Structural-EM for Learning PDG Models from Incomplete Data

Abstract

Probabilistic Decision Graphs (PDGs) are a class of graphical models that can naturally encode some context specific independencies that cannot always be efficiently captured by other popular models, such as Bayesian Networks. Furthermore, inference can be carried out efficiently over a PDG, in time linear in the size of the model. The problem of learning PDGs from data has been studied in the literature, but only for the case of complete data. In this paper we propose an algorithm for learning PDGs in the presence of missing data. The proposed method is based on the EM algorithm for estimating the structure of the model as well as the parameters. We test our proposal on artificially generated data with different rates of missing cells, showing a reasonable performance.