Application of double asymptotics and random matrix theory in error estimation of regularized linear discriminant analysis

Amin Zollanvari, Edward R. Dougherty

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

1 Citation (Scopus)

Abstract

The theory of double asymptotics and random matrices has been employed to construct a nearly unbiased estimator of true error rate of linear discriminant analysis with ridge estimator of inverse covariance matrix in the multivariate Gaussian model. In such a scenario, the performance of the constructed estimator, as measured by Root-Mean-Square (RMS) error, shows improvement over well-known estimators of true error.

Original languageEnglish
Title of host publication2013 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2013 - Proceedings
Pages57-59
Number of pages3
DOIs
Publication statusPublished - 2013
Externally publishedYes
Event2013 1st IEEE Global Conference on Signal and Information Processing, GlobalSIP 2013 - Austin, TX, United States
Duration: Dec 3 2013Dec 5 2013

Other

Other2013 1st IEEE Global Conference on Signal and Information Processing, GlobalSIP 2013
CountryUnited States
CityAustin, TX
Period12/3/1312/5/13

Fingerprint

Discriminant analysis
Error analysis
Covariance matrix
Mean square error

Keywords

  • Double asymptotics
  • Kolmogorov asymptotics
  • Linear discriminant analysis
  • Random matrix theory
  • Small-Sample

ASJC Scopus subject areas

  • Information Systems
  • Signal Processing

Cite this

Zollanvari, A., & Dougherty, E. R. (2013). Application of double asymptotics and random matrix theory in error estimation of regularized linear discriminant analysis. In 2013 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2013 - Proceedings (pp. 57-59). [6736811] https://doi.org/10.1109/GlobalSIP.2013.6736811

Application of double asymptotics and random matrix theory in error estimation of regularized linear discriminant analysis. / Zollanvari, Amin; Dougherty, Edward R.

2013 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2013 - Proceedings. 2013. p. 57-59 6736811.

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

Zollanvari, A & Dougherty, ER 2013, Application of double asymptotics and random matrix theory in error estimation of regularized linear discriminant analysis. in 2013 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2013 - Proceedings., 6736811, pp. 57-59, 2013 1st IEEE Global Conference on Signal and Information Processing, GlobalSIP 2013, Austin, TX, United States, 12/3/13. https://doi.org/10.1109/GlobalSIP.2013.6736811
Zollanvari A, Dougherty ER. Application of double asymptotics and random matrix theory in error estimation of regularized linear discriminant analysis. In 2013 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2013 - Proceedings. 2013. p. 57-59. 6736811 https://doi.org/10.1109/GlobalSIP.2013.6736811
Zollanvari, Amin ; Dougherty, Edward R. / Application of double asymptotics and random matrix theory in error estimation of regularized linear discriminant analysis. 2013 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2013 - Proceedings. 2013. pp. 57-59
@inproceedings{ba9e27eaec2d4855a7cd812179bfc2e8,
title = "Application of double asymptotics and random matrix theory in error estimation of regularized linear discriminant analysis",
abstract = "The theory of double asymptotics and random matrices has been employed to construct a nearly unbiased estimator of true error rate of linear discriminant analysis with ridge estimator of inverse covariance matrix in the multivariate Gaussian model. In such a scenario, the performance of the constructed estimator, as measured by Root-Mean-Square (RMS) error, shows improvement over well-known estimators of true error.",
keywords = "Double asymptotics, Kolmogorov asymptotics, Linear discriminant analysis, Random matrix theory, Small-Sample",
author = "Amin Zollanvari and Dougherty, {Edward R.}",
year = "2013",
doi = "10.1109/GlobalSIP.2013.6736811",
language = "English",
isbn = "9781479902484",
pages = "57--59",
booktitle = "2013 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2013 - Proceedings",

}

TY - GEN

T1 - Application of double asymptotics and random matrix theory in error estimation of regularized linear discriminant analysis

AU - Zollanvari, Amin

AU - Dougherty, Edward R.

PY - 2013

Y1 - 2013

N2 - The theory of double asymptotics and random matrices has been employed to construct a nearly unbiased estimator of true error rate of linear discriminant analysis with ridge estimator of inverse covariance matrix in the multivariate Gaussian model. In such a scenario, the performance of the constructed estimator, as measured by Root-Mean-Square (RMS) error, shows improvement over well-known estimators of true error.

AB - The theory of double asymptotics and random matrices has been employed to construct a nearly unbiased estimator of true error rate of linear discriminant analysis with ridge estimator of inverse covariance matrix in the multivariate Gaussian model. In such a scenario, the performance of the constructed estimator, as measured by Root-Mean-Square (RMS) error, shows improvement over well-known estimators of true error.

KW - Double asymptotics

KW - Kolmogorov asymptotics

KW - Linear discriminant analysis

KW - Random matrix theory

KW - Small-Sample

UR - http://www.scopus.com/inward/record.url?scp=84897729837&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84897729837&partnerID=8YFLogxK

U2 - 10.1109/GlobalSIP.2013.6736811

DO - 10.1109/GlobalSIP.2013.6736811

M3 - Conference contribution

SN - 9781479902484

SP - 57

EP - 59

BT - 2013 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2013 - Proceedings

ER -