The statistical literature discusses different types of Markov properties for chain graphs that lead to four possible classes of chain graph Markov models. The different models are rather well understood when the observations are continuous and multivariate normal, and it is also known that one model class, referred to as models of LWF (Lauritzen–Wermuth–Frydenberg) or block concentration type, yields discrete models for categorical data that are smooth. This paper considers the structural properties of the discrete models based on the three alternative Markov properties. It is shown by example that two of the alternative Markov properties can lead to non-smooth models. The remaining model class, which can be viewed as a discrete version of multivariate regressions, is proven to comprise only smooth models. The proof employs a simple change of coordinates that also reveals that the model’s likelihood function is unimodal if the chain components of the graph are complete sets.
References
[1] Andersson, S.A., Madigan, D. and Perlman, M.D. (2001). Alternative Markov properties for chain graphs. Scand. J. Statist. 28 33–85.
[2] Bergsma, W.P. and Rudas, T. (2002). Marginal models for categorical data. Ann. Statist. 30 140–159.
[3] Cox, D., Little, J. and O’Shea, D. (1997). Ideals, Varieties, and Algorithms, 2nd ed. New York: Springer.
[4] Cox, D.R. and Wermuth, N. (1996). Multivariate Dependencies. London: Chapman & Hall.
[5] Drton, M. (2008). Iterative conditional fitting for discrete chain graph models. In COMPSTAT 2008 – Proceedings in Computational Statistics 93–104. Heidelberg: Physica.
[6] Drton, M. (2009). Likelihood ratio tests and singularities. Ann. Statist. 37 979–1012.
[7] Drton, M. and Perlman, M.D. (2008). A SINful approach to Gaussian graphical model selection. J. Statist. Plann. Inference 138 1179–1200.
[8] Drton, M. and Richardson, T.S. (2008). Binary models for marginal independence. J. R. Stat. Soc. Ser. B Stat. Methodol. 70 287–309.
[9] Drton, M. and Sullivant, S. (2007). Algebraic statistical models. Statist. Sinica 17 1273–1297.
[10] Ekholm, A., Jokinen, J., McDonald, J.W. and Smith, P.W.F. (2003). Joint regression and association modeling of longitudinal ordinal data. Biometrics 59 795–803.
[11] Fitzmaurice, G.M., Laird, N.M. and Rotnitzky, A.G. (1993). Regression models for discrete longitudinal responses. Statist. Sci. 8 284–309.
[12] Frydenberg, M. (1990). The chain graph Markov property. Scand. J. Statist. 17 333–353.
[13] Geiger, D., Heckerman, D., King, H. and Meek, C. (2001). Stratified exponential families: Graphical models and model selection. Ann. Statist. 29 505–529.
[14] Goodman, L.A. (1973). The analysis of multidimensional contingency tables when some variables are posterior to others: A modified path analysis approach. Biometrika 60 179–192.
[15] Greuel, G.-M., Pfister, G. and Schönemann, H. (2005). Singular 3.0. A computer algebra system for polynomial computations. Centre for Computer Algebra, University of Kaiserslautern. Available at http://www.singular.uni-kl.de.
[16] Jokinen, J. (2006). Fast estimation algorithm for likelihood-based analysis of repeated categorical responses. Comput. Statist. Data Anal. 51 1509–1522.
[17] Kass, R.E. and Vos, P.W. (1997). Geometrical Foundations of Asymptotic Inference. New York: Wiley.
[18] Kauermann, G. (1996). On a dualization of graphical Gaussian models. Scand. J. Statist. 23 105–116.
[19] Kauermann, G. (1997). A note on multivariate logistic models for contingency tables. Austral. J. Statist. 39 261–276.
[20] Koster, J.T.A. (1999). On the validity of the Markov interpretation of path diagrams of Gaussian structural equation systems with correlated errors. Scand. J. Statist. 26 413–431.
[21] Lauritzen, S.L. (1996). Graphical Models. New York: Oxford Univ. Press.
[22] Lauritzen, S.L. and Wermuth, N. (1989). Graphical models for associations between variables, some of which are qualitative and some quantitative. Ann. Statist. 17 31–57.
[23] Lupparelli, M., Marchetti, G.M. and Bergsma, W.P. (2008). Parameterizations and fitting of bi-directed graph models to categorical data. Available at arXiv:0801.1440.
[24] Pachter, L. and Sturmfels, B., eds. (2005). Algebraic Statistics for Computational Biology. New York: Cambridge Univ. Press.
[25] Richardson, T.S. (2003). Markov properties for acyclic directed mixed graphs. Scand. J. Statist. 30 145–157.
[26] Richardson, T.S. and Spirtes, P. (2002). Ancestral graph Markov models. Ann. Statist. 30 962–1030.
[27] Wermuth, N. and Cox, D.R. (2004). Joint response graphs and separation induced by triangular systems. J. R. Stat. Soc. Ser. B Stat. Methodol. 66 687–717.
