Improving principal component analysis using Bayesian estimation

M. N. Nounou; B. R. Bakshi; P. K. Goel; X. Shen

doi:10.1109/acc.2001.946204

Improving principal component analysis using Bayesian estimation

M. N. Nounou^*
, B. R. Bakshi
, P. K. Goel
, X. Shen

^*Corresponding author for this work

Ohio State University

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

Abstract

Bayesian estimation is used in this paper to derive a new PCA modeling algorithm that improves the estimation accuracy by incorporating prior knowledge about the data and model. It is shown that the algorithm is more general than existing methods, PCA and MLPCA, and reduces to these techniques when a uniform prior is used. It is also shown that when no external information is available, an empirically estimated prior from the available data can still provide improved accuracy over non-Bayesian methods.

Original language	English
Title of host publication	Proceedings of the 2001 American Control Conference, ACC 2001
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	3666-3671
Number of pages	6
ISBN (Print)	0780364953
DOIs	https://doi.org/10.1109/acc.2001.946204
Publication status	Published - 2001
Externally published	Yes
Event	2001 American Control Conference, ACC 2001 - Arlington, VA, United States Duration: 25 Jun 2001 → 27 Jun 2001

Publication series

Name	Proceedings of the American Control Conference
Volume	5
ISSN (Print)	0743-1619

Conference

Conference	2001 American Control Conference, ACC 2001
Country/Territory	United States
City	Arlington, VA
Period	25/06/01 → 27/06/01

Access to Document

10.1109/acc.2001.946204

Cite this

@inproceedings{22c231dc9a8b434abe0742d9299b4c4e,

title = "Improving principal component analysis using Bayesian estimation",

abstract = "Bayesian estimation is used in this paper to derive a new PCA modeling algorithm that improves the estimation accuracy by incorporating prior knowledge about the data and model. It is shown that the algorithm is more general than existing methods, PCA and MLPCA, and reduces to these techniques when a uniform prior is used. It is also shown that when no external information is available, an empirically estimated prior from the available data can still provide improved accuracy over non-Bayesian methods.",

author = "Nounou, \{M. N.\} and Bakshi, \{B. R.\} and Goel, \{P. K.\} and X. Shen",

year = "2001",

doi = "10.1109/acc.2001.946204",

language = "English",

isbn = "0780364953",

series = "Proceedings of the American Control Conference",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

pages = "3666--3671",

booktitle = "Proceedings of the 2001 American Control Conference, ACC 2001",

address = "United States",

note = "2001 American Control Conference, ACC 2001 ; Conference date: 25-06-2001 Through 27-06-2001",

}

Nounou, MN, Bakshi, BR, Goel, PK & Shen, X 2001, Improving principal component analysis using Bayesian estimation. in Proceedings of the 2001 American Control Conference, ACC 2001. Proceedings of the American Control Conference, vol. 5, Institute of Electrical and Electronics Engineers Inc., pp. 3666-3671, 2001 American Control Conference, ACC 2001, Arlington, VA, United States, 25/06/01. https://doi.org/10.1109/acc.2001.946204

Improving principal component analysis using Bayesian estimation. / Nounou, M. N.; Bakshi, B. R.; Goel, P. K. et al.
Proceedings of the 2001 American Control Conference, ACC 2001. Institute of Electrical and Electronics Engineers Inc., 2001. p. 3666-3671 (Proceedings of the American Control Conference; Vol. 5).

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

TY - GEN

T1 - Improving principal component analysis using Bayesian estimation

AU - Nounou, M. N.

AU - Bakshi, B. R.

AU - Goel, P. K.

AU - Shen, X.

PY - 2001

Y1 - 2001

N2 - Bayesian estimation is used in this paper to derive a new PCA modeling algorithm that improves the estimation accuracy by incorporating prior knowledge about the data and model. It is shown that the algorithm is more general than existing methods, PCA and MLPCA, and reduces to these techniques when a uniform prior is used. It is also shown that when no external information is available, an empirically estimated prior from the available data can still provide improved accuracy over non-Bayesian methods.

AB - Bayesian estimation is used in this paper to derive a new PCA modeling algorithm that improves the estimation accuracy by incorporating prior knowledge about the data and model. It is shown that the algorithm is more general than existing methods, PCA and MLPCA, and reduces to these techniques when a uniform prior is used. It is also shown that when no external information is available, an empirically estimated prior from the available data can still provide improved accuracy over non-Bayesian methods.

UR - https://www.scopus.com/pages/publications/0034848185

U2 - 10.1109/acc.2001.946204

DO - 10.1109/acc.2001.946204

M3 - Conference contribution

AN - SCOPUS:0034848185

SN - 0780364953

T3 - Proceedings of the American Control Conference

SP - 3666

EP - 3671

BT - Proceedings of the 2001 American Control Conference, ACC 2001

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 2001 American Control Conference, ACC 2001

Y2 - 25 June 2001 through 27 June 2001

ER -

Improving principal component analysis using Bayesian estimation

Abstract

Publication series

Conference

Access to Document

Other files and links

Fingerprint

Cite this