Compatibility of prior specifications across linear models

DOI10.1214/08-STS258zbMATH Open1329.62331arXiv1102.2981MaRDI QIDQ908147FDOQ908147

Authors: Guido Consonni, Piero Veronese 

Publication date: 3 February 2016

Published in: Statistical Science (Search for Journal in Brave)

Abstract: Bayesian model comparison requires the specification of a prior distribution on the parameter space of each candidate model. In this connection two concerns arise: on the one hand the elicitation task rapidly becomes prohibitive as the number of models increases; on the other hand numerous prior specifications can only exacerbate the well-known sensitivity to prior assignments, thus producing less dependable conclusions. Within the subjective framework, both difficulties can be counteracted by linking priors across models in order to achieve simplification and compatibility; we discuss links with related objective approaches. Given an encompassing, or full, model together with a prior on its parameter space, we review and summarize a few procedures for deriving priors under a submodel, namely marginalization, conditioning, and Kullback--Leibler projection. These techniques are illustrated and discussed with reference to variable selection in linear models adopting a conventional $g$-prior; comparisons with existing standard approaches are provided. Finally, the relative merits of each procedure are evaluated through simulated and real data sets.

Full work available at URL: https://arxiv.org/abs/1102.2981

Recommendations

• Joint specification of model space and parameter space prior distributions
• Compatible priors for Bayesian model comparison with an application to the Hardy-Weinberg equilibrium model
• Bayesian model selection using encompassing priors
• Benchmark priors for Bayesian model averaging.
• Expected-posterior prior distributions for model selection

zbMATH Keywords

variable selectionhypothesis testingBayes factorconjugate prior\(g\)-priornested modelcompatible priorKullback-Leibler projection

Mathematics Subject Classification ID

Bayesian inference (62F15) Analysis of variance and covariance (ANOVA) (62J10) Bayesian problems; characterization of Bayes procedures (62C10)

Cites Work

• Title not available (Why is that?)
• Model choice: a minimum posterior predictive loss approach
• Nonparametric regression using Bayesian variable selection
• Model uncertainty
• Optimal predictive model selection.
• Title not available (Why is that?)
• Title not available (Why is that?)
• Extending conventional priors for testing general hypotheses in linear models
• Mixtures of g Priors for Bayesian Variable Selection
• Title not available (Why is that?)
• Model choice in generalised linear models: a Bayesian approach via Kullback-Leibler projections
• Title not available (Why is that?)
• Title not available (Why is that?)
• Bayesian Model Averaging for Linear Regression Models
• Calibration and empirical Bayes variable selection
• The Intrinsic Bayes Factor for Model Selection and Prediction
• The Weighted Likelihood Ratio, Linear Hypotheses on Normal Location Parameters
• Efficient Empirical Bayes Variable Selection and Estimation in Linear Models
• Title not available (Why is that?)
• Title not available (Why is that?)
• Benchmark priors for Bayesian model averaging.
• Comparison of Bayesian objective procedures for variable selection in linear regression
• Statistical Models
• Objective Bayesian Variable Selection
• Training samples in objective Bayesian model selection.
• Expected-posterior prior distributions for model selection
• Bayesian thinking, modeling and computation.
• Bayesian Hypothesis Testing: A Reference Approach
• Prediction Via Orthogonalized Model Mixing
• The Variable Selection Problem
• Assessing robustness of intrinsic tests of independence in two-way contingency tables
• Compatible Prior Distributions for Directed Acyclic Graph Models
• Parsimonious estimation of multiplicative interaction in analysis of variance using Kullback-Leibler information
• Title not available (Why is that?)
• Variable selection in qualitative models via an entropic explanatory power
• A Bayesian approach to selecting covariates for prediction
• Compatible priors for Bayesian model comparison with an application to the Hardy-Weinberg equilibrium model
• Tests based on intrinsic priors for the equality of two correlated proportions
• Title not available (Why is that?)
• Predictive specification of prior model probabilities in variable selection
• Bayes factors for nonlinear hypotheses and likelihood distributions

Cited In (33)

• Objective methods for graphical structural learning
• Joint specification of model space and parameter space prior distributions
• Compatible priors for model selection of high-dimensional Gaussian DAGs
• Penalising model component complexity: a principled, practical approach to constructing priors
• On consistency and optimality of Bayesian variable selection based on \(g\)-prior in normal linear regression models
• Power-expected-posterior priors for variable selection in Gaussian linear models
• Bayesian sample size determination for causal discovery
• Nonparametric Bayes classification and hypothesis testing on manifolds
• Bayesian effect estimation accounting for adjustment uncertainty
• Prior distributions for objective Bayesian analysis
• An encompassing prior generalization of the Savage-Dickey density ratio
• Learning Markov equivalence classes of directed acyclic graphs: an objective Bayes approach
• Power-expected-posterior priors for generalized linear models
• Priors via imaginary training samples of sufficient statistics for objective Bayesian hypothesis testing
• Objective Bayesian comparison of constrained analysis of variance models
• An evaluation of alternative methods for testing hypotheses, from the perspective of Harold Jeffreys
• On the correspondence from Bayesian log-linear modelling to logistic regression modelling with \(g\)-priors
• Bayesian model comparison based on expected posterior priors for discrete decomposable graphical models
• On resolving the Savage-Dickey paradox
• On the invariance of conditioning procedures for the specification of prior distributions for nested DAG models
• Information consistency of the Jeffreys power-expected-posterior prior in Gaussian linear models
• Compatible priors for Bayesian model comparison with an application to the Hardy-Weinberg equilibrium model
• A Generalization of the Savage–Dickey Density Ratio for Testing Equality and Order Constrained Hypotheses
• Shrinkage priors via random imaginary data
• Conjugate and conditional conjugate Bayesian analysis of discrete graphical models of marginal independence
• Objective Bayesian comparison of order-constrained models in contingency tables
• A comparison of power-expected-posterior priors in shrinkage regression
• Objective Bayesian transformation and variable selection using default Bayes factors
• Bayesian estimation and hypothesis tests for a circular generalized linear model
• A general approach to prior transformation
• Power-expected-posterior priors as mixtures of \(g\)-priors in normal linear models
• A general approximation to nested Bayes factors with informed priors
• Power-Expected-Posterior Methodology with Baseline Shrinkage Priors

This page was built for publication: Compatibility of prior specifications across linear models

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q908147)