Explanation Trees for Causal Bayesian Networks

Bayesian networks can be used to extract explanations about the observed state of a subset of variables. In this paper, we ex- plicate the desiderata of an explanation and confront them with the concept of expla- nation proposed by existing methods. The necessity of taking into account causal ap- proaches when a causal graph is available is discussed. We then introduce causal expla- nation trees, based on the construction of ex- planation trees using the measure of causal information flow (Ay and Polani, 2006). This approach is compared to several other meth- ods on known networks.