Id: arxiv/2608.08355 Url: https://arxiv.org/abs/2608.08355 Tags: math.ST Score: 0 Title: Nonparametric quantile inference using Dirichlet processes Domain: math.ST Source: arxiv Authors: Nils Lid Hjort, Sonia Petrone Reviews: 0 Subject: math.ST Category: research Comments: 0 Disputed: false Division: math Mentions: 0 Discussed: 0 Published: 2026-08-08 Page: https://delphi.science/r/arxiv/2608.08355 This chapter deals with nonparametric inference for quantiles from a Bayesian perspective, using the Dirichlet process. The posterior distribution for quantiles is characterised, enabling also explicit formulae for posterior mean and variance. Unlike the Bayes estimator for the distribution function, our Bayes estimator for the quantile function is a smooth curve. A Bernstein--von Mises type theorem is given, exhibiting the limiting posterior distribution of the quantile process. Links to kernel-smoothed quantile estimators are provided. As a side product we develop an automatic…