A1587
Title: Analyze of count longitudinal data with random effects using R packages, cold, lme4 and glmmML
Authors: Maria Helena Goncalves - FCiencias.ID, Associacao para a Investigacao e Desenvolvimento de Ciencias (Portugal) (Portugal) [presenting]
Maria Salome Cabral - CEAUL and FCUL (Portugal)
Abstract: Longitudinal count data are commonly encountered in both experimental and observational studies across all disciplines. In these studies, repeated measurements are made on the same subject across occasions in one or more treatment groups, and correlation is usually present among response variables for a given subject. The generalized linear mixed models (GLMMs) account for that correlation by the inclusion of random effects in the linear predictor. However, in GLMMs it is assumed that the observations of the same subject are independent conditional to the random effects and covariates, which may be not true. For fitting GLMMs, R has available the packages lme4 and glmmML, at the least. The methodology implemented in cold R package inference is based on the likelihood approach, serial dependence is assumed to be of Markovian type, and it is considered as the basic stochastic mechanism. The serial dependence AR1 incorporated in the random effects model in cold allows dependence between repeated measures in terms of numerical analysis, which is ignored in the traditional approach (GLMM) implemented in the lme4 and glmmML. The R packages lme4 and glmmML only allow an independent structure, and glmmML only allows a random effect in the intercept. A real dataset is used to compare the aforementioned R packages.
