LONGITUDINAL DATA ANALYSIS USING GENERALIZED MAXIMUM ENTROPY APPROACH

Marginal generalized linear models are frequently used for the analysis of repeated measurements and longitudinal data. During the last three decades, researchers used parametric, nonparametric as well as Bayesian methods as useful approaches to model such kind of data. The correlation among the repeated measurements is considered a vital factor to increase the estimation efficiency of the model’s parameters for different correlation structures. This article suggests using the generalized maximum entropy (GME) as an efficient method for the joint modelling of mean and correlation parameters that permits the estimation with minimum distributional assumptions. Moreover, we present a simulation study to compare the performance of the GME method with a set of well known estimation methods in the longitudinal data literatures.