On the toric algebra of graphical models
Geiger, Dan ; Meek, Christopher ; Sturmfels, Bernd
Ann. Statist., Tome 34 (2006) no. 1, p. 1463-1492 / Harvested from Project Euclid
We formulate necessary and sufficient conditions for an arbitrary discrete probability distribution to factor according to an undirected graphical model, or a log-linear model, or other more general exponential models. For decomposable graphical models these conditions are equivalent to a set of conditional independence statements similar to the Hammersley–Clifford theorem; however, we show that for nondecomposable graphical models they are not. We also show that nondecomposable models can have nonrational maximum likelihood estimates. These results are used to give several novel characterizations of decomposable graphical models.
Publié le : 2006-06-14
Classification:  Conditional independence,  factorization,  graphical models,  decomposable models,  factorization of discrete distributions,  Hammersley–Clifford theorem,  Gröbner bases,  60E05,  62H99,  13P10,  14M25,  68W30
@article{1152540755,
     author = {Geiger, Dan and Meek, Christopher and Sturmfels, Bernd},
     title = {On the toric algebra of graphical models},
     journal = {Ann. Statist.},
     volume = {34},
     number = {1},
     year = {2006},
     pages = { 1463-1492},
     language = {en},
     url = {http://dml.mathdoc.fr/item/1152540755}
}
Geiger, Dan; Meek, Christopher; Sturmfels, Bernd. On the toric algebra of graphical models. Ann. Statist., Tome 34 (2006) no. 1, pp.  1463-1492. http://gdmltest.u-ga.fr/item/1152540755/