A0444
Title: A general Bayesian approach to multiple-output quantile regression
Authors: Annika Camehl - Erasmus University Rotterdam (Netherlands) [presenting]
Dennis Fok - Erasmus University Rotterdam (The Netherlands)
Kathrin Gruber - Erasmus University Rotterdam (Netherlands)
Abstract: A general Bayesian approach is proposed to multiple-output quantile regression. We place a prior on the collection of regression functions and consider the joint distribution of the output-vector given the covariates to be a finite mixture of multivariate Gaussians. The resulting model provides an extremely flexible framework to approximate various symmetric and asymmetric distributions. We show how to construct conditional quantiles for the marginal effects within a unified model. It can handle correlation across disturbance terms, guarantees non-crossing quantiles and yields easily interpretable results.
