A regularized profile likelihood approach to covariance matrix estimation

被引：3

作者：

Banerjee, Samprit ^{[1
]}

Monni, Stefano ^{[2
]}

Wells, Martin T. ^{[1
]}

机构：

[1] Cornell Univ, Ithaca, NY 14853 USA

[2] Amer Univ Beirut, Beirut, Lebanon

来源：

JOURNAL OF STATISTICAL PLANNING AND INFERENCE | 2016年 / 179卷

基金：

美国国家科学基金会;

关键词：

Covariance matrix; Eigenvalues; Isotonic regression; Laplace approximation; Modified profile likelihood; LATENT ROOTS;

D O I：

10.1016/j.jspi.2016.06.004

中图分类号：

O21 [概率论与数理统计]; C8 [统计学];

学科分类号：

020208 ; 070103 ; 0714 ;

摘要：

Two new orthogonally equivariant estimators of the covariance matrix are proposed. The estimates of the population eigenvalues are isotonized maximum likelihood estimates of the modified profile likelihood obtained from the Wishart distribution, in one case, and of a penalized form of such a likelihood function, in the other, with a penalty that constrains the trace of the sample covariance matrix. Properties of these estimators are studied and numerical risk comparisons with six other well-known estimators are presented to demonstrate the robustness of the proposed estimators for various real and simulated covariance structures. (C) 2016 Elsevier B.V. All rights reserved.

引用

页码：36 / 59

页数：24

共 35 条

[1] AN ASYMPTOTIC-EXPANSION FOR THE DISTRIBUTION OF THE LATENT ROOTS OF THE ESTIMATED COVARIANCE-MATRIX [J].