M. Studeny, D. Haws, R. Hemmecke, S. Lindner:
Polyhedral approach to statistical learning graphical
models. In Harmony of Groebner Bases and the
Modern Industrial Society: the 2nd CREST--SBM
International Conference (T. Hibi ed.),
World Scientific, Singapore 2012, pages 346-372.
The statistical task to learn graphical models of Bayesian network structure
from data leads to the study of special polyhedra. In the paper, we
offer an overview of our polyhedral approach to learning these
statistical models. First, we report on the results on this topic
from our recent papers. The second part of the paper brings some
specific additional results inspired by this approach.
- AMS classification 68T30
- Bayesian network structure
- standard imset
- characteristic imset
- polyhedral geometry
pdf version (306kB) is available.