Bayesian LASSO-Regularized quantile regression for linear regression models with autoregressive errors

ABSTRACT

Quantile regression (QR) is a natural alternative for depicting the impact of covariates on the conditional distributions of a outcome variable instead of the mean. In this paper, we investigate Bayesian regularized QR for the linear models with autoregressive errors. LASSO-penalized type priors are forced on regression coefficients and autoregressive parameters of the model. Gibbs sampler algorithm is employed to draw the full posterior distributions of unknown parameters. Finally, the proposed procedures are illustrated by some simulation studies and applied to a real data analysis of the electricity consumption.

KEYWORDS:

• Bayesian QR
• LASSO-penalized priors
• Gibbs sampler
• Autoregressive error
• Regression model

MATHEMATICS SUBJECT CLASSIFICATION:

• 62F15
• 62J05
• 62J07

Acknowledgements

The work is supported by National Natural Science Foundation of China (No. 11501167, 11601126), China Postdoctoral Science Foundation (No. 2017M610156) and Young academic leaders project of Henan University of Science and Technology (No. 13490008), Key Program of National Philosophy and Social Science Foundation Grant (No. 13AZD064), Renmin University of China: the special developing and guiding fund for building world-class universities (disciplines) (No. 15XNL008).

Additional information

Funding

15XNL008