Gildas Mazo (MaIAGE - INRA), 9 mars 2018

A constrained kernel density estimator for location-scale mixture models based on copulas
vendredi 9 mars 2018

In this communication we shall present copula-based semiparametric mixture models as a way to model heterogeneous populations. Copulas can cope with complex dependence structures while the nonparametric estimation of the marginals alleviate one’s effort in the modeling task. Estimation is performed by two EM-like algorithms and one of them will be shown to perform better by taking into account the inherent structure of the problem into account.