Topic: Contributed on Track: Bayesian semi- and nonparametric modelling Title: Compound random measures Authors:  Fabrizio Leisen - University of Kent (United Kingdom) [presenting]
Jim Griffin - University of Kent (United Kingdom)
Abstract: Many nonparametric priors have been proposed for related distributions and have found a wide range of applications in statistics and machine learning. We describe a new class of dependent random measures which we call compound random measures. These priors can be constructed with gamma, stable and generalized gamma process marginals and their dependence can be characterized using both the Levy copula and correlation function. Normalized version of these random measures can be used as dependent priors for related distributions and inference can be made using a slice sampling algorithm or a Polya Urn approach.