Bayesian Nonparametrics (2012)

Ramses Mena Chavez

Geometric weights priors for nonparametric modeling

A random probability measure constructed via a species sampling model with geometric weights is presented for its use in Bayesian nonparametric methods. Although this model is considerably less complex than others available in the literature, it is equally robust and, in cases, more suitable for real implementations. Some properties, computational aspects and applications in regression, stochastic processes and Bayesian nonparametric mixture modeling will be discussed.