The Dirichlet distribution I chose is merely the uniform distribution over the unit simplex. I.e., it means that all possible h-vectors get equal weight. It's

The Dirichlet distribution I chose is merely the uniform distribution over the unit simplex. I.e., it means that all possible h-vectors get equal weight. It's an attempt to choose as uninformative a prior on h as possible.