Explaining Bayesian Networks using Argumentation

Abstract

Qualitative and quantitative systems to deal with uncertainty coexist. Bayesian networks are a well known tool in probabilistic reasoning. For non-statistical experts, however, Bayesian networks may be hard to interpret. Especially since the inner workings of Bayesian networks are complicated they may appear as black box models. Argumentation models, on the contrary, emphasise the derivation of results. However, they have notorious difficulty dealing with probabilities. In this paper we formalise a two-phase method to extract probabilistically supported arguments from a Bayesian network. First, from a BN we construct a support graph, and, second, given a set of observations we build arguments from that support graph. Such arguments can facilitate the correct interpretation and explanation of the evidence modelled in the Bayesian network.