The Masked Sample Covariance Estimator: An Analysis via the Matrix Laplace Transform

Abstract

Covariance estimation becomes challenging in the regime where the number p of vari- ables outstrips the number n of samples available to construct the estimate. One way to circumvent this problem is to assume that the covariance matrix is nearly sparse and to focus on estimating only the signi cant entries. To analyze this approach, Levina and Vershynin (2011) introduce a formalism called masked covariance estimation, where each entry of the sample covariance estimator is reweighed to re ect an a priori assessment of its importance. This paper provides a new analysis of the masked sample covariance estimator based on the matrix Laplace transform method. The main result applies to general subgaussian distributions. Specialized to the case of a Gaussian distribution, the theory o ers qualitative improvements over earlier work. For example, the new results show that n = O(B log2 p) samples su ce to estimate a banded covariance matrix with bandwidth B up to a relative spectral-norm error, in contrast to the sample complexity n = O(B log5 p) obtained by Levina and Vershynin.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Feb 01, 2012
Accession Number
ADA563050

Entities

People

  • Alex Gittens
  • Joel A. Tropp
  • Richard Y. Chen

Organizations

  • California Institute of Technology

Tags

Communities of Interest

  • Energy and Power Technologies

DTIC Thesaurus Topics

  • Bandwidth
  • Covariance
  • Data Science
  • Estimators
  • Factor Analysis
  • Gaussian Distributions
  • Information Science
  • Mathematical Filters
  • Mathematics
  • New York
  • Normal Distribution
  • Probability
  • Random Variables
  • Statistical Algorithms
  • Statistical Analysis
  • Statistics
  • Surveys

Fields of Study

  • Mathematics

Readers

  • Adaptive Control and Estimation with Uncertainty in Dynamic Systems.
  • Linear Algebra
  • Snow Cover Descriptors for Reptiles and Their Illustrations.