Determining full conditional independence by low-order conditioning

DOI10.3150/09-BEJ193zbMATH Open1206.60009arXiv0705.1613OpenAlexW2045131719MaRDI QIDQ605892FDOQ605892

Authors: Dhafer Malouche 

Publication date: 15 November 2010

Published in: Bernoulli (Search for Journal in Brave)

Abstract: A concentration graph associated with a random vector is an undirected graph where each vertex corresponds to one random variable in the vector. The absence of an edge between any pair of vertices (or variables) is equivalent to full conditional independence between these two variables given all the other variables. In the multivariate Gaussian case, the absence of an edge corresponds to a zero coefficient in the precision matrix, which is the inverse of the covariance matrix. It is well known that this concentration graph represents some of the conditional independencies in the distribution of the associated random vector. These conditional independencies correspond to the "separations" or absence of edges in that graph. In this paper we assume that there are no other independencies present in the probability distribution than those represented by the graph. This property is called the perfect Markovianity of the probability distribution with respect to the associated concentration graph. We prove in this paper that this particular concentration graph, the one associated with a perfect Markov distribution, can be determined by only conditioning on a limited number of variables. We demonstrate that this number is equal to the maximum size of the minimal separators in the concentration graph.

Full work available at URL: https://arxiv.org/abs/0705.1613

Recommendations

• Advances in the understanding and use of conditional independence
• A linear-algebraic tool for conditional independence inference
• Testing conditional independence via empirical likelihood
• Symbolic and Quantitative Approaches to Reasoning with Uncertainty
• Totally ordered conditional independence models
• Independence for full conditional probabilities: structure, factorization, non-uniqueness, and Bayesian networks
• scientific article; zbMATH DE number 3915395
• Efficient algorithms for conditional independence inference

zbMATH Keywords

graphical modelsconditional independenceundirected graphsconcentration graph modelsMarkov propertiesseparability in graphs

Mathematics Subject Classification ID

Random graphs (graph-theoretic aspects) (05C80) Combinatorial probability (60C05) Graph representations (geometric and intersection representations, etc.) (05C62)

Cites Work

• Causation, prediction, and search
• Introduction to Graphical Modelling
• Estimating high-dimensional directed acyclic graphs with the PC-algorithm
• Title not available (Why is that?)
• Title not available (Why is that?)
• Wishart distributions for decomposable graphs
• Flexible covariance estimation in graphical Gaussian models
• Logical and algorithmic properties of conditional independence and graphical models
• Title not available (Why is that?)
• A robust procedure for Gaussian graphical model search from microarray data with \(p\) larger than \(n\)
• Estimation of a covariance matrix with zeros
• Graphical methods for efficient likelihood inference in Gaussian covariance models
• Low-Order Conditional Independence Graphs for Inferring Genetic Networks
• Applications of HUGIN to diagnosis and control of autonomous vehicles

Cited In (5)

• Implications of faithfulness in graphical models
• Random variables with completely independent subcollections
• Gaussian covariance faithful Markov trees
• Title not available (Why is that?)
• Independence for full conditional measures and their graphoid properties

Uses Software

• pcalg
• TETRAD
• MIM

This page was built for publication: Determining full conditional independence by low-order conditioning

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q605892)