Robust estimators for nondecomposable elliptical graphical models

D. Vogel*, D. E. Tyler

*Corresponding author for this work

Research output: Contribution to journalArticle

3 Citations (Scopus)

Abstract

Robust estimators of the restricted covariance matrices associated with elliptical graphical models are studied. General asymptotic results, which apply to both decomposable and nondecomposable graphical models, are presented for robust plug-in type estimators. These extend results previously established only for the decomposable case. Furthermore, a class of graphical M-estimators for the restricted covariance matrices is introduced and compared with the corresponding plug-in M-estimators. The two approaches are shown to be asymptotically equivalent under random sampling from an elliptical distribution. A simulation study demonstrates the superiority of the graphical M-estimators for small samples.

Original languageEnglish
Pages (from-to)865-882
Number of pages18
JournalBiometrika
Volume101
Issue number4
Early online date17 Oct 2014
DOIs
Publication statusPublished - Dec 2014

Keywords

  • affine equivariance
  • deviance test
  • Gaussian graphical model
  • M-estimator
  • partial correlation
  • Maximum-likelihood-estimation
  • maultivariate time-series
  • covariance-matrix
  • asymptotic-behavior
  • Gaussian Models
  • scatter
  • selection
  • location
  • distributions
  • efficiency

Cite this

Robust estimators for nondecomposable elliptical graphical models. / Vogel, D.; Tyler, D. E.

In: Biometrika, Vol. 101, No. 4, 12.2014, p. 865-882.

Research output: Contribution to journalArticle

Vogel, D. ; Tyler, D. E. / Robust estimators for nondecomposable elliptical graphical models. In: Biometrika. 2014 ; Vol. 101, No. 4. pp. 865-882.
@article{4f43fcd16ba54c8abd83ae071115737b,
title = "Robust estimators for nondecomposable elliptical graphical models",
abstract = "Robust estimators of the restricted covariance matrices associated with elliptical graphical models are studied. General asymptotic results, which apply to both decomposable and nondecomposable graphical models, are presented for robust plug-in type estimators. These extend results previously established only for the decomposable case. Furthermore, a class of graphical M-estimators for the restricted covariance matrices is introduced and compared with the corresponding plug-in M-estimators. The two approaches are shown to be asymptotically equivalent under random sampling from an elliptical distribution. A simulation study demonstrates the superiority of the graphical M-estimators for small samples.",
keywords = "affine equivariance, deviance test, Gaussian graphical model, M-estimator, partial correlation, Maximum-likelihood-estimation, maultivariate time-series, covariance-matrix, asymptotic-behavior, Gaussian Models, scatter, selection, location, distributions, efficiency",
author = "D. Vogel and Tyler, {D. E.}",
year = "2014",
month = "12",
doi = "10.1093/biomet/asu041",
language = "English",
volume = "101",
pages = "865--882",
journal = "Biometrika",
issn = "0006-3444",
publisher = "Oxford University Press",
number = "4",

}

TY - JOUR

T1 - Robust estimators for nondecomposable elliptical graphical models

AU - Vogel, D.

AU - Tyler, D. E.

PY - 2014/12

Y1 - 2014/12

N2 - Robust estimators of the restricted covariance matrices associated with elliptical graphical models are studied. General asymptotic results, which apply to both decomposable and nondecomposable graphical models, are presented for robust plug-in type estimators. These extend results previously established only for the decomposable case. Furthermore, a class of graphical M-estimators for the restricted covariance matrices is introduced and compared with the corresponding plug-in M-estimators. The two approaches are shown to be asymptotically equivalent under random sampling from an elliptical distribution. A simulation study demonstrates the superiority of the graphical M-estimators for small samples.

AB - Robust estimators of the restricted covariance matrices associated with elliptical graphical models are studied. General asymptotic results, which apply to both decomposable and nondecomposable graphical models, are presented for robust plug-in type estimators. These extend results previously established only for the decomposable case. Furthermore, a class of graphical M-estimators for the restricted covariance matrices is introduced and compared with the corresponding plug-in M-estimators. The two approaches are shown to be asymptotically equivalent under random sampling from an elliptical distribution. A simulation study demonstrates the superiority of the graphical M-estimators for small samples.

KW - affine equivariance

KW - deviance test

KW - Gaussian graphical model

KW - M-estimator

KW - partial correlation

KW - Maximum-likelihood-estimation

KW - maultivariate time-series

KW - covariance-matrix

KW - asymptotic-behavior

KW - Gaussian Models

KW - scatter

KW - selection

KW - location

KW - distributions

KW - efficiency

U2 - 10.1093/biomet/asu041

DO - 10.1093/biomet/asu041

M3 - Article

VL - 101

SP - 865

EP - 882

JO - Biometrika

JF - Biometrika

SN - 0006-3444

IS - 4

ER -