Joint modeling of dependent diverse format data: A review from several case-studies

Overview

Abstract

This talk reviews some fundamental and practical issues related to the formulation and analysis of joint models of mixed types of outcomes with latent variables, with particular emphasis on both several case-studies in applied statistics and their computational implementation. For instance, longitudinal CD4-count and survival HIV/AIDS data study has aroused much interest in recent years, where biomarker repeated measures and time to event outcomes had been analyzed separately, in particular for dependent diverse format data. However, considering that such data are observed in the same experimental unit, the joint modeling of the associated response variables seems to be more appropriate when compared to the separate modeling. In that case, latent processes can be employed to link the response sub-models and data analysis can also be done from a Bayesian perspective. 

Brief Biography

Giovani Silva got his PhD in Mathematics (IST) at the University of Lisbon. He was also a student at Federal University of Ceara and University of Sao Paulo, and completed a post-doctoral internship in Statistics at Simon Fraser University. He is Assistant Professor at the Department of Mathematics at Instituto Superior Técnico (IST) and member of the Center for Statistics and Applications (CEAUL), both at the University of Lisbon. He is the co-author of three books on Generalized Linear Models, Spatiotemporal Models for Multistate Rates, Proportions and Processes, and Bayesian Statistics. He is currently the co-editor of REVSTAT - Statistical Journal and associate editor of the Brazilian Journal of Probability and Statistics and Chilean Journal of Statistics.