Selecting dynamic graphical models with hidden variables from data
- 330 Downloads
Selecting graphical models for a set of variables from data consists of finding the graphical structure and its associated probability distribution which best fit the data. In this paper we propose a new method for selecting Markovian dynamic graphical models from data and, in particular, we develop a new Bayesian technique for selecting graphical hidden Markov models, depicted by a chain graph, from an incomplete data set where values corresponding to hidden or latent variables are not present in data. The proposed method is illustrated by a case study.
KeywordsGraphical Models Hidden Markov Models Bayesian Learning
- Castillo, E.; Gutiérrez, J.M. & Hadi, A.S. (1997), Expert systems and probabilistic network models. Springer.Google Scholar
- Geiger, D. & Heckerman, D. (1994), ‘Learning Gaussian networks’, in ‘Proceedings of the 10th Conference on Uncertainty in Artificial Intelligence (UAI-94)’, Standford University, 29–31th July, 235–243.Google Scholar
- Heckerman, D. (1995), A tutorial on learning with Bayesian networks. Technical Report MSR-TR-95-06. ftp://ftp.research.microsoft.com/pub/techreports/winter94-95/tr-95-06.ps.
- Heckerman, D. & Geiger, D. (1995), ‘Learning Bayesian networks: A unification for discrete and Gaussian domains’, in ‘Proceedings of the 11th Conference on Uncertainty in Artificial Intelligence (UAI-95)’, MCGill University, Montreal, Quebec, Canada, 18–20th August, 274–284.Google Scholar
- Lacruz, B. (1998), Procedimientos secuenciales de modelización gráfica, inferencia y aprendizaje de un sistema dinámico parcialmente observado. PhD Thesis, University of Zaragoza, Spain.Google Scholar
- Lacruz, B.; Lasala, P. & Lekuona, A. (2000), ‘Dynamic graphical models and nonhomogeneous hidden Markov models’, Statistics and Probability Letters (in press).Google Scholar