Improving Bayesian Network Structure Search with Random Variable Aggregation Hierarchies

  • John Burge
  • Terran Lane
Part of the Lecture Notes in Computer Science book series (LNCS, volume 4212)


Bayesian network structure identification is known to be NP-Hard in the general case. We demonstrate a heuristic search for structure identification based on aggregation hierarchies. The basic idea is to perform initial exhaustive searches on composite “high-level” random variables (RVs) that are created via aggregations of atomic RVs. The results of the high-level searches then constrain a refined search on the atomic RVs. We demonstrate our methods on a challenging real-world neuroimaging domain and show that they consistently yield higher scoring networks when compared to traditional searches, provided sufficient topological complexity is permitted. On simulated data, where ground truth is known and controllable, our methods yield improved classification accuracy and structural precision, but can also result in reduced structural recall on particularly noisy datasets.


Bayesian network structure search hierarchy fMRI 


  1. 1.
    Anderson, C., Domingos, P., Weld, D.: Relational Markov models and their application to adaptive web navigation. In: KDD, pp.143–152 (2002)Google Scholar
  2. 2.
    Burge, J., Lane, T.: Learning Class-Discriminative Dynamic Bayesian Networks. ICML 22, 97–104 (2005)CrossRefGoogle Scholar
  3. 3.
    Buckner, R.L., Snyder, A., Sanders, A., Marcus, R., Morris, J.: Functional Brain Imaging of Young, Nondemented, and Demented Older Adults. Journal of Cognitive Neuroscience 12(2), 24–34 (2000)CrossRefGoogle Scholar
  4. 4.
    Chickering, D., Geiger, D., Heckerman, D.: Learning Bayesian Networks is NP-Hard. Technical Report MSR-TR-94-17, Microsoft (1994)Google Scholar
  5. 5.
    Clark, V.P., Friedman, L., Manoach, D., Ho, B.C., Lim, K., Andreasen, N.: A collaborative fMRI study of the novelty oddball task in schizophrenia: Effects of illness duration. Society for Neuroscience Abstracts, 474.9 (2005)Google Scholar
  6. 6.
    Dartmouth fMRI Data Center, The (2006),
  7. 7.
    Fine, S., Singer, Y., Tishby, N.: The Hierarchical Hidden Markov Model: Analysis and Applications. Machine Learning 32, 41–62 (1998)zbMATHCrossRefGoogle Scholar
  8. 8.
    Grossman, D., Domingos, P.: Learning Bayesian Network Classifiers by Maximizing Conditional Likelihood. In: ICML, vol. 21, pp. 361–368 (2004)Google Scholar
  9. 9.
    Gyftodimos, E., Flach, P.: Hierarchical Bayesian Networks: An Approach to Classification and Learning for Structured Data. In: Proceedings of Methods and Applications of Artificial Intelligence, Third Hellenic Conference in AI, pp. 291–300 (2004)Google Scholar
  10. 10.
    Heckerman, D., Geiger, D., Chickering, D.M.: Learning Bayesian Networks: the Combination of Knowledge and Statistical Data. Machine Learning 20, 197–243 (1995)zbMATHGoogle Scholar
  11. 11.
    Kiehl, K.: An event-related functional magnetic resonance imaging study of an auditory oddball task in schizophrenia. Schizophrenia Research 48, 159–171 (2001)CrossRefGoogle Scholar
  12. 12.
    Lancaster, J.L., Woldorff, M.G., Parsons, L.M., Liotti, M., Freitas, C.S., Rainey, L., Kochunov, P.V., Nickerson, D., Mikiten, S.A., Fox, P.T.: Automated Talairach Atlas labels for functional brain mapping. Human Brain Mapping 10, 120–131 (2000)CrossRefGoogle Scholar
  13. 13.
    Lam, W., Bacchus, F.: Learning Bayesian belief networks: An approach based on the MDL principle. Computational Intelligence 10, 269–293 (1994)CrossRefGoogle Scholar
  14. 14.
    Margaritis, D., Thrun, S.: Bayesian Network Induction via Local Neighborhoods. In: Advances in Neural Information Processing Systems 12, Denver, CO (1999)Google Scholar
  15. 15.
    McCallum, A., Rosenfeld, R., Mitchell, T., Ng, A.Y.: Improving Text Classification by Shrinkage in a Hierarchy of Classes. In: ICML (1998)Google Scholar
  16. 16.
    The MIND Institute (2006),
  17. 17.
    Pearl, J.: Fusion, Propagation and Structuring in Belief Networks. AI 29(3), 241–288 (1986)zbMATHMathSciNetGoogle Scholar
  18. 18.
    Schwarz, G.: Estimating the dimension of a model. Annals of Stats. 6, 461–464 (1978)zbMATHCrossRefGoogle Scholar
  19. 19.
    Sutton, R.S., Precup, D., Singh, S.: Between MDPs and semi-MDPs: A Framework for Temporal Abstraction in Reinforcement Learning. AI 112, 181–211 (1999)zbMATHMathSciNetGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2006

Authors and Affiliations

  • John Burge
    • 1
  • Terran Lane
    • 1
  1. 1.Department of Computer ScienceUniversity of New Mexico 

Personalised recommendations