Workshop SDS: Registration
View Submission - SDS2022
A0163
Title: Semiparametric finite mixture of regression models with Bayesian P-splines Authors:  Marco Berrettini - University of Bologna (Italy) [presenting]
Giuliano Galimberti - University of Bologna (Italy)
Saverio Ranciati - Universita di Bologna (Italy)
Abstract: Mixture models provide a useful tool to account for unobserved heterogeneity and are at the basis of many model-based clustering methods. To gain additional flexibility, some model parameters can be expressed as functions of concomitant covariates. A semiparametric finite mixture of regression models is defined, with concomitant information assumed to influence both the component weights and the conditional means. The proposed contribution replaces the linear predictors with a smooth function of the covariate considered. An estimation procedure within the Bayesian paradigm is suggested, where the smoothness of the covariate effects is controlled by suitable choices for the prior distributions of the spline coefficients. A data augmentation scheme based on different random utility models is exploited to describe the mixture weights as functions of the covariate. The performance of the proposed methodology is investigated via simulation experiments, and applications to real datasets are discussed.