Minimax optimal estimation of general bandable covariance matrices

Lingzhou Xue, Hui Zou

Research output: Contribution to journalArticle

5 Citations (Scopus)

Abstract

Cai etal. (2010). [4] have studied the minimax optimal estimation of a collection of large bandable covariance matrices whose off-diagonal entries decay to zero at a polynomial rate. They have shown that the minimax optimal procedures are fundamentally different under Frobenius and spectral norms, regardless of the rate of polynomial decay. To gain more insight into this interesting problem, we study minimax estimation of large bandable covariance matrices over a parameter space characterized by a general positive decay function. We obtain explicit results to show how the decay function determines the minimax rates of convergence and the optimal procedures. From the general minimax analysis we find that for certain decay functions there is a tapering estimator that simultaneously attains the minimax optimal rates of convergence under the two norms. Moreover, we show that under the ultra-high dimension scenario it is possible to achieve adaptive minimax optimal estimation under the spectral norm. These new findings complement previous work.

Original languageEnglish (US)
Pages (from-to)45-51
Number of pages7
JournalJournal of Multivariate Analysis
Volume116
DOIs
StatePublished - Apr 1 2013

Fingerprint

Minimax Estimation
Optimal Estimation
Covariance matrix
Minimax
Decay
Spectral Norm
Polynomials
Minimax Rate
Polynomial Decay
Tapering
Frobenius norm
Optimal Rate of Convergence
Higher Dimensions
Parameter Space
Rate of Convergence
Complement
Estimator
Norm
Scenarios
Polynomial

All Science Journal Classification (ASJC) codes

  • Statistics, Probability and Uncertainty
  • Numerical Analysis
  • Statistics and Probability

Cite this

@article{b4f0e5cfb1214164806644fc497d9196,
title = "Minimax optimal estimation of general bandable covariance matrices",
abstract = "Cai etal. (2010). [4] have studied the minimax optimal estimation of a collection of large bandable covariance matrices whose off-diagonal entries decay to zero at a polynomial rate. They have shown that the minimax optimal procedures are fundamentally different under Frobenius and spectral norms, regardless of the rate of polynomial decay. To gain more insight into this interesting problem, we study minimax estimation of large bandable covariance matrices over a parameter space characterized by a general positive decay function. We obtain explicit results to show how the decay function determines the minimax rates of convergence and the optimal procedures. From the general minimax analysis we find that for certain decay functions there is a tapering estimator that simultaneously attains the minimax optimal rates of convergence under the two norms. Moreover, we show that under the ultra-high dimension scenario it is possible to achieve adaptive minimax optimal estimation under the spectral norm. These new findings complement previous work.",
author = "Lingzhou Xue and Hui Zou",
year = "2013",
month = "4",
day = "1",
doi = "10.1016/j.jmva.2012.11.003",
language = "English (US)",
volume = "116",
pages = "45--51",
journal = "Journal of Multivariate Analysis",
issn = "0047-259X",
publisher = "Academic Press Inc.",

}

Minimax optimal estimation of general bandable covariance matrices. / Xue, Lingzhou; Zou, Hui.

In: Journal of Multivariate Analysis, Vol. 116, 01.04.2013, p. 45-51.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Minimax optimal estimation of general bandable covariance matrices

AU - Xue, Lingzhou

AU - Zou, Hui

PY - 2013/4/1

Y1 - 2013/4/1

N2 - Cai etal. (2010). [4] have studied the minimax optimal estimation of a collection of large bandable covariance matrices whose off-diagonal entries decay to zero at a polynomial rate. They have shown that the minimax optimal procedures are fundamentally different under Frobenius and spectral norms, regardless of the rate of polynomial decay. To gain more insight into this interesting problem, we study minimax estimation of large bandable covariance matrices over a parameter space characterized by a general positive decay function. We obtain explicit results to show how the decay function determines the minimax rates of convergence and the optimal procedures. From the general minimax analysis we find that for certain decay functions there is a tapering estimator that simultaneously attains the minimax optimal rates of convergence under the two norms. Moreover, we show that under the ultra-high dimension scenario it is possible to achieve adaptive minimax optimal estimation under the spectral norm. These new findings complement previous work.

AB - Cai etal. (2010). [4] have studied the minimax optimal estimation of a collection of large bandable covariance matrices whose off-diagonal entries decay to zero at a polynomial rate. They have shown that the minimax optimal procedures are fundamentally different under Frobenius and spectral norms, regardless of the rate of polynomial decay. To gain more insight into this interesting problem, we study minimax estimation of large bandable covariance matrices over a parameter space characterized by a general positive decay function. We obtain explicit results to show how the decay function determines the minimax rates of convergence and the optimal procedures. From the general minimax analysis we find that for certain decay functions there is a tapering estimator that simultaneously attains the minimax optimal rates of convergence under the two norms. Moreover, we show that under the ultra-high dimension scenario it is possible to achieve adaptive minimax optimal estimation under the spectral norm. These new findings complement previous work.

UR - http://www.scopus.com/inward/record.url?scp=84871642314&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84871642314&partnerID=8YFLogxK

U2 - 10.1016/j.jmva.2012.11.003

DO - 10.1016/j.jmva.2012.11.003

M3 - Article

AN - SCOPUS:84871642314

VL - 116

SP - 45

EP - 51

JO - Journal of Multivariate Analysis

JF - Journal of Multivariate Analysis

SN - 0047-259X

ER -