Inferring Block Structure of Graphical Models in Exponential Families
Siqi Sun, Hai Wang, Jinbo Xu · 2015
Learning the structure of a graphical model is a fundamental problem and it is used extensively to infer the relationship between random vari-ables. In many real world applications, we usu-ally have some prior knowledge about the under-lying graph structure, such as degree distribution and block structure. In this paper, we propose a novel generative model for describing the block structure in general exponential families, and op-timize it by an Expectation-Maximization(EM) algorithm with variational Bayes. Experimental results show that our method performs well on both synthetic and real data. Furthermore, our method can predict overlapping block structure of a graphical model in general exponential fam-ilies. 1