On a Loss-based prior for the number of components in mixture models

Clara Grazian, Cristiano Villa, Brunero Liseo

We propose a prior distribution for the number of components of a finite mixture model. The novelty is that the prior distribution is obtained by considering the loss one would inc…