|
Abstract:
|
In the last few years, the analysis of non-exchangeable data has attracted considerable interest in the Bayesian nonparametric literature. The talk will discuss a model based on dependent discrete random probability measures that may be used, among others, to deal with multiple-sample data. Predictive distributions and posterior characterizations will be presented, along with algorithms that allow to determine approximate relevant Bayesian inferences. The discussion will be completed by illustrations with simulated and real data.
|
