Quasi-independence models with rational maximum likelihood estimator

Abstract We classify the two-way quasi-independence models (independence with structural zeros) that have rational maximum likelihood estimators, or MLEs. give a necessary and sufficient condition on bipartite graph associated to model for MLE be rational. In this case, we an explicit formula in terms of combinatorial features graph. also use Horn uniformization show general log-linear M MLE, any obtained by restricting face cone statistics has MLE.