Bayesian Adaptive Lasso Quantile Regression And Applications Of EVA Analysis

The classical least squares method focuses on the function that describes how the mean of the response changes with the vector of covariates.However,quantile regression is a nonparametric statistical method and considers a much more detailed insight in the effects of the covariates on the different quantiles of the response distribution.Compared with the frequentist approach,bayesian approach is robust and make up the default of the classical least squares method to a certain degree.In this paper,we propose adaptive Lasso quantile regression method from a Bayesian perspective and apply it in the empirical research for influence factors of EVA in the listed companies.The first chapter is an introduction that points out the significance of research and conducted an extensive literature review of bayesian adaptive Lasso quantile regression.Finally,the paper summarizes works and innovation.Second,this chapter describe basic idea and principle of bayesian adaptive Lasso quantile regression and derive the full conditional posterior distribution of bayesian adaptive Lasso quantile regression based on the skewed Laplace distribution assumption of residual term and derive its corresponding Gibbs sampler.Third,in this section,we carry out Monte Carlo simulations to study the performance of bayesian adaptive Lasso quantile regression with comparison to bayesian Lasso quantile regression and quantile regression.Furthermore,we consider different choices for the distribution of residual term and different sample size to study properties of bayesian adaptive Lasso quantile regression.The simulations show that bayesian adaptive Lasso quantile regression based on the skewed Laplace distribution assumption of residual term have better effect.Fourth,this chapter builds the bayesian adaptive Lasso quantile regression model and apply it in the empirical research for influence factors of EVA in the listed companies of china.And the result indicates that EVA of the listed companies is influenced by following factors including Net asset yield,Total asset yield Fixed assets ratio,Inventory turnover,Long-term capital debt ratio,Number of times interest earned.Adaptive Lasso variable selection method eliminate several unnecessary variables.And some recommendations were made based on the results from this study.Finally,this part summarizes study's conclusions and puts forward the research prospects in the field considering the limitation of bayesian adaptive Lasso quantile regression.