New Bayesian Regularized for quantile regression

Abstract

The statistical analysis provided by quantile regression of the links between random variables is more in-depth. The economic field has made considerable use of quantile regression techniques to study the factors that affect wages, the results of discrimination, and the development of income inequality. Using the Bayesian lasso penalty technique, this article estimates and selects variables in quantile regression models. The Laplace prior distribution of the vector of parameters will be represented by a scale mixture of normal distributions mixing Rayleigh density, and a Bayesian hierarchical model will be constructed to estimate its parameters. Simulation examples and real-world data are taken into account to assess the suggested method's effectiveness and to compare it to other current approaches.