A probabilistic reasoning algorithm for Bayesian networks by simplifying their structures

This article describes a new probabilistic reasoning algorithm for Bayesian networks, one of the stochastic models. The proposed method, called Extended LBPC, takes advantage of existing reasoning methods and statistical techniques such as hypothesis testing and interval estimate. Several computer simulations demonstrate that the proposed algorithm can perform appropriate and adaptable probabilistic inference depending on the complexity of problems in terms of both accuracy and computational time, compared to other conventional methods.