On expressiveness of the chain graph interpretations

Highlights

• We study the expressiveness of the different chain graph interpretations in terms of representable independence models.

• We show how an MCMC sampling algorithm can be used to sample chain graph models.

• We show that chain graphs can represent exponentially many more independence models compared to Bayesian networks.

• We show (numerically) that the average number of chain graphs per chain graph model appears to converge.

• We show that AMP and MVR chain graphs are more expressive than LWF chain graphs.

Abstract

In this article we study the expressiveness of the different chain graph interpretations. Chain graphs is a class of probabilistic graphical models that can contain two types of edges, representing different types of relationships between the variables in question. Chain graphs is also a superclass of directed acyclic graphs, i.e. Bayesian networks, and can thereby represent systems more accurately than this less expressive class of models. Today there do however exist several different ways of interpreting chain graphs and what conditional independences they encode, giving rise to different so-called chain graph interpretations. Previous research has approximated the number of representable independence models for the Lauritzen–Wermuth–Frydenberg and the multivariate regression chain graph interpretations using an MCMC based approach. In this article we use a similar approach to approximate the number of models representable by the latest chain graph interpretation in research, the Andersson–Madigan–Perlman interpretation. Moreover we summarize and compare the different chain graph interpretations with each other. Our results confirm previous results that directed acyclic graphs only can represent a small fraction of the models representable by chain graphs, even for a low number of nodes. The results also show that the Andersson–Madigan–Perlman and multivariate regression interpretations can represent about the same amount of models and twice the amount of models compared to the Lauritzen–Wermuth–Frydenberg interpretation. However, at the same time almost all models representable by the latter interpretation can only be represented by that interpretation while the former two have a large intersection in terms of representable models.