Abstract
A Markov field represents dependencies in the form of an undirected graph and is a natural representation for specific types of knowledge, such as networks of concepts. A Bayesian learning algorithm is presented for inferring structure and parameters from a database of independent and identically distributed cases from the distribution. We use a conjugate family of prior distributions for model parameters conditional on connectivity structure, and the Bayesian Information Criterion to approximate the posterior probabilities for structures. The procedure is conceived for the general setting of experimental situations, where the correlation structure of a system of attributes has to be learned from examples. Correlations of order two and higher correspond to concepts. The graph induced by drawing cliques corresponding to the detected correlations and/or interactions provides information on the rules to be extracted.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
References
J.J. Hopfield, “Neural networks and Physics systems with emergent collective computational abilities, Proceedings of the National Academy of Science, USA (1984): 2554–2558
L. Martignon, H.v. Hasseln, S. Grün, A. Aertsen and G. Palm, “Detecting higher order interactions among the spiking events in a group of neurons” to appear in Biological Cybernetics (1995)
G. Hinton, J. McClelland, D. Rumelhart.“Distributed Representations” Parallel distributed processing, D.Rumelhart and J. McClelland (eds) (The MIT Press, Cambridge, 1986 )
G. Hinton, T. Sejnowski,“Learning and relearning in Boltzmann machines”in Parallel distributed processing,D. Rumelhart and J. Mc Clelland (eds) ( The MIT Press, Cambridge, 1986 )
K.B. Laskey, “Adapting connectionist learning to Bayes networks”, International Journal of Approximate Reasoning, May (1990)
S. Amari, “Dualistic geometry of the Manifold of Higher—Order Neurons”, Neural Networks, 4, (1991), p. 443–451
S. Amari, K. Kurata, and H. Nagaoka,“Information Geometry of Boltzmann machines”, IEEE Trans. Neural networks, 3, (1992) no. 2, p. 260–271
S. Amari and T. SunHan, “Statistical Inference under Multiterminal Rate Restrictions: A Differential Geometric Approach”, IEEE Trans.Inf.Theory, 35, (1989), no. 2, p. 217–227
Y. Bishop, S. Fienberg, and P. Holland, Discrete multivariate analysis: Theory and Practice, (The MIT Press, Cambridge, 1989)
R.E. Kass, A. Raftery, “Bayes factors and model uncertainty”,Carnegie Mellon University Dept. of Statistics, Technical Report Nr.571(1993)
M.H. de Groot, Optimal Statistical Decisions,(Mc.Graw Hill, New York,1970)
K.B. Laskey and L. Martignon, “ Bayesian Learning of Markov Fields Models”submitted to Neural Networks
W.B. Poland, and R. Schachter,“Three approaches to probability model selection”, Uncertainty in artificial intelligence:Proceedings of the tenth conference eds. R.Lopez de Mantara, and D.Poole (eds.)(Morgan Kaufman, San Mateo,1994)
S. Grün, A. Aertsen, M. Abeles, G. Gerstein, and G. Palm,“Behavior— related neuron group activity in the cortex”, Proc. 17th Ann. Meeting of the European neurosc. Ass.(Oxford University Press, ENA), (1994)
G. Cooper and E. Herskovits,“A Bayesian method for the induction of probabilistic networks from data”,Machine Learning 9 (1992),p. 309–347
D. Heckerman, D. Geiger, D. M. Chickering,“Learning Bayesian networks:The combination of knowledge and statistical data”,Artificial Intelligence“(to appear).
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 1995 Springer-Verlag London Limited
About this paper
Cite this paper
Martignon, L., Laskey, K.B. (1995). Bayesian Strategies for Machine Learning: Rule Extraction and Concept Detection. In: Kappen, B., Gielen, S. (eds) Neural Networks: Artificial Intelligence and Industrial Applications. Springer, London. https://doi.org/10.1007/978-1-4471-3087-1_17
Download citation
DOI: https://doi.org/10.1007/978-1-4471-3087-1_17
Publisher Name: Springer, London
Print ISBN: 978-3-540-19992-2
Online ISBN: 978-1-4471-3087-1
eBook Packages: Springer Book Archive