Bayes factors and the geometry of discrete hierarchical loglinear models

DOI10.1214/12-AOS974zbMATH Open1274.62391arXiv1103.5381OpenAlexW2963398855MaRDI QIDQ447834FDOQ447834

Authors: Gérard Letac, Hélène Massam 

Publication date: 29 August 2012

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

Abstract: A standard tool for model selection in a Bayesian framework is the Bayes factor which compares the marginal likelihood of the data under two given different models. In this paper, we consider the class of hierarchical loglinear models for discrete data given under the form of a contingency table with multinomial sampling. We assume that the Diaconis-Ylvisaker conjugate prior is the prior distribution on the loglinear parameters and the uniform is the prior distribution on the space of models. Under these conditions, the Bayes factor between two models is a function of their prior and posterior normalizing constants. These constants are functions of the hyperparameters $(m,alpha)$ which can be interpreted respectively as marginal counts and the total count of a fictive contingency table. We study the behaviour of the Bayes factor when $alpha$ tends to zero. In this study two mathematical objects play a most important role. They are, first, the interior $C$ of the convex hull $$ of the support of the multinomial distribution for a given hierarchical loglinear model together with its faces and second, the characteristic function $mathbb{J}_C$ of this convex set $C$. We show that, when $alpha$ tends to 0, if the data lies on a face $F_i$ of $$ of dimension $k_i$, the Bayes factor behaves like $alph{a}^{k_1-k_2}$. This implies in particular that when the data is in $C_1$ and in $C_2$, i.e. when $k_i$ equals the dimension of model $J_i$, the sparser model is favored, thus confirming the idea of Bayesian regularization.

Full work available at URL: https://arxiv.org/abs/1103.5381

Recommendations

• A conjugate prior for discrete hierarchical log-linear models
• scientific article; zbMATH DE number 4074232
• Bayesian selection of log‐linear models
• A hierarchical Bayesian model for contingency tables
• Markov chain Monte Carlo model determination for hierarchical and graphical log-linear models

zbMATH Keywords

BayesDirichlet distributionsparse contingency tablesconjugate priorscharacteristic function of a convex set

Mathematics Subject Classification ID

Contingency tables (62H17) Generalized linear models (logistic models) (62J12) Bayesian problems; characterization of Bayes procedures (62C10)

Cites Work

• Estimating the dimension of a model
• Title not available (Why is that?)
• A conjugate prior for discrete hierarchical log-linear models
• Conjugate priors for exponential families
• Hyper Markov laws in the statistical analysis of decomposable graphical models
• A fast procedure for model search in multidimensional contingency tables
• Title not available (Why is that?)
• Title not available (Why is that?)
• Bayes factors and the geometry of discrete hierarchical loglinear models
• On the toric algebra of graphical models
• On the choice of a model to fit data from an exponential family
• Title not available (Why is that?)
• Additive and multiplicative models and interactions
• Hypercube embedding of generalized bipartite metrics
• Gröbner bases and polyhedral geometry of reducible and cyclic models.
• Polyhedral conditions for the nonexistence of the MLE for hierarchical log-linear models

Cited In (9)

• Optimal Gaussian approximations to the posterior for log-linear models with Diaconis-Ylvisaker priors
• Modeling association in microbial communities with clique loglinear models
• Title not available (Why is that?)
• Loglinear model selection and human mobility
• A Bayesian graphical model for genome-wide association studies (GWAS)
• Confidence intervals for discrete log-linear models when MLE does not exist
• Local conditional and marginal approach to parameter estimation in discrete graphical models
• Bayes factors and the geometry of discrete hierarchical loglinear models
• Composite mixture of log-linear models with application to psychiatric studies

This page was built for publication: Bayes factors and the geometry of discrete hierarchical loglinear models

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q447834)