Variational extensions to EM and multinomial PCA

Wray Buntine

Variational extensions to EM and multinomial PCA

Wray Buntine

Department of Data Science & AI

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

65 Citations (Scopus)

Abstract

Several authors in recent years have proposed discrete analogues to principle component analysis intended to handle discrete or positive only data, for instance suited to analyzing sets of documents. Methods include non-negative matrix factorization, probabilistic latent semantic analysis, and latent Dirichlet allocation. This paper begins with a review of the basic theory of the variational extension to the expectation-maximization algorithm, and then presents discrete component finding algorithms in that light. Experiments are conducted on both bigram word data and document bag-of-word to expose some of the subtleties of this new class of algorithms.

Original language	English
Title of host publication	Machine Learning
Subtitle of host publication	ECML 2002 - 13th European Conference on Machine Learning, Proceedings
Editors	Tapio Elomaa, Heikki Mannila, Hannu Toivonen
Publisher	Springer-Verlag London Ltd.
Pages	23-34
Number of pages	12
ISBN (Print)	9783540440369
Publication status	Published - 1 Jan 2002
Event	European Conference on Machine Learning 2002 - Helsinki, Finland Duration: 19 Aug 2002 → 23 Aug 2002 Conference number: 13th https://link.springer.com/book/10.1007/3-540-36755-1 (Proceedings)

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	2430
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	European Conference on Machine Learning 2002
Abbreviated title	ECML 2002
Country/Territory	Finland
City	Helsinki
Period	19/08/02 → 23/08/02
Internet address	https://link.springer.com/book/10.1007/3-540-36755-1 (Proceedings)

Cite this

Buntine, W. (2002). Variational extensions to EM and multinomial PCA. In T. Elomaa, H. Mannila, & H. Toivonen (Eds.), Machine Learning: ECML 2002 - 13th European Conference on Machine Learning, Proceedings (pp. 23-34). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 2430). Springer-Verlag London Ltd..

Buntine, Wray. / Variational extensions to EM and multinomial PCA. Machine Learning: ECML 2002 - 13th European Conference on Machine Learning, Proceedings. editor / Tapio Elomaa ; Heikki Mannila ; Hannu Toivonen. Springer-Verlag London Ltd., 2002. pp. 23-34 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{3c0a74a224a0484cb2b604b01ebae4b4,

title = "Variational extensions to EM and multinomial PCA",

abstract = "Several authors in recent years have proposed discrete analogues to principle component analysis intended to handle discrete or positive only data, for instance suited to analyzing sets of documents. Methods include non-negative matrix factorization, probabilistic latent semantic analysis, and latent Dirichlet allocation. This paper begins with a review of the basic theory of the variational extension to the expectation-maximization algorithm, and then presents discrete component finding algorithms in that light. Experiments are conducted on both bigram word data and document bag-of-word to expose some of the subtleties of this new class of algorithms.",

author = "Wray Buntine",

year = "2002",

month = jan,

day = "1",

language = "English",

isbn = "9783540440369",

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

publisher = "Springer-Verlag London Ltd.",

pages = "23--34",

editor = "Tapio Elomaa and Heikki Mannila and Hannu Toivonen",

booktitle = "Machine Learning",

address = "Germany",

note = "European Conference on Machine Learning 2002, ECML 2002 ; Conference date: 19-08-2002 Through 23-08-2002",

url = "https://link.springer.com/book/10.1007/3-540-36755-1",

}

Buntine, W 2002, Variational extensions to EM and multinomial PCA. in T Elomaa, H Mannila & H Toivonen (eds), Machine Learning: ECML 2002 - 13th European Conference on Machine Learning, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 2430, Springer-Verlag London Ltd., pp. 23-34, European Conference on Machine Learning 2002, Helsinki, Finland, 19/08/02.

Variational extensions to EM and multinomial PCA. / Buntine, Wray.
Machine Learning: ECML 2002 - 13th European Conference on Machine Learning, Proceedings. ed. / Tapio Elomaa; Heikki Mannila; Hannu Toivonen. Springer-Verlag London Ltd., 2002. p. 23-34 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 2430).

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

TY - GEN

T1 - Variational extensions to EM and multinomial PCA

AU - Buntine, Wray

N1 - Conference code: 13th

PY - 2002/1/1

Y1 - 2002/1/1

N2 - Several authors in recent years have proposed discrete analogues to principle component analysis intended to handle discrete or positive only data, for instance suited to analyzing sets of documents. Methods include non-negative matrix factorization, probabilistic latent semantic analysis, and latent Dirichlet allocation. This paper begins with a review of the basic theory of the variational extension to the expectation-maximization algorithm, and then presents discrete component finding algorithms in that light. Experiments are conducted on both bigram word data and document bag-of-word to expose some of the subtleties of this new class of algorithms.

AB - Several authors in recent years have proposed discrete analogues to principle component analysis intended to handle discrete or positive only data, for instance suited to analyzing sets of documents. Methods include non-negative matrix factorization, probabilistic latent semantic analysis, and latent Dirichlet allocation. This paper begins with a review of the basic theory of the variational extension to the expectation-maximization algorithm, and then presents discrete component finding algorithms in that light. Experiments are conducted on both bigram word data and document bag-of-word to expose some of the subtleties of this new class of algorithms.

UR - http://www.scopus.com/inward/record.url?scp=84945258136&partnerID=8YFLogxK

M3 - Conference Paper

AN - SCOPUS:84945258136

SN - 9783540440369

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

SP - 23

EP - 34

BT - Machine Learning

A2 - Elomaa, Tapio

A2 - Mannila, Heikki

A2 - Toivonen, Hannu

PB - Springer-Verlag London Ltd.

T2 - European Conference on Machine Learning 2002

Y2 - 19 August 2002 through 23 August 2002

ER -

Buntine W. Variational extensions to EM and multinomial PCA. In Elomaa T, Mannila H, Toivonen H, editors, Machine Learning: ECML 2002 - 13th European Conference on Machine Learning, Proceedings. Springer-Verlag London Ltd. 2002. p. 23-34. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

Variational extensions to EM and multinomial PCA

Abstract

Publication series

Conference

Other files and links

Cite this