The Bayesian analysis of complex, high-dimensional models: can it be CODA?

DOI10.1214/14-STS483MaRDI QIDQ252815zbMATH OpenOpenAlexFDO

Authors B. J. K. Kleijn, P. J. Bickel, A. C. Gamst, Yaacov Ritov

Publication date 4 March 2016

Published in Statistical Science (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1203.5471, https://projecteuclid.org/euclid.ss/1421330550

zbMATH Keywords

Bayesian inferencesemiparametricsstopping timeCODAfoundationsfunctional estimationpartial linear modelwhite noise models

Mathematics Subject Classification ID

Bayesian inference (62F15) Nonparametric regression and quantile regression (62G08) Linear regression; mixed models (62J05) Estimation in multivariate analysis (62H12) Diffusion processes (60J60)

Abstract: We consider the Bayesian analysis of a few complex, high-dimensional models and show that intuitive priors, which are not tailored to the fine details of the model and the estimated parameters, produce estimators which perform poorly in situations in which good, simple frequentist estimators exist. The models we consider are: stratified sampling, the partial linear model, linear and quadratic functionals of white noise and estimation with stopping times. We present a strong version of Doob's consistency theorem which demonstrates that the existence of a uniformly $sqrtn$-consistent estimator ensures that the Bayes posterior is $sqrtn$-consistent for values of the parameter in subsets of prior probability 1. We also demonstrate that it is, at least, in principle, possible to construct Bayes priors giving both global and local minimax rates, using a suitable combination of loss functions. We argue that there is no contradiction in these apparently conflicting findings.

Recommendations

• An introduction to Bayesian analysis. Theory and methods.
• Bayesian estimation in a high dimensional parameter framework
• Bayesian aspects of some nonparametric problems
• Bayesian model selection in high-dimensional settings
• Necessary and sufficient conditions for high-dimensional posterior consistency under \(g\)-priors

Cites work

• scientific article; zbMATH DE number 3844864 (Why is no real title available?)
• scientific article; zbMATH DE number 4107941 (Why is no real title available?)
• scientific article; zbMATH DE number 3549966 (Why is no real title available?)
• scientific article; zbMATH DE number 1220667 (Why is no real title available?)
• scientific article; zbMATH DE number 1181283 (Why is no real title available?)
• scientific article; zbMATH DE number 2038320 (Why is no real title available?)
• scientific article; zbMATH DE number 6193702 (Why is no real title available?)
• scientific article; zbMATH DE number 3390151 (Why is no real title available?)
• scientific article; zbMATH DE number 3195782 (Why is no real title available?)
• scientific article; zbMATH DE number 3080020 (Why is no real title available?)
• A Bayesian analysis of some nonparametric problems
• A difference based approach to the semiparametric partial linear model
• A statistical analysis of multiple temperature proxies: are reconstructions of surface temperatures over the last 1000 years reliable?
• Adapting to Unknown Smoothness via Wavelet Shrinkage
• An analysis of Bayesian inference for nonparametric regression
• Are Bayesian inferences weak for Wasserman's example?
• Asymptotic Inference for Mixture Models by Using Data-Dependent Priors
• Asymptotic equivalence of density estimation and Gaussian white noise
• Asymptotic equivalence of nonparametric regression and white noise
• Asymptotics in statistics: some basic concepts
• Bayesian aspects of some nonparametric problems
• Consistency of Bayes estimates for nonparametric regression: normal theory
• Convergence rates of posterior distributions.
• Data-driven efficient estimators for a partially linear model
• Minimax estimation of the mean of a normal distribution when the parameter space is restricted
• Minimax risk over \(l_ p\)-balls for \(l_ q\)-error
• Misspecification in infinite-dimensional Bayesian statistics
• Nonparametric binary regression: A Bayesian approach
• Nonparametric estimators which can be ``plugged-in.
• On efficient estimation in regression models
• On the Asymptotic Behavior of Bayes' Estimates in the Discrete Case
• On the Bernstein-von Mises theorem with infinite-dimensional parameters
• Rejoinder: ``The case for objective Bayesian analysis
• Semiparametric minimax rates
• Subjective Bayesian analysis: principles and practice
• The case for objective Bayesian analysis
• The horseshoe estimator: posterior concentration around nearly black vectors
• The interplay of Bayesian and frequentist analysis
• The semiparametric Bernstein-von Mises theorem

Cited in

(9)

• Prior and posterior checking of implicit causal assumptions
• ValidP-Values Behave Exactly as They Should: Some Misleading Criticisms ofP-Values and Their Resolution WithS-Values
• Consistency of Bayesian inference for multivariate max-stable distributions
• Semiparametric Bayesian causal inference
• Are Bayesian inferences weak for Wasserman's example?
• Assumption-lean falsification tests of rate double-robustness of double-machine-learning estimators
• Criteria for posterior consistency and convergence at a rate
• On nearly assumption-free tests of nominal confidence interval coverage for causal parameters estimated by machine learning
• Bayesian estimation in a high dimensional parameter framework

This page was built for publication: The Bayesian analysis of complex, high-dimensional models: can it be CODA?

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q252815)