Bayesian multivariate quantile regression using Dependent Dirichlet Process prior

In this article, we consider a non-parametric Bayesian approach to multivariate quantile regression. The proposed involves modeling of related conditional distributions response vector given the covariates using Dependent Dirichlet Process (DDP) prior. DDP is used introduce dependence across covariates. flexible covariate-dependent mixture Gaussian kernels gives rise an induced posterior for desired quantile. For computations, use truncated stick-breaking representation DDP, and block Gibbs sampler estimating model parameters. We illustrate our method with simulation studies, data containing blood pressures 40 women. Finally, provide theoretical justification through consistency support properties