Mixture Data-Dependent Priors

Leonardo Egidi, Francesco Pauli, Nicola Torelli

We propose a two-component mixture of a noninformative (diffuse) and an informative prior distribution, weighted through the data in such a way to prefer the first component if a p…