A simple method to losslessly decompose Bayesian Networks
Yanfeng Zhang · Yunnan Daxue xuebao. Shehui kexue ban · 2009
As Bayesian Networks becomes popular tools for common knowledge representation and reasoning of partial beliefs under uncertainty,Bayesian Networks have been successfully applied to a variety of problem domains.Confronted with many real-world applications,Bayesian Networks established become larger and more complex.It is not realistic to infer directly on these models.Thus,losslessly decomposing large and complex Bayesian Networks into smaller submodels is to be an alternative solution.Based on the properties of marginal models.A simpler method to losslessly decompose the Bayesian network into a set of smaller Bayesian networks is given.