An Explanation Mechanism for Bayesian Inferencing Systems