A0985
Title: Bayesian nonparametric inference by means of stick-breaking priors with dependent lenght variables
Authors: Ramses Mena - Universidad Nacional Autonoma De Mexico (Mexico)
Pierpaolo De Blasi - University of Torino and Collegio Carlo Alberto (Italy)
Maria Fernanda Gil-Leyva Villa - IIMAS,UNAM (Mexico) [presenting]
Abstract: The general classes of exchangeable stick-breaking processes (ESB) and Markov stick-breaking processes (MSB) are studied. In particular, the aim is to show how other well-known random probability measures, such as Dirichlet, Geometric and Pitman-Yor processes, can be recovered through ESBs and MSBs. The objective is to explain how to implement mixture models with an ESB or an MSB mixing priors by means of a novel Gibbs sampler method, and the performance of the models in clustering and density estimation is evaluated.