On the Markov equivalence of maximal ancestral graphs (Q2574688)

The paper is devoted to the class of maximal ancestral graphs as an appropriate extention of directed acyclic graphs. The reader may find useful to see the recent paper by \textit{T. S. Richardson} and \textit{P. Spirtes} [Ann. Stat. 30, 962--1030 (2002; Zbl 1033.60008)]. Among the elements of the graphical model is a family of probability functions describing marginal distributions and conditional independence. The goal is to study the relationship between such graphs and in particular to find conditions for a property called Markov equivalence. According to the authors, their results are useful in graphical models, Bayesian networks and exploratory data analysis.