The Library

Asymptotic model selection and identifiability of directed tree models with hidden variables

Tools

Zwiernik, Piotr (2010) Asymptotic model selection and identifiability of directed tree models with hidden variables. Working Paper. Coventry: University of Warwick. Centre for Research in Statistical Methodology. Working papers, Vol.2010 (No.6).

Preview

PDF
WRAP_Zwiernik_10-06w.pdf - Published Version - Requires a PDF viewer such as GSview, Xpdf or Adobe Acrobat Reader
Download (599Kb)

Official URL: http://www2.warwick.ac.uk/fac/sci/statistics/crism...

Abstract

The standard Bayesian Information Criterion (BIC) is derived under
some regularity conditions which are not always satisfied by the graphical
models with hidden variables. In this paper we derive the BIC score for
Bayesian networks in the case when the data is binary and the underlying
graph is a rooted tree and all the inner nodes represent hidden variables. This
provides a direct generalization of a similar formula given by Rusakov and
Geiger in [10]. Geometric results obtained in this paper are complementary to
the results in the previous paper [18] extending our understanding of this class
of models. The main tool used in this paper is the connection between asymptotic
approximation of Laplace integrals and the real log-canonical threshold.

Item Type:

Working or Discussion Paper (Working Paper)

Subjects:

Q Science > QA Mathematics

Divisions:

Faculty of Science > Statistics

Library of Congress Subject Headings (LCSH):

Bayesian statistical decision theory, Trees (Graph theory)

Series Name:

Working papers

Publisher:

University of Warwick. Centre for Research in Statistical Methodology

Place of Publication:

Coventry

Official Date:

2010

Dates:

Date	Event
2010	Published

Volume:

Vol.2010

Number:

No.6

Number of Pages:

Status:

Not Peer Reviewed

Access rights to Published version:

Open Access

References:

[1] V. Arnold, S. Gusein-Zade, and A. Varchenko, Singularities of differentiable maps,
vol. II, Birkhauser, 1988.
[2] E. Gawrilow and M. Joswig, Geometric reasoning with polymake. arXiv:math.CO/0507273,
2005.
[3] D. Geiger, D. Heckerman, H. King, and C. Meek, Stratified exponential families: graph-
ical models and model selection, Ann. Statist., 29 (2001), pp. 505{529.
[4] D. Haughton, On the choice of a model to fit data from an exponential family, Ann. Statist,
16 (1988), pp. 342{355.
[5] J. Kollar, Singularities of pairs, in Proceedings of Symposia in Pure Mathematics, vol. 62,
1997, pp. 221{288.
[6] R. Lazarsfeld, Positivity in algebraic geometry, A Series of Modern Surveys in Mathematics,
Springer Verlag, 2004.
[7] S. Lin, Asymptotic approximation of marginal likelihood integrals. in preparation, 2010.
[8] R. Mihaescu and L. Pachter, Combinatorics of least-squares trees, Proceedings of the
National Academy of Sciences of the United States of America, 105 (2008), p. 13206.
[9] V. Moulton and M. Steel, Peeling phylogenetic 'oranges', Advances in Applied Mathematics,
33 (2004), pp. 710{727.
[10] D. Rusakov and D. Geiger, Asymptotic model selection for naive Bayesian networks, J.
Mach. Learn. Res., 6 (2005), pp. 1{35 (electronic).
[11] M. Saito, On real log canonical thresholds, Arxiv preprint arXiv:0707.2308 v, 1 (2007).
[12] G. Schwarz, Estimating the dimension of a model, Annals of Statistics, 6 (1978), pp. 461{
464.
[13] C. Semple and M. Steel, Phylogenetics, vol. 24 of Oxford Lecture Series in Mathematics
and its Applications, Oxford University Press, Oxford, 2003.
[14] R. P. Stanley, Enumerative combinatorics. Volume I, no. 49 in Cambridge Studies in Advanced
Mathematics, Cambridge University Press, 2002.
[15] S. Watanabe, Algebraic Analysis for Nonidentifiable Learning Machines, Neural Computation,
13 (2001), pp. 899{933.
[16] S. Watanabe, Algebraic Geometry and Statistical Learning Theory, no. 25 in Cambridge
Monographs on Applied and Computational Mathematics, Cambridge University Press, 2009.
ISBN-13: 9780521864671.
[17] K. Yamazaki and S. Watanabe, Newton diagram and stochastic complexity in mixture of
binomial distributions, in Algorithmic learning theory, vol. 3244 of Lecture Notes in Comput.
Sci., Springer, Berlin, 2004, pp. 350{364.
[18] P. Zwiernik and J. Q. Smith, The geometry of conditional independence tree models with
hidden variables. Arxiv preprint arXiv:0904.1980, January 2010.