A0325
Title: Quantile regression inference processes and choices in sparsity estimation
Authors: Thomas Parker - University of Waterloo (Canada) [presenting]
Abstract: The aim is to investigate uniform inference for conditional quantile functions, in the spirit of a prior study. The focus is on estimation of the derivative of the conditional quantile process, or the sparsity function, beyond the choices proposed by prior studies. Recent technical advances have proposed methods to establish the uniform consistency of density estimates that can be adapted to this setting, allowing researchers to use a wide variety of sparsity estimators for inference. A small simulation experiment compares the finite sample performance of a few such estimators with the well-established sparsity estimates.
