Open Access
August 2014 Markovian acyclic directed mixed graphs for discrete data
Robin J. Evans, Thomas S. Richardson
Ann. Statist. 42(4): 1452-1482 (August 2014). DOI: 10.1214/14-AOS1206


Acyclic directed mixed graphs (ADMGs) are graphs that contain directed ($\rightarrow$) and bidirected ($\leftrightarrow$) edges, subject to the constraint that there are no cycles of directed edges. Such graphs may be used to represent the conditional independence structure induced by a DAG model containing hidden variables on its observed margin. The Markovian model associated with an ADMG is simply the set of distributions obeying the global Markov property, given via a simple path criterion (m-separation). We first present a factorization criterion characterizing the Markovian model that generalizes the well-known recursive factorization for DAGs. For the case of finite discrete random variables, we also provide a parameterization of the model in terms of simple conditional probabilities, and characterize its variation dependence. We show that the induced models are smooth. Consequently, Markovian ADMG models for discrete variables are curved exponential families of distributions.


Download Citation

Robin J. Evans. Thomas S. Richardson. "Markovian acyclic directed mixed graphs for discrete data." Ann. Statist. 42 (4) 1452 - 1482, August 2014.


Published: August 2014
First available in Project Euclid: 7 August 2014

zbMATH: 1302.62148
MathSciNet: MR3262457
Digital Object Identifier: 10.1214/14-AOS1206

Primary: 62M45

Keywords: Acyclic directed mixed graph , Conditional independence , curved exponential family , Graphical model , m-separation , parameterization

Rights: Copyright © 2014 Institute of Mathematical Statistics

Vol.42 • No. 4 • August 2014
Back to Top