Improvement of Adaptive Lasso in Binary Quantile Regression

Abstract

In order to avoid the over-fitting of the model, the adaptive LASSO method was used to the variables selection of the binary quantile regression model. Bayesian method is use to construct the Gibbs sampling algorithm and the constraint condition that does not affect the predictive result is used to improve the stability of the sampling value. That the improved model has better parameter estimation efficiency and variable selection effect and classification ability are illustrated in the numerical simulation.