A two-part finite mixture quantile regression model for semi-continuous longitudinal data

This paper develops a two-part finite mixture quantile regression model for semi-continuous longitudinal data. The components of the finite mixture are associatedwith homogeneous individuals in the population sharing common values of the model parameters. The proposed methodology allows heterogeneity sources that influence the first level decision process, that is, the model for the binary response variable, to influence also the distribution of the positive outcomes. Estimation is carried out through an EM algorithm without parametric assumptions on the random effects distribution. A penalized version of the EM algorithm is also presented to tackle the problem of variable selection. The suggested modelling framework has been discussed using the extensively investigated RAND Health Insurance Experiment dataset in the random intercept case