|
Abstract:
|
There has been a growing interest among researchers in the use of Latent Class Growth mixture modeling techniques for applications in the field of developmental psychology, social and behavioral science. Latent Class Growth Mixture (LCGM) model has been increasingly recognized for its effectiveness to identify distinct subgroups of trajectories based on longitudinal data. Current LCGM methods (e.g., using Mplus and SAS Proc Traj) for Zero-inflated count data, have problems with convergence issue, non-identification, local maxima etc.. This paper focuses on developing multilevel LCGM for Zero-inflated Poisson data. The goal is to model longitudinal growth curves on counts of rare events and to classify individuals into distinct latent groups that represent the unique phenotype of each individual. Bayesian Metropolis algorithm is used to estimate the latent variables and regression parameters in this proposed model. The methodological details are illustrated through a simulation study and are applied to both individual-level data and community-level data on cigarette smoking from the National Longitudinal Study of Adolescent to Adult Health.
|
