Asymptotic normality and optimalities in estimation of large Gaussian graphical models

Abstract

The Gaussian graphical model, a popular paradigm for studying relationship among variables in a wide range of applications, has attracted great attention in recent years. This paper considers a fundamental question: When is it possible to estimate low-dimensional parameters at parametric square-root rate in a large Gaussian graphical model? A novel regression approach is proposed to obtain asymptotically efficient estimation of each entry of a precision matrix under a sparseness condition relative to the sample size. When the precision matrix is not sufficiently sparse, or equivalently the sample size is not sufficiently large, a lower bound is established to show that it is no longer possible to achieve the parametric rate in the estimation of each entry. This lower bound result, which provides an answer to the delicate sample size question, is established with a novel construction of a subset of sparse precision matrices in an application of Le Cam's lemma. Moreover, the proposed estimator is proven to have optimal convergence rate when the parametric rate cannot be achieved, under a minimal sample requirement. The proposed estimator is applied to test the presence of an edge in the Gaussian graphical model or to recover the support of the entire model, to obtain adaptive rate-optimal estimation of the entire precision matrix as measured by the matrix ${\ell }_q$ operator norm and to make inference in latent variables in the graphical model. All of this is achieved under a sparsity condition on the precision matrix and a side condition on the range of its spectrum. This significantly relaxes the commonly imposed uniform signal strength condition on the precision matrix, irrepresentability condition on the Hessian tensor operator of the covariance matrix or the ${\ell }_1$ constraint on the precision matrix. Numerical results confirm our theoretical findings. The ROC curve of the proposed algorithm, Asymptotic Normal Thresholding (ANT), for support recovery significantly outperforms that of the popular GLasso algorithm.

Published at http://dx.doi.org/10.1214/14-AOS1286 in the Annals of Statistics (http://www.imstat.org/aos/) by the Institute of Mathematical Statistics (http://www.imstat.org)

References in corpus (13)

• Regularized estimation of large covariance matrices
• Covariance regularization by thresholding
• Sparse permutation invariant covariance estimation
• The sparsity and bias of the Lasso selection in high-dimensional linear regression
• Optimal rates of convergence for covariance matrix estimation
• Operator norm consistent estimation of large-dimensional sparse covariance matrices
• Optimal rates of convergence for sparse covariance matrix estimation
• The Dantzig selector and sparsity oracle inequalities
• Some sharp performance bounds for least squares regression with $L_1$ regularization
• L0 Sparse Inverse Covariance Estimation
• Discussion: Latent variable graphical model selection via convex optimization
• Gaussian Graphical Model Estimation with False Discovery Rate Control
• Estimating Sparse Precision Matrix: Optimal Rates of Convergence and Adaptive Estimation

Cited by in corpus (57)

• Confidence Intervals and Hypothesis Testing for High-Dimensional Regression
• The Hardness of Conditional Independence Testing and the Generalised Covariance Measure
• Inference and Uncertainty Quantification for Noisy Matrix Completion
• Global and Simultaneous Hypothesis Testing for High-Dimensional Logistic Regression Models
• High-dimensional inference in misspecified linear models
• Goodness of fit tests for high-dimensional linear models
• Confidence regions for entries of a large precision matrix
• Sparse CCA via Precision Adjusted Iterative Thresholding
• Estimating sparse precision matrices
• DRAGON: Determining Regulatory Associations using Graphical models on multi-Omic Networks
• Local and Global Inference for High Dimensional Nonparanormal Graphical Models
• On Semiparametric Exponential Family Graphical Models
• Two Sample Inference for Populations of Graphical Models with Applications to Functional Connectivity
• Post-Regularization Inference for Time-Varying Nonparanormal Graphical Models
• Estimating Differential Latent Variable Graphical Models with Applications to Brain Connectivity
• Confidence Intervals for Causal Effects with Invalid Instruments using Two-Stage Hard Thresholding with Voting
• Inference for Sparse Conditional Precision Matrices
• Robust Regression via Mutivariate Regression Depth
• Constrained High Dimensional Statistical Inference
• Online Structural Change-point Detection of High-dimensional Streaming Data via Dynamic Sparse Subspace Learning
• A Projection Based Conditional Dependence Measure with Applications to High-dimensional Undirected Graphical Models
• De-biased graphical Lasso for high-frequency data
• Inference for Heteroskedastic PCA with Missing Data
• Confidence Intervals for High-Dimensional Linear Regression: Minimax Rates and Adaptivity
• General framework for projection structures
• High-Temperature Structure Detection in Ferromagnets
• Adaptive Inferential Method for Monotone Graph Invariants
• RSVP-graphs: Fast High-dimensional Covariance Matrix Estimation under Latent Confounding
• Tensor Graphical Model: Non-convex Optimization and Statistical Inference
• On the efficiency of the de-biased Lasso
• The Benefit of Group Sparsity in Group Inference with De-biased Scaled Group Lasso
• Optimal Estimation of Co-heritability in High-dimensional Linear Models
• Graph Estimation for Matrix-variate Gaussian Data
• Tackling small eigen-gaps: Fine-grained eigenvector estimation and inference under heteroscedastic noise
• Berry-Esseen Bounds for Projection Parameters and Partial Correlations with Increasing Dimension
• Minimax Posterior Convergence Rates and Model Selection Consistency in High-dimensional DAG Models based on Sparse Cholesky Factors
• Semiparametric Gaussian copula classification
• Innovated scalable efficient estimation in ultra-large Gaussian graphical models
• A Stepwise Approach for High-Dimensional Gaussian Graphical Models
• Multi-level Thresholding Test for High Dimensional Covariance Matrices
• Tuning-Free Heterogeneity Pursuit in Massive Networks
• Joint Nonparametric Precision Matrix Estimation with Confounding
• Minimax Estimation of Bandable Precision Matrices
• A unified precision matrix estimation framework via sparse column-wise inverse operator under weak sparsity
• Optimal statistical decision for Gaussian graphical model selection
• Minimax Estimation of Large Precision Matrices with Bandable Cholesky Factor
• Bayesian Bandwidth Test and Selection for High-dimensional Banded Precision Matrices
• Estimation of Graphical Models using the $L_{1,2}$ Norm
• Exponential Family Graphical Models: Correlated Replicates and Unmeasured Confounders, with Applications to fMRI Data
• Hub Detection in Gaussian Graphical Models
• StarTrek: Combinatorial Variable Selection with False Discovery Rate Control
• Bayesian inference for high-dimensional decomposable graphs
• Gaussian Graphical Model exploration and selection in high dimension low sample size setting
• High-dimensional Gaussian graphical model for network-linked data
• High-dimensional Two-sample Precision Matrices Test: An Adaptive Approach through Multiplier Bootstrap
• Graphical Fermat's Principle and Triangle-Free Graph Estimation
• Testing Endogeneity with High Dimensional Covariates