Discriminant analysis

A unified approach

Peng Zhang, Jing Peng, Norbert Riedel

Research output: Chapter in Book/Report/Conference proceedingConference contributionResearchpeer-review

11 Citations (Scopus)

Abstract

Linear discriminant analysis (LDA) as a dimension reduction method is widely used in data mining and machine learning. It however suffers from the small sample size (SSS) problem when data dimensionality is greater than the sample size. Many modified methods have been proposed to address some aspect of this difficulty from a particular viewpoint. A. comprehensive framework that provides a complete solution to the SSS problem is still missing. In this paper, we provide a unified approach to LDA, and investigate the SSS problem in the framework of statistical learning theory. In such a unified approach, our analysis results in a deeper understanding of LDA. We demonstrate that LDA (and its nonlinear extension) belongs to the same framework -where powerful classifiers such as support vector machines (SVMs) are formulated. In addition, this approach allows us to establish an error bound for LDA. Finally our experiments validate our theoretical analysis results.

Original languageEnglish
Title of host publicationProceedings - Fifth IEEE International Conference on Data Mining, ICDM 2005
Pages514-521
Number of pages8
DOIs
StatePublished - 1 Dec 2005
Event5th IEEE International Conference on Data Mining, ICDM 2005 - Houston, TX, United States
Duration: 27 Nov 200530 Nov 2005

Other

Other5th IEEE International Conference on Data Mining, ICDM 2005
CountryUnited States
CityHouston, TX
Period27/11/0530/11/05

Fingerprint

Discriminant analysis
Support vector machines
Data mining
Learning systems
Classifiers
Experiments

Cite this

Zhang, P., Peng, J., & Riedel, N. (2005). Discriminant analysis: A unified approach. In Proceedings - Fifth IEEE International Conference on Data Mining, ICDM 2005 (pp. 514-521). [1565719] https://doi.org/10.1109/ICDM.2005.51
Zhang, Peng ; Peng, Jing ; Riedel, Norbert. / Discriminant analysis : A unified approach. Proceedings - Fifth IEEE International Conference on Data Mining, ICDM 2005. 2005. pp. 514-521
@inproceedings{9a1d6074c230406fbc98a2592a307dbe,
title = "Discriminant analysis: A unified approach",
abstract = "Linear discriminant analysis (LDA) as a dimension reduction method is widely used in data mining and machine learning. It however suffers from the small sample size (SSS) problem when data dimensionality is greater than the sample size. Many modified methods have been proposed to address some aspect of this difficulty from a particular viewpoint. A. comprehensive framework that provides a complete solution to the SSS problem is still missing. In this paper, we provide a unified approach to LDA, and investigate the SSS problem in the framework of statistical learning theory. In such a unified approach, our analysis results in a deeper understanding of LDA. We demonstrate that LDA (and its nonlinear extension) belongs to the same framework -where powerful classifiers such as support vector machines (SVMs) are formulated. In addition, this approach allows us to establish an error bound for LDA. Finally our experiments validate our theoretical analysis results.",
author = "Peng Zhang and Jing Peng and Norbert Riedel",
year = "2005",
month = "12",
day = "1",
doi = "10.1109/ICDM.2005.51",
language = "English",
isbn = "0769522785",
pages = "514--521",
booktitle = "Proceedings - Fifth IEEE International Conference on Data Mining, ICDM 2005",

}

Zhang, P, Peng, J & Riedel, N 2005, Discriminant analysis: A unified approach. in Proceedings - Fifth IEEE International Conference on Data Mining, ICDM 2005., 1565719, pp. 514-521, 5th IEEE International Conference on Data Mining, ICDM 2005, Houston, TX, United States, 27/11/05. https://doi.org/10.1109/ICDM.2005.51

Discriminant analysis : A unified approach. / Zhang, Peng; Peng, Jing; Riedel, Norbert.

Proceedings - Fifth IEEE International Conference on Data Mining, ICDM 2005. 2005. p. 514-521 1565719.

Research output: Chapter in Book/Report/Conference proceedingConference contributionResearchpeer-review

TY - GEN

T1 - Discriminant analysis

T2 - A unified approach

AU - Zhang, Peng

AU - Peng, Jing

AU - Riedel, Norbert

PY - 2005/12/1

Y1 - 2005/12/1

N2 - Linear discriminant analysis (LDA) as a dimension reduction method is widely used in data mining and machine learning. It however suffers from the small sample size (SSS) problem when data dimensionality is greater than the sample size. Many modified methods have been proposed to address some aspect of this difficulty from a particular viewpoint. A. comprehensive framework that provides a complete solution to the SSS problem is still missing. In this paper, we provide a unified approach to LDA, and investigate the SSS problem in the framework of statistical learning theory. In such a unified approach, our analysis results in a deeper understanding of LDA. We demonstrate that LDA (and its nonlinear extension) belongs to the same framework -where powerful classifiers such as support vector machines (SVMs) are formulated. In addition, this approach allows us to establish an error bound for LDA. Finally our experiments validate our theoretical analysis results.

AB - Linear discriminant analysis (LDA) as a dimension reduction method is widely used in data mining and machine learning. It however suffers from the small sample size (SSS) problem when data dimensionality is greater than the sample size. Many modified methods have been proposed to address some aspect of this difficulty from a particular viewpoint. A. comprehensive framework that provides a complete solution to the SSS problem is still missing. In this paper, we provide a unified approach to LDA, and investigate the SSS problem in the framework of statistical learning theory. In such a unified approach, our analysis results in a deeper understanding of LDA. We demonstrate that LDA (and its nonlinear extension) belongs to the same framework -where powerful classifiers such as support vector machines (SVMs) are formulated. In addition, this approach allows us to establish an error bound for LDA. Finally our experiments validate our theoretical analysis results.

UR - http://www.scopus.com/inward/record.url?scp=34548588868&partnerID=8YFLogxK

U2 - 10.1109/ICDM.2005.51

DO - 10.1109/ICDM.2005.51

M3 - Conference contribution

SN - 0769522785

SN - 9780769522784

SP - 514

EP - 521

BT - Proceedings - Fifth IEEE International Conference on Data Mining, ICDM 2005

ER -

Zhang P, Peng J, Riedel N. Discriminant analysis: A unified approach. In Proceedings - Fifth IEEE International Conference on Data Mining, ICDM 2005. 2005. p. 514-521. 1565719 https://doi.org/10.1109/ICDM.2005.51