Context-specific independence in graphical log-linear models

DOI10.1007/S00180-015-0606-6zbMATH Open1348.65036arXiv1409.2713OpenAlexW1507811118MaRDI QIDQ333386FDOQ333386

Jukka Corander, Henrik Nyman, Johan Pensar, Timo Koski

Publication date: 28 October 2016

Published in: Computational Statistics (Search for Journal in Brave)

Abstract: Log-linear models are the popular workhorses of analyzing contingency tables. A log-linear parameterization of an interaction model can be more expressive than a direct parameterization based on probabilities, leading to a powerful way of defining restrictions derived from marginal, conditional and context-specific independence. However, parameter estimation is often simpler under a direct parameterization, provided that the model enjoys certain decomposability properties. Here we introduce a cyclical projection algorithm for obtaining maximum likelihood estimates of log-linear parameters under an arbitrary context-specific graphical log-linear model, which needs not satisfy criteria of decomposability. We illustrate that lifting the restriction of decomposability makes the models more expressive, such that additional context-specific independencies embedded in real data can be identified. It is also shown how a context-specific graphical model can correspond to a non-hierarchical log-linear parameterization with a concise interpretation. This observation can pave way to further development of non-hierarchical log-linear models, which have been largely neglected due to their believed lack of interpretability.

Full work available at URL: https://arxiv.org/abs/1409.2713

zbMATH Keywords

parameter estimationgraphical modellog-linear modelcontext-specific interaction model

Mathematics Subject Classification ID

Computational methods for problems pertaining to statistics (62-08) Contingency tables (62H17) Graphical methods in statistics (62A09)

Cites Work

• Title not available (Why is that?)
• Title not available (Why is that?)
• Title not available (Why is that?)
• Title not available (Why is that?)
• Estimating the dimension of a model
• A fast procedure for model search in multidimensional contingency tables
• Labeled directed acyclic graphs: a generalization of context-specific independence in directed graphical models
• Labelled Graphical Models
• Statistical Inference in Context Specific Interaction Models for Contingency Tables
• I-divergence geometry of probability distributions and minimization problems
• Algorithmic graph theory and perfect graphs
• Information projections revisited
• A new algorithm for the maximum likelihood estimation of graphical log-linear models
• Marginal and simultaneous predictive classification using stratified graphical models
• Stratified graphical models -- context-specific independence in graphical models
• Split models for contingency tables

Cited In (8)

• On the role of the overall effect in exponential families
• Structure Learning of Contextual Markov Networks using Marginal Pseudo‐likelihood
• A Logical Approach to Context-Specific Independence
• Hierarchical Aitchison-Silvey models for incomplete binary sample spaces
• Context-Specific Independencies Embedded in Chain Graph Models of Type I
• Context-Specific and Local Independence in Markovian Dependence Structures
• Context-specific independencies in stratified chain regression graphical models
• Context-specific independencies in hierarchical multinomial marginal models

Recommendations

• Labeled directed acyclic graphs: a generalization of context-specific independence in directed graphical models 👍 👎
• Logical and algorithmic properties of conditional independence and graphical models 👍 👎
• CONDITIONAL INDEPENDENCE STRUCTURES AND GRAPHICAL MODELS 👍 👎
• A graphical characterization of lattice conditional independence models 👍 👎
• Context-specific independencies in stratified chain regression graphical models 👍 👎
• Context-specific independencies in hierarchical multinomial marginal models 👍 👎
• Marginal log-linear parameterization of conditional independence models 👍 👎
• Log‐mean Linear Parameterization for Discrete Graphical Models of Marginal Independence and the Analysis of Dichotomizations 👍 👎
• Stratified graphical models -- context-specific independence in graphical models 👍 👎

This page was built for publication: Context-specific independence in graphical log-linear models

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q333386)