Improved minimax predictive densities under Kullback-Leibler loss

DOI10.1214/009053606000000155zbMATH Open1091.62003arXivmath/0605432OpenAlexW2056860084MaRDI QIDQ2493546FDOQ2493546

Authors: Edward I. George, Feng Liang, Xinyi Xu 

Publication date: 21 June 2006

Published in: The Annals of Statistics (Search for Journal in Brave)

Abstract: Let $X|musim{N}_p(mu,v_{x}I)$ and $Y|musim{N}_p(mu,v_{y}I)$ be independent p-dimensional multivariate normal vectors with common unknown mean $mu$. Based on only observing $X=x$, we consider the problem of obtaining a predictive density $hatp(y|x)$ for $Y$ that is close to $p(y|mu)$ as measured by expected Kullback--Leibler loss. A natural procedure for this problem is the (formal) Bayes predictive density $hat{p}_{mathrmU}(y|x)$ under the uniform prior $p{i}_{mathrmU}(mu)equiv1$, which is best invariant and minimax. We show that any Bayes predictive density will be minimax if it is obtained by a prior yielding a marginal that is superharmonic or whose square root is superharmonic. This yields wide classes of minimax procedures that dominate $hat{p}_{mathrmU}(y|x)$, including Bayes predictive densities under superharmonic priors. Fundamental similarities and differences with the parallel theory of estimating a multivariate normal mean under quadratic loss are described.

Full work available at URL: https://arxiv.org/abs/math/0605432

Recommendations

• Admissible predictive density estimation
• Proper Bayes and minimax predictive densities related to estimation of a normal mean matrix
• Asymptotically minimax Bayes predictive densities
• On efficient prediction and predictive density estimation for normal and spherically symmetric models
• PREDICTIVE DENSITY ESTIMATION FOR MULTIPLE REGRESSION

zbMATH Keywords

multiple shrinkageheat equationmultivariate normalinadmissibilityshrinkage estimationunbiased risk estimationBayes rulessuperharmonic priorssuperharmonic marginals

Mathematics Subject Classification ID

Bayesian inference (62F15) Bayesian problems; characterization of Bayes procedures (62C10) Minimax procedures in statistical decision theory (62C20)

Cites Work

• Estimation of the mean of a multivariate normal distribution
• Title not available (Why is that?)
• Statistical decision theory and Bayesian analysis. 2nd ed
• Title not available (Why is that?)
• Predictive fit for natural exponential families
• The maximum likelihood prior
• Simultaneous prediction of independent Poisson observables
• Asymptotically minimax Bayes predictive densities
• Nonsubjective priors via predictive relative entropy regret
• A shrinkage predictive distribution for multivariate normal observables
• On asymptotic properties of predictive distributions
• Goodness of prediction fit
• On the construction of Bayes minimax estimators
• Title not available (Why is that?)
• Proper Bayes Minimax Estimators of the Multivariate Normal Mean
• The heat equation and Stein's identity: connections, applications
• Exact Minimax Strategies for Predictive Density Estimation, Data Compression, and Model Selection
• Admissible Estimators, Recurrent Diffusions, and Insoluble Boundary Value Problems
• Minimax multiple shrinkage estimation
• Combining Minimax Shrinkage Estimators
• On the estimation of parametric density functions
• A Note on the Estimation of Probability Density Functions
• A formal bayes multiple shrinkage estimator

Cited In (59)

• Bayesian analysis of hypothesis testing problems for general population: A Kullback-Leibler alternative
• On predictive density estimation for location families under integrated squared error loss
• Bayesian shrinkage prediction for the regression problem
• Enriched standard conjugate priors and the right invariant prior for Wishart distributions
• Estimation, prediction and the Stein phenomenon under divergence loss
• Improved prediction for a multivariate normal distribution with unknown mean and variance
• On the Stein phenomenon under divergence loss and an unknown variance-covariance matrix
• Admissible predictive density estimation
• Asymptotic properties of Bayesian predictive densities when the distributions of data and target variables are different
• Proper Bayes and minimax predictive densities related to estimation of a normal mean matrix
• Stochastic domination in predictive density estimation for ordered normal means under \(\alpha\)-divergence loss
• Quantifying the fraction of missing information for hypothesis testing in statistical and genetic studies
• Parametric bootstrap approximation to the distribution of EBLUP and related prediction intervals in linear mixed models
• Estimation in a linear regression model under the Kullback-Leibler loss and its application to model selection
• Asymptotic minimax risk of predictive density estimation for non-parametric regression
• Density prediction and the Stein phenomenon
• On predictive density estimation under \(\alpha\)-divergence loss
• Bayesian predictive distribution for a negative binomial model
• Pitman closeness properties of point estimators and predictive densities with parametric constraints
• On improved predictive density estimation with parametric constraints
• Exact minimax estimation of the predictive density in sparse Gaussian models
• Minimaxity in predictive density estimation with parametric constraints
• A superharmonic prior for the autoregressive process of the second-order
• A class of proper priors for Bayesian simultaneous prediction of independent Poisson observ\-a\-bles
• Shrinkage priors for single-spiked covariance models
• Title not available (Why is that?)
• Asymptotically minimax Bayesian predictive densities for multinomial models
• Simultaneous prediction for independent Poisson processes with different durations
• From minimax shrinkage estimation to minimax shrinkage prediction
• Inadmissibility of the best equivariant predictive density in the unknown variance case
• Minimax predictive density for sparse count data
• On discrete priors and sparse minimax optimal predictive densities
• Pitman closeness properties of Bayes shrinkage procedures in estimation and prediction
• Asymptotically minimax Bayes predictive densities
• Bayesian predictive densities based on superharmonic priors for the 2-dimensional Wishart model
• Transport analysis of infinitely deep neural network
• Shrinkage priors for Bayesian prediction
• Larry Brown's work on admissibility
• Empirical Bayes predictive densities for high-dimensional normal models
• Bayesian predictive density estimation for a chi-squared model using information from a normal observation with unknown mean and variance
• Bayesian inference and prediction for mean-mixtures of normal distributions
• Asymptotic expansion of the risk difference of the Bayesian spectral density in the autoregressive moving average model
• Predictive density estimation under the Wasserstein loss
• Bayesian predictive distribution for a Poisson model with a parametric restriction
• PREDICTIVE DENSITY ESTIMATION FOR MULTIPLE REGRESSION
• On efficient prediction and predictive density estimation for normal and spherically symmetric models
• On minimax optimality of sparse Bayes predictive density estimates
• On predictive density estimation with additional information
• On predictive density estimation for gamma models with parametric constraints
• Hierarchical empirical Bayes estimation of two sample means under divergence loss
• A numerical comparison of some prediction densities for the normal linear model
• Prediction risk for the horseshoe regression
• Predictive densities for multivariate normal models based on extended models and shrinkage Bayes methods
• Title not available (Why is that?)
• Matrix variate density estimation with additional information
• Predictive density estimators with integrated \(L_1\) loss
• Title not available (Why is that?)
• Optimal shrinkage estimation of predictive densities under \(\alpha\)-divergences
• Title not available (Why is that?)

This page was built for publication: Improved minimax predictive densities under Kullback-Leibler loss

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2493546)