Feature extraction via generalized uncorrelated linear discriminant analysis

Jieping Ye, Ravi Janardan, Qi Li, Haesun Park

Research output: Chapter in Book/Report/Conference proceedingConference contribution

30 Citations (Scopus)

Abstract

Feature extraction is important in many applications, such as text and image retrieval, because of high dimensionality. Uncorrelated Linear Discriminant Analysis (ULDA) was recently proposed for feature extraction. The extracted features via ULDA were shown to be statistically uncorrelated, which is desirable for many applications. In this paper, we will first propose the ULDA/QR algorithm to simplify the previous implementation of ULDA. Then we propose the ULDA/GSVD algorithm, based on a novel optimization criterion, to address the singularity problem. It is applicable for undersampled problem, where the data dimension is much larger than the data size, such as text and image retrieval. The novel criterion used in ULDA/GSVD is the perturbed version of the one from ULDA/QR, while surprisingly, the solution to ULDA/GSVD is shown to be independent of the amount of perturbation applied. We did extensive experiments on text and face image data to show the effectiveness of ULDA/GSVD and compare with other popular feature extraction algorithms.

Original languageEnglish (US)
Title of host publicationProceedings, Twenty-First International Conference on Machine Learning, ICML 2004
EditorsR. Greiner, D. Schuurmans
Pages895-902
Number of pages8
StatePublished - 2004
Externally publishedYes
EventProceedings, Twenty-First International Conference on Machine Learning, ICML 2004 - Banff, Alta, Canada
Duration: Jul 4 2004Jul 8 2004

Other

OtherProceedings, Twenty-First International Conference on Machine Learning, ICML 2004
CountryCanada
CityBanff, Alta
Period7/4/047/8/04

Fingerprint

Discriminant analysis
Feature extraction
Image retrieval

ASJC Scopus subject areas

  • Engineering(all)

Cite this

Ye, J., Janardan, R., Li, Q., & Park, H. (2004). Feature extraction via generalized uncorrelated linear discriminant analysis. In R. Greiner, & D. Schuurmans (Eds.), Proceedings, Twenty-First International Conference on Machine Learning, ICML 2004 (pp. 895-902)

Feature extraction via generalized uncorrelated linear discriminant analysis. / Ye, Jieping; Janardan, Ravi; Li, Qi; Park, Haesun.

Proceedings, Twenty-First International Conference on Machine Learning, ICML 2004. ed. / R. Greiner; D. Schuurmans. 2004. p. 895-902.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Ye, J, Janardan, R, Li, Q & Park, H 2004, Feature extraction via generalized uncorrelated linear discriminant analysis. in R Greiner & D Schuurmans (eds), Proceedings, Twenty-First International Conference on Machine Learning, ICML 2004. pp. 895-902, Proceedings, Twenty-First International Conference on Machine Learning, ICML 2004, Banff, Alta, Canada, 7/4/04.
Ye J, Janardan R, Li Q, Park H. Feature extraction via generalized uncorrelated linear discriminant analysis. In Greiner R, Schuurmans D, editors, Proceedings, Twenty-First International Conference on Machine Learning, ICML 2004. 2004. p. 895-902
Ye, Jieping ; Janardan, Ravi ; Li, Qi ; Park, Haesun. / Feature extraction via generalized uncorrelated linear discriminant analysis. Proceedings, Twenty-First International Conference on Machine Learning, ICML 2004. editor / R. Greiner ; D. Schuurmans. 2004. pp. 895-902
@inproceedings{9c004c14faec430b9fd4c38afaa92a00,
title = "Feature extraction via generalized uncorrelated linear discriminant analysis",
abstract = "Feature extraction is important in many applications, such as text and image retrieval, because of high dimensionality. Uncorrelated Linear Discriminant Analysis (ULDA) was recently proposed for feature extraction. The extracted features via ULDA were shown to be statistically uncorrelated, which is desirable for many applications. In this paper, we will first propose the ULDA/QR algorithm to simplify the previous implementation of ULDA. Then we propose the ULDA/GSVD algorithm, based on a novel optimization criterion, to address the singularity problem. It is applicable for undersampled problem, where the data dimension is much larger than the data size, such as text and image retrieval. The novel criterion used in ULDA/GSVD is the perturbed version of the one from ULDA/QR, while surprisingly, the solution to ULDA/GSVD is shown to be independent of the amount of perturbation applied. We did extensive experiments on text and face image data to show the effectiveness of ULDA/GSVD and compare with other popular feature extraction algorithms.",
author = "Jieping Ye and Ravi Janardan and Qi Li and Haesun Park",
year = "2004",
language = "English (US)",
isbn = "1581138385",
pages = "895--902",
editor = "R. Greiner and D. Schuurmans",
booktitle = "Proceedings, Twenty-First International Conference on Machine Learning, ICML 2004",

}

TY - GEN

T1 - Feature extraction via generalized uncorrelated linear discriminant analysis

AU - Ye, Jieping

AU - Janardan, Ravi

AU - Li, Qi

AU - Park, Haesun

PY - 2004

Y1 - 2004

N2 - Feature extraction is important in many applications, such as text and image retrieval, because of high dimensionality. Uncorrelated Linear Discriminant Analysis (ULDA) was recently proposed for feature extraction. The extracted features via ULDA were shown to be statistically uncorrelated, which is desirable for many applications. In this paper, we will first propose the ULDA/QR algorithm to simplify the previous implementation of ULDA. Then we propose the ULDA/GSVD algorithm, based on a novel optimization criterion, to address the singularity problem. It is applicable for undersampled problem, where the data dimension is much larger than the data size, such as text and image retrieval. The novel criterion used in ULDA/GSVD is the perturbed version of the one from ULDA/QR, while surprisingly, the solution to ULDA/GSVD is shown to be independent of the amount of perturbation applied. We did extensive experiments on text and face image data to show the effectiveness of ULDA/GSVD and compare with other popular feature extraction algorithms.

AB - Feature extraction is important in many applications, such as text and image retrieval, because of high dimensionality. Uncorrelated Linear Discriminant Analysis (ULDA) was recently proposed for feature extraction. The extracted features via ULDA were shown to be statistically uncorrelated, which is desirable for many applications. In this paper, we will first propose the ULDA/QR algorithm to simplify the previous implementation of ULDA. Then we propose the ULDA/GSVD algorithm, based on a novel optimization criterion, to address the singularity problem. It is applicable for undersampled problem, where the data dimension is much larger than the data size, such as text and image retrieval. The novel criterion used in ULDA/GSVD is the perturbed version of the one from ULDA/QR, while surprisingly, the solution to ULDA/GSVD is shown to be independent of the amount of perturbation applied. We did extensive experiments on text and face image data to show the effectiveness of ULDA/GSVD and compare with other popular feature extraction algorithms.

UR - http://www.scopus.com/inward/record.url?scp=14344266778&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=14344266778&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:14344266778

SN - 1581138385

SN - 9781581138382

SP - 895

EP - 902

BT - Proceedings, Twenty-First International Conference on Machine Learning, ICML 2004

A2 - Greiner, R.

A2 - Schuurmans, D.

ER -