Skip to main content

Table 3 Models combining block models and topic models

From: A review of stochastic block models and extensions for graph clustering

ArticleType of graphTextualInferenceClusteringNumber of groups/topicsLongitudinal
Zhou et al. (2006)ValueddirectedEdgesMCMCSoftFixed (6 groups; 20 topics)No
McCallum et al. (2007)ValueddirectedEdgesMCMCSoftCriterion (Perplexity)No
Pathak et al. (2008)ValueddirectedEdgesMCMCSoftFixed (8 groups; 25 topics)No
Liu et al. (2009)BinarydirectedNodesVariationalSoftCriterion (Perplexity)No
Chang and Blei (2010)BinaryundirectedNodesVariationalSoftFixed (5,10,15,20,25 topics)No
Ho et al. (2012)BinarydirectedNodesMCMCSoftDirichlet processNo
Sachan et al. (2012)ValueddirectedEdgesMCMCSoft; SoftCriterion (Perplexity)No
Bouveyron et al. (2016)ValueddirectedEdgesVariationalSoft; HardCriterion (BIC-like)No
Corneli et al. (2018)ValueddirectedEdgesVariationalSoft; HardCriterion (ICL)Yes
  1. Under the column “clustering”, the first approach refers to the topics. If there is a second approach, it refers to the groups