Qing Li
|
Department of Mathematics,
|
Title: Bayesian Regularized Quantile Regression
|
Abstract:
Regularization has been shown to be effective in quantile regression in
improving the prediction accuracy. This talk studies regularization in
quantile regressions from a Bayesian perspective. By proposing a hierarchical
model framework, we give a generic treatment to a set of regularization
approaches, including lasso, group lasso and elastic net penalties. Gibbs
samplers are derived for all cases. Both simulated and real data examples
show that Bayesian regularized quantile regression methods often outperform
quantile regression without regularization and their non-Bayesian
counterparts with regularization. |