Covariance selection by thresholding the sample correlation matrix

Binyan Jiang

doi:10.1016/j.spl.2013.07.008

Covariance selection by thresholding the sample correlation matrix

Binyan Jiang

Research output: Journal article publication › Journal article › Academic research › peer-review

10 Citations (Scopus)

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 constantC1, we can obtain covariance selection consistency by thresholding the sample correlation matrix. Furthermore, the rate( log p / n) 1 / 2is shown to be optimal.

Original language	English
Pages (from-to)	2492-2498
Number of pages	7
Journal	Statistics and Probability Letters
Volume	83
Issue number	11
DOIs	https://doi.org/10.1016/j.spl.2013.07.008
Publication status	Published - 1 Nov 2013
Externally published	Yes

Keywords

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

ASJC Scopus subject areas

Statistics and Probability
Statistics, Probability and Uncertainty

More information

10.1016/j.spl.2013.07.008

Cite this

@article{ae6f16f212164b4f92943aa4f8366d11,

title = "Covariance selection by thresholding the sample correlation matrix",

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 constantC1, we can obtain covariance selection consistency by thresholding the sample correlation matrix. Furthermore, the rate( log p / n) 1 / 2is shown to be optimal.",

keywords = "Bernstein type inequality, Covariance selection, Large correlation matrix, Large covariance matrix, Thresholding",

author = "Binyan Jiang",

year = "2013",

month = nov,

day = "1",

doi = "10.1016/j.spl.2013.07.008",

language = "English",

volume = "83",

pages = "2492--2498",

journal = "Statistics and Probability Letters",

issn = "0167-7152",

publisher = "Elsevier",

number = "11",

}

TY - JOUR

T1 - Covariance selection by thresholding the sample correlation matrix

AU - Jiang, Binyan

PY - 2013/11/1

Y1 - 2013/11/1

N2 - 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 constantC1, we can obtain covariance selection consistency by thresholding the sample correlation matrix. Furthermore, the rate( log p / n) 1 / 2is shown to be optimal.

AB - 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 constantC1, we can obtain covariance selection consistency by thresholding the sample correlation matrix. Furthermore, the rate( log p / n) 1 / 2is shown to be optimal.

KW - Bernstein type inequality

KW - Covariance selection

KW - Large correlation matrix

KW - Large covariance matrix

KW - Thresholding

UR - http://www.scopus.com/inward/record.url?scp=84882764055&partnerID=8YFLogxK

U2 - 10.1016/j.spl.2013.07.008

DO - 10.1016/j.spl.2013.07.008

M3 - Journal article

SN - 0167-7152

VL - 83

SP - 2492

EP - 2498

JO - Statistics and Probability Letters

JF - Statistics and Probability Letters

IS - 11

ER -

Covariance selection by thresholding the sample correlation matrix

Abstract

Keywords

ASJC Scopus subject areas

More information

Other files and links

Fingerprint

Cite this