Penalizing function based bandwidth choice in nonparametric quantile regression

In nonparametric mean regression various methods for bandwidth choice exist. These methods can roughly be divided into plug-in methods and methods based on penalizing functions. This paper uses the approach based on penalizing functions and adapt it to nonparametric quantile regression estimation, where bandwidth choice is still an unsolved problem. Various criteria for bandwitdth choice are defined and compared in some simulation examples.

MoreLess

Year of publication:	2001
Authors:	Abberger, Klaus
Publisher:	Konstanz : University of Konstanz, Center of Finance and Econometrics (CoFE)
Subject:	Regression \| Nichtparametrisches Verfahren \| Theorie \| nonparametric quantile regression \| bandwidth choice \| crossvalidation \| penalizing functions

More details

Series:	CoFE Discussion Paper ; 01/10
Type of publication:	Book / Working Paper
Type of publication (narrower categories):	Working Paper
Language:	English
Other identifiers:	870179217 [GVK] hdl:10419/85240 [Handle] RePEc:zbw:cofedp:0110 [RePEc]
Source:	EconStor

Persistent link: https://www.econbiz.de/10010324105