Explaining Predictions in Bayesian Networks and Influence Diagrams