An automated MCEM algorithm for hierarchical models with multivariate and multitype response variables.

Vera Georgescu (1) , Nicolas Desassis (2) , Samuel S. Soubeyrand (1) , André Kretzschmar (1) , Rachid Senoussi (1)

1 BioSP - Biostatistique et Processus Spatiaux
2 GEOSCIENCES - Centre de Géosciences

In this paper, we consider a model allowing the analysis of multivariate data, which can contain data attributes of different types (e.g. continuous, discrete, binary). This model is a two-level hierarchical model which supports a wide range of correlation structures and can accommodate overdispersed data. Maximum likelihood estimation of the model parameters is achieved with an automated Monte Carlo Expectation Maximization (MCEM) algorithm. Our method is tested in a simulation study in the bivariate case and applied to a dataset dealing with beehive activity.

Mots clés: Count data; Mixed mode data; Monte Carlo EM; Overdispersion

Mots clés: Continuous data; Poisson-log normal distribution

Domaines: Mathématiques [math]; Statistiques [math.ST]

Vera Georgescu, Nicolas Desassis, Samuel S. Soubeyrand, André Kretzschmar, Rachid Senoussi. An automated MCEM algorithm for hierarchical models with multivariate and multitype response variables.. Communications in Statistics - Theory and Methods, 2014, 43 (17), pp.3698-3719. ⟨10.1080/03610926.2012.700372⟩. ⟨hal-01115524⟩