Finding Explanations in Bayesian Networks

Changhe Yuan, Tsai-Ching Lu


Maximum a Posteriori (MAP) and Most Probable Explanation (MPE) are popular approaches of finding explanations for given observations in Bayesian networks. One of their shortcomings is that they have to find a complete assignment to a set of target variables. However, it is often the case that only few of the target variables are most relevant in explaining the observations. In this paper, we formulate the problem of finding the most relevant variables as explanatory MAP (eMAP), which considers all subsets of the target variables and finds a partial assignment that maximizes a chosen quality measure. We consider two quality measures for eMAP: Bayes Factor (or Weight of Evidence) and likelihood function. We then illustrate the proposed methodology on a circuit diagnosis problem in literature.



author = {Changhe Yuan and Tsai-Ching Lu},
title = {Finding Explanations in {B}ayesian Networks},
booktitle = {Proceedings of the 18th International Workshop on Principles of Diagnosis (DX-07)},
year = {2007},
pages = {414--419}