Learning Bayesian networks. I. A theory based on MAP-MDL criteria

Heping Pan1
1Digital Intelligence Research Centre School of Remote Sensing and Information Engineering, Wuhan University of China, Wuhan, China

Tóm tắt

Bayesian networks provide a powerful architecture for information fusion of multiple disparate variables. A theory of learning discrete Bayesian networks from data is presented in this paper. The theory is based on a joint criterion of maximizing the joint probability or interchangeably minimizing the joint description length of the data and the Bayesian network model including the network structure and the probability distribution parameters. The computable formalisms for the data likelihood given a structure, the description length of a structure, and the estimation of the parameters given a structure are derived. EM algorithms are constructed for handling incomplete and soft data. The theory leads to a computational algorithm described in a companion paper.

Từ khóa

#Bayesian methods #Probability distribution #NP-hard problem #Intelligent networks #Intelligent structures #Remote sensing #Power engineering and energy #Floors #Buildings #Parameter estimation

Tài liệu tham khảo

pan, 2002, Learning bayesian networks: Ii - A computational algorithm, 5th International Conference on Information Fusion Submitted 10.1016/0005-1098(78)90005-5 pan, 1992, An mdl-principled evolutionary mechanism to automatic architec- turing of pattern recognition neural network, IEEE Proc 11th Int Conference of Int Association of Pattern Recognition(IAPR) 10.1016/B978-1-55860-332-5.50018-3 friedman, 1998, The bayesian structural em algorithm, Proc Fourteenth Conference on Uncertainty in Artificial Intelligence(UAI'98) krause, 1998, Learning Probabilistic Networks 10.1109/69.494161 heckerraan, 1996, A Tutorial on Learning with Bayesian Networks (Revised) 10.1111/j.1467-8640.1994.tb00166.x geiger, 1995, A characterization of the dirichlet distribution with application to learning bayesian networks, Proc Eleventh Conference on Uncertainty in Artificial Intelli- gence(UAI'95) 10.1007/BF00994110 10.1142/S021800140000060X 10.1007/978-94-011-5014-9 10.1016/B978-1-4832-1451-1.50037-8 rissanen, 1989, Stochastic Complexity in Statistical Inquiry