Mixed effect modelling and variable selection for quantile regression

It is known that the estimating equations for quantile regression (QR) can be solved using an EM algorithm in which M-step computed via weighted least squares, with weights at E-step as expectation of independent generalized inverse-Gaussian variables. This fact exploited here to extend QR allow random effects linear predictor. Convergence this setting established by showing it a alternating minimization (GAM) procedure. Another modification also allows us adapt recently proposed method variable selection mean models setting. Simulations show resulting significantly outperforms lasso penalty. Applications real data include frailty analysis hospital stays, and age onset lung cancer riboflavin production rate high-dimensional gene expression arrays prediction.