Rational partition models under iterative proportional scaling

Abstract

In this work we investigate partition models, the subset of log-linear models for which one can perform the iterative proportional scaling (IPS) algorithm to numerically compute the maximum likelihood estimate (MLE). Partition models include families of models such as hierarchical models and balanced, stratified staged trees. We define a sufficient condition, called the Generalized Running Intersection Property (GRIP), on the matrix representation of a partition model under which IPS algorithm produces the exact MLE in one cycle. Additionally we connect the GRIP to the toric fiber product and to previous results for hierarchical models and balanced, stratified staged trees. This leads to a characterization of balanced, stratified staged trees in terms of the GRIP.

arXiv comment41 pages, 8 figures

References in corpus (10)

• On the toric algebra of graphical models
• The maximum likelihood degree of a very affine variety
• The Maximum Likelihood Degree of Toric Varieties
• Multigraded Commutative Algebra of Graph Decompositions
• Equations defining probability tree models
• Maximum Likelihood Estimation of Toric Fano Varieties
• Gröbner bases for staged trees
• Staged tree models with toric structure
• Toric invariant theory for maximum likelihood estimation in log-linear models
• Iterative proportional scaling via decomposable submodels for contingency tables

Cited by in corpus (1)

• Geometry of rational quasi-independence models as toric fiber products