Binary Models for Marginal Independence

Abstract

Log-linear models are a classical tool for the analysis of contingency tables. In particular, the subclass of graphical log-linear models provides a general framework for modelling conditional independences. However, with the exception of special structures, marginal independence hypotheses cannot be accommodated by these traditional models. Focusing on binary variables, we present a model class that provides a framework for modelling marginal independences in contingency tables. The approach taken is graphical and draws on analogies to multivariate Gaussian models for marginal independence. For the graphical model representation we use bi-directed graphs, which are in the tradition of path diagrams. We show how the models can be parameterized in a simple fashion, and how maximum likelihood estimation can be performed using a version of the Iterated Conditional Fitting algorithm. Finally we consider combining these models with symmetry restrictions.

Cited by in corpus (24)

• Graphical modelling of multivariate time series
• Conjunctive Bayesian networks
• Discrete chain graph models
• Markov properties for mixed graphs
• Chain graph models of multivariate regression type for categorical data
• Marginal log-linear parameters for graphical Markov models
• Implicit inequality constraints in a binary tree model
• Log-mean linear models for binary data
• Parameter and Structure Learning in Nested Markov Models
• A factorization criterion for acyclic directed mixed graphs
• Sparse Nested Markov models with Log-linear Parameters
• Mixed Cumulative Distribution Networks
• A Transformational Characterization of Markov Equivalence for Directed Acyclic Graphs with Latent Variables
• Bayesian Analysis of Marginal Log-Linear Graphical Models for Three Way Contingency Tables
• On the Impact of Dimension Reduction on Graphical Structures
• Causal Structural Learning Via Local Graphs
• Smooth, identifiable supermodels of discrete DAG models with latent variables
• Learning from Pairwise Marginal Independencies
• Marginal Independence Models
• Parameterizations and fitting of bi-directed graph models to categorical data
• Bayesian Inference in Cumulative Distribution Fields
• Random Networks, Graphical Models, and Exchangeability
• Probability Based Independence Sampler for Bayesian Quantitative Learning in Graphical Log-Linear Marginal Models
• Graphical methods for efficient likelihood inference in Gaussian covariance models