Marginalization and conditioning for LWF chain graphs (Q309746)

This paper deals with graphical models for chain graphs (CG) with a specific interpretation of conditional independence, generally known as the Lauritzen-Wermuth-Frydenberg (LWF) interpretation. The author focuses on the problem of marginalization over and conditioning on two disjoint subsets of the node set of CG -- with LWF Markov property. Finally he also discusses ``the implications of the results for probabilistic independence models that are faithful to LWF CGs, and possible ways to generalize the parametrizations existing in the literature for chain mixed graphs and anterial graphs''.