B0572
Title: Multivariate species sampling models
Authors: Beatrice Franzolini - Bocconi University (Italy)
Antonio Lijoi - Bocconi University (Italy)
Igor Pruenster - Bocconi University (Italy)
Giovanni Rebaudo - University of Turin and Collegio Carlo Alberto (Italy)
Beatrice Franzolini - Bocconi University (Italy) [presenting]
Abstract: Species sampling models provide a general framework for random discrete distributions and are tailored for exchangeable data. However, they fall short when used to model heterogeneous data collected from related sources or distinct experimental conditions. To address this limitation, partial exchangeability serves as the ideal probabilistic invariance condition. While numerous models exist for partially exchangeable observations, a unifying framework, similar to species sampling models, is currently absent. Multivariate species sampling models are introduced, which are a general class of models characterized by their partially exchangeable partition probability function. These models encompass existing nonparametric models for partial exchangeable data, thereby highlighting their core distributional properties and induced learning mechanisms. The results enable an in-depth comprehension of the induced dependence structure as well as facilitate the development of new models.