conference, ICA 2006, Proceedings",

note = "6th International Conference on Independent Component Analysis and Blind Signal Separation, ICA 2006 ; Conference date: 05-03-2006 Through 08-03-2006",

}

TY - GEN

T1 - A novel dimension reduction procedure for searching non-Gaussian subspaces

AU - Kawanabe, Motoaki

AU - Blanchard, Gilles

AU - Sugiyama, Masashi

AU - Spokoiny, Vladimir

AU - Müller, Klaus Robert

PY - 2006

Y1 - 2006

N2 - In this article, we consider high-dimensional data which contains a low-dimensional non-Gaussian structure contaminated with Gaussian noise and propose a new linear method to identify the non-Gaussian subspace. Our method NGCA (Non-Gaussian Component Analysis) is based on a very general semiparametric framework and has a theoretical guarantee that the estimation error of finding the non-Gaussian components tends to zero at a parametric rate. NGCA can be used not only as preprocessing for ICA, but also for extracting and visualizing more general structures like clusters. A numerical study demonstrates the usefulness of our method.

AB - In this article, we consider high-dimensional data which contains a low-dimensional non-Gaussian structure contaminated with Gaussian noise and propose a new linear method to identify the non-Gaussian subspace. Our method NGCA (Non-Gaussian Component Analysis) is based on a very general semiparametric framework and has a theoretical guarantee that the estimation error of finding the non-Gaussian components tends to zero at a parametric rate. NGCA can be used not only as preprocessing for ICA, but also for extracting and visualizing more general structures like clusters. A numerical study demonstrates the usefulness of our method.

UR - http://www.scopus.com/inward/record.url?scp=33745711559&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=33745711559&partnerID=8YFLogxK

U2 - 10.1007/11679363_19

DO - 10.1007/11679363_19

M3 - Conference contribution

AN - SCOPUS:33745711559

SN - 3540326308

SN - 9783540326304

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

SP - 149

EP - 156

BT - Independent Component Analysis and Blind Signal Separation - 6th International Conference, ICA 2006, Proceedings

T2 - 6th International Conference on Independent Component Analysis and Blind Signal Separation, ICA 2006

Y2 - 5 March 2006 through 8 March 2006

ER -

A novel dimension reduction procedure for searching non-Gaussian subspaces

Abstract

Publication series

Other

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this