Mixture Approximations to Bayesian Networks

Volker Tresp, Michael Haft, Reimar Hofmann

Structure and parameters in a Bayesian network uniquely specify the probability distribution of the modeled domain. The locality of both structure and probabilistic information are…