Please read our student and staff community guidance on COVID-19
Skip to main content Skip to navigation

Paper No. 06-13

Download 06-13

JE Griffin

On the Bayesian analysis of species sampling mixture models for density estimation

Abstract: The mixture of normals model has been extensively applied to density estimation problems. This paper proposes an alternative parameterisation that naturally leads to new forms of prior distribution. The parameters can be interpreted as the location, scale and smoothness and default choices lead to automatic Bayesian density estimation. The ideas are extended to of the density. Priors on these parameters are often easier to specify. Alternatively, improper
multivariate density estimation.

Keywords: Density Estimation, Species sampling models, Dirichlet process mixture models, Mixtures of normals.