Publication Type

Journal Article

Version

submittedVersion

Publication Date

11-2013

Abstract

This article shows that when the nonzero coefficients of the population correlation matrix are all greater in absolute value than (C1logp/n)1/2 for some constant C1, we can obtain covariance selection consistency by thresholding the sample correlation matrix. Furthermore, the rate (logp/n)1/2 is shown to be optimal.

Keywords

Bernstein type inequality, Covariance selection, Large correlation matrix, Large covariance matrix, Thresholding

Discipline

Databases and Information Systems | Numerical Analysis and Scientific Computing

Publication

Statistics and Probability Letters

Volume

83

Issue

11

First Page

2492

Last Page

2498

ISSN

0167-7152

Identifier

10.1016/j.spl.2013.07.008

Publisher

Elsevier

Embargo Period

2-20-2014

Additional URL

http://dx.doi.org/10.1016/j.spl.2013.07.008

Share

COinS