“…GoM models belong to the general class of mixed membership models (Erosheva, 2002; Erosheva et al, 2004), in which individuals are allowed membership in more than one class simultaneously. Models from this family have been proposed for a wide range of applications, including the study of disability among elders (Manton et al, 1994; Erosheva et al, 2007), network analysis (Airoldi et al, 2008), electoral preferences analysis (Gormley, 2006; Gormley and Murphy, 2008), estimation of judgment accuracy (Cooil and Varki, 2003), estimation of population sizes (Manrique-Vallier and Fienberg, 2008), genetic composition analysis (Pritchard et al, 2000) and text classification (Erosheva et al, 2004; Blei et al, 2003; Blei and Lafferty, 2007). Like other latent structure models, mixed membership models offer an approach to the analysis of large, sparse contingency tables with complex interactions.…”