Mixtures provide a useful approach for relaxing parametric assumptions. Discrete mixture models induce clusters, typically with the same cluster allocation for each parameter in multivariate cases. As a more flexible approach that facilitates sparse nonparamet... ...
