B0948
Title: Faithfulness of probability distributions to graphs
Authors: Kayvan Sadeghi - University of Cambridge (United Kingdom) [presenting]
Abstract: We provide sufficient conditions for a given probability distribution to be Markov to a graph with the minimum possible number of edges, and more importantly, necessary and sufficient conditions for a given probability distribution to be faithful to a graph. We present our results briefly for the general case of mixed graphs, but in more details, specialize the definitions and results to the subclass directed acyclic graphs, which are essential elements of causal inference. Based on the results, we discuss new methods and algorithms for selecting graphical models that capture a given set of conditional independence statements.
