Estimation of a Covariance Matrix with Zeros

Abstract

We consider estimation of the covariance matrix of a multivariate random vector under the constraint that certain covariances are zero. We first present an algorithm, which we call Iterative Conditional Fitting, for computing the maximum likelihood estimator of the constrained covariance matrix, under the assumption of multivariate normality. In contrast to previous approaches, this algorithm has guaranteed convergence properties. Dropping the assumption of multivariate normality, we show how to estimate the covariance matrix in an empirical likelihood approach. These approaches are then compared via simulation and on an example of gene expression.

25 pages

References in corpus (2)

• Statistical applications of the multivariate skew-normal distribution
• Maximum Likelihood Estimation in Gaussian Chain Graph Models under the Alternative Markov Property

Cited by in corpus (27)

• Sparse permutation invariant covariance estimation
• Multiple Testing and Error Control in Gaussian Graphical Model Selection
• High-dimensional covariance estimation based on Gaussian graphical models
• Binary Models for Marginal Independence
• Scaling It Up: Stochastic Search Structure Learning in Graphical Models
• Wishart distributions for decomposable covariance graph models
• Concave Penalized Estimation of Sparse Gaussian Bayesian Networks
• Computing maximum likelihood estimates in recursive linear models with correlated errors
• Bayesian sparse graphical models for classification with application to protein expression data
• Estimating linear covariance models with numerical nonlinear algebra
• High dimensional sparse covariance estimation via directed acyclic graphs
• Learning to Estimate Without Bias
• Perturbations and projections of Kalman-Bucy semigroups
• Determining full conditional independence by low-order conditioning
• Maximum likelihood degree of the two-dimensional linear Gaussian covariance model
• Unobserved classes and extra variables in high-dimensional discriminant analysis
• On Testing Marginal versus Conditional Independence
• A random covariance model for bi-level graphical modeling with application to resting-state fMRI data
• Sparse Cholesky covariance parametrization for recovering latent structure in ordered data
• Detecting the large entries of a sparse covariance matrix in sub-quadratic time
• Introducing Gaussian covariance graph models in genome-wide prediction
• MLE of Jointly Constrained Mean-Covariance of Multivariate Normal Distributions
• A new method for the estimation of variance matrix with prescribed zeros in nonlinear mixed effects models
• Graphical methods for efficient likelihood inference in Gaussian covariance models
• Clique Matrices for Statistical Graph Decomposition and Parameterising Restricted Positive Definite Matrices
• A Proximal Distance Algorithm for Likelihood-Based Sparse Covariance Estimation
• A new approach to Cholesky-based covariance regularization in high dimensions