Fig. 5.

Fabric Fault Data; posterior distributions $\text{Pr}(K_{+}|\mathbf{y})$ of the number of clusters $K_{+}$ for mixture of Possion GLMs (left-hand side) as well as mixtures of negative binomial GLMs (right-hand side); top: based on sparse finite mixtures (SFM), bottom: based on Dirichlet process mixtures (DPM) under various hyper priors