This paper proposes two new penalized quantile regression methods for the longitudinal data model with multiple random effects. By applying Lasso and adaptive Lasso penalties to the quantile regression coefficients, both methods can be used to automatically select the independent variables in the model. The paper also designs the iterative algorithm for parameter estimation and discusses the optimal penalty parameter selection method. The results of Monte Carlo simulation studies show that the two new methods can not only estimate and select the quantile regression coefficient precisely, but a...
