Frequentist coverage of adaptive nonparametric Bayesian credible sets

Abstract

We investigate the frequentist coverage of Bayesian credible sets in a nonparametric setting. We consider a scale of priors of varying regularity and choose the regularity by an empirical Bayes method. Next we consider a central set of prescribed posterior probability in the posterior distribution of the chosen regularity. We show that such an adaptive Bayes credible set gives correct uncertainty quantification of "polished tail" parameters, in the sense of high probability of coverage of such parameters. On the negative side, we show by theory and example that adaptation of the prior necessarily leads to gross and haphazard uncertainty quantification for some true parameters that are still within the hyperrectangle regularity scale.

Published at http://dx.doi.org/10.1214/14-AOS1270 in the Annals of Statistics (http://www.imstat.org/aos/) by the Institute of Mathematical Statistics (http://www.imstat.org)

References in corpus (4)

• Confidence bands in density estimation
• On adaptive inference and confidence bands
• Adaptive nonparametric confidence sets
• Adaptive confidence balls

Cited by in corpus (44)

• On the Bernstein-von Mises phenomenon for nonparametric Bayes procedures
• Contributed Discussion to Uncertainty Quantification for the Horseshoe by Stéphanie van der Pas, Botond Szabó and Aad van der Vaart
• Nonparametric Bayesian posterior contraction rates for discretely observed scalar diffusions
• Adaptive Bernstein-von Mises theorems in Gaussian white noise
• Supremum Norm Posterior Contraction and Credible Sets for Nonparametric Multivariate Regression
• Frequentist coverage and sup-norm convergence rate in Gaussian process regression
• Bayesian Quantile Regression Using Random B-spline Series Prior
• An asymptotic analysis of distributed nonparametric methods
• Posterior consistency and convergence rates for Bayesian inversion with hypoelliptic operators
• A hierarchical Bayesian perspective on majorization-minimization for non-convex sparse regression: application to M/EEG source imaging
• Consistency of Bayesian inference with Gaussian process priors for a parabolic inverse problem
• Empirical priors and coverage of posterior credible sets in a sparse normal mean model
• A Bayesian nonparametric approach to log-concave density estimation
• Adaptive confidence bands for Markov chains and diffusions: Estimating the invariant measure and the drift
• Empirical priors and posterior concentration rates for a monotone density
• General framework for projection structures
• Posterior concentration rates for empirical Bayes procedures, with applications to Dirichlet Process mixtures
• The Bayes Lepski's Method and Credible Bands through Volume of Tubular Neighborhoods
• Adaptive inference over Besov spaces in the white noise model using $p$-exponential priors
• On frequentist coverage errors of Bayesian credible sets in moderately high dimensions
• Testing whether a Learning Procedure is Calibrated
• Comparing Scale Parameter Estimators for Gaussian Process Interpolation with the Brownian Motion Prior: Leave-One-Out Cross Validation and Maximum Likelihood
• Nonparametric Bayesian estimation of a Hölder continuous diffusion coefficient
• On coverage and local radial rates of DDM-credible sets
• Discussion of "Frequentist coverage of adaptive nonparametric Bayesian credible sets"
• On Semi-parametric Bernstein-von Mises Theorems for BART
• A sharp adaptive confidence ball for self-similar functions
• Sketching in Bayesian High Dimensional Regression With Big Data Using Gaussian Scale Mixture Priors
• Smoothness Estimation for Whittle-Matérn Processes on Closed Riemannian Manifolds
• Rejoinder to discussions of "Frequentist coverage of adaptive nonparametric Bayesian credible sets"
• Optimal regularized hypothesis testing in statistical inverse problems
• Spatially Adaptive Ensemble Learning with Calibrated Predictive Uncertainty
• A nonparametrically corrected likelihood for Bayesian spectral analysis of multivariate time series
• Variable Selection Consistency of Gaussian Process Regression
• Posterior Contraction and Credible Sets for Filaments of Regression Functions
• AutoCP: Automated Pipelines for Accurate Prediction Intervals
• Adaptive Non-parametric Estimation of Mean and Autocovariance in Regression with Dependent Errors
• Finite sample Bernstein-von Mises theorems for functionals and spectral projectors of the covariance matrix
• On frequentist coverage of Bayesian credible sets for estimation of the mean under constraints
• Uncertainty quantification for distributed regression
• Quasi-Bayesian Dual Instrumental Variable Regression
• Asymptotically optimal inference in sparse sequence models with a simple data-dependent measure
• Convergence Rates of Empirical Bayes Posterior Distributions: A Variational Perspective
• Minimax $L_2$-Separation Rate in Testing the Sobolev-Type Regularity of a function