A Bayesian Allocation Model Based Approach to Mixed Membership Stochastic Blockmodels

Although detecting communities in networks has attracted considerable recent attention, estimating the number of is still an open problem. In this paper, we propose a model, which replicates generative process mixed-membership stochastic block model (MMSB) within generic allocation framework Bayesian (BAM) and BAM-MMSB. contrast to traditional blockmodels, BAM-MMSB considers observations as Poisson counts generated by base marks according MMSB. Moreover, optimal for estimated computing variational approximations marginal likelihood each order. Experiments on synthetic real data sets show that proposed approach promises generalized selection solution can choose not only size but also most appropriate decomposition.