Abstract
This paper addresses the problem of efficiently finding an optimal Bayesian network structure w.r.t. maximizing the posterior probability and minimizing the description length. In particular, we focus on the branch and bound strategy to save computational effort. To obtain an efficient search, a larger lower bound of the score is required (when we seek its minimum). We generalize an existing lower bound (Campose and Ji, 2011) for the Bayesian Dirichlet BDeu (Bayesian Dirichlet equivalent uniform) to one for the BD (Bayesian Dirichlet) and mathematically prove that the number of variables in each parent set cannot be bounded for maximizing the posterior probability.
Notes
- 1.
We denote \(X\perp \!\!\!\perp Y|Z\) if X and Y are independent given Z.
- 2.
The idea of using dynamic programing was invented by A. P. Singh & A. W. Moore (2005).
- 3.
At the same conference (Uncertainty in Artificial Intelligence 1993), Wai and Bucchus [20] presented another approach for MDL-based Bayesian network learning.
References
Akaike, H.: Information theory and an extension of the maximum likelihood principle. In: 2nd International Symposium on Information Theory, Budapest, Hungary (1973)
Buntine, W.: Theory refinement on Bayesian networks. In: Uncertainty in Artificial Intelligence, Los Angels, CA pp. 52–60 (1991)
Cussens, J., Bartlett, M.: GOBNILP 1.6.2 User/Developer Manual1, University of York (2015)
Chickering, D.M., Meek, C., Heckerman, D.: Large-sample learning of Bayesian networks is NP-hard. In: Uncertainty in Artificial Intelligence, Acapulco, Mexico, pp. 124–133 (2003)
Cooper, G.F., Herskovits, E.: A Bayesian method for the induction of probabilistic networks from data. Mach. Learn. 9(4), 309–347 (1992)
de Campos, C.P., Ji, Q.: Efficient structure learning of bayesian networks using constraints. J. Mach. Learn. Res. 12, 663–689 (2011)
Krichevsky, R.E., Trofimov, V.K.: The performance of universal encoding. IEEE Trans. Inf. Theory IT–27(2), 199–207 (1981)
Rissanen, J.: Modeling by shortest data description. Automatica 14, 465–471 (1978)
Silander, T., Myllymaki, P.: A simple approach for finding the globally optimal bayesian network structure. In: Uncertainty in Artificial Intelligence (2006)
Singh, A.P., Moore, A.W.: Finding optimal Bayesian networks by dynamic programming. Research Showcase@CMU (2005)
Spirtes, P., Glymour, C., Scheines, R.: Causation, Prediction and Search. Springer, Berlin (1993)
Pearl, J.: Probabilistic Reasoning in Intelligent Systems: Networks of Plausible Inference (Representation and Reasoning), 2nd edn. Morgan Kaufmann Pub., San Mateo (1988)
Suzuki, J.: Learning bayesian belief networks based on the minimum description length principle: an efficient algorithm using the B & B technique. In: International Conference on Machine Learning, pp. 462–470 (1996)
Suzuki, J.: A construction of bayesian networks from databases on an MDL principle. In: The Ninth Conference on Uncertainty in Artificial Intelligence, Washington D.C., pp. 266–273 (1993)
Suzuki, J.: The bayesian Chow-Liu algorithm. In: The Proceedings of the Sixth European Workshop on Probabilistic Graphical Models, Granada, Spain (2012)
Suzuki, J.: Consistency of learning bayesian network structures with continuous variables: an information theoretic approach. Entropy 2015 17(8), 5752–5770 (2015)
Tian, J.: A branch-and-bound algorithm for MDL learning Bayesian networks. In: Uncertainty in Artificial Intelligence, Palo Alto, pp. 580–588 (2000)
Ueno, M.: Robust learning Bayesian networks for prior belief. In: Uncertainty in Artificial Intelligence, Corvallis, Oregon, pp. 698–707 (2011)
Yuan, C., Malone, B.: Learning optimal bayesian networks: a shortest path perspective. J. Mach. Learn. Res. 48, 23–65 (2013)
Lam, W., Bacchus, F.: Using causal information and local measures to learn bayesian networks. In: The Ninth Conference on Uncertainty in Artificial Intelligence, Washington D.C., pp. 243–250 (1993)
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2015 Springer International Publishing Switzerland
About this paper
Cite this paper
Suzuki, J. (2015). Efficiently Learning Bayesian Network Structures Based on the B&B Strategy: A Theoretical Analysis. In: Suzuki, J., Ueno, M. (eds) Advanced Methodologies for Bayesian Networks. AMBN 2015. Lecture Notes in Computer Science(), vol 9505. Springer, Cham. https://doi.org/10.1007/978-3-319-28379-1_1
Download citation
DOI: https://doi.org/10.1007/978-3-319-28379-1_1
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-28378-4
Online ISBN: 978-3-319-28379-1
eBook Packages: Computer ScienceComputer Science (R0)