Quantile Sheets

S.K. Schnabel, P.H.C. Eilers

Research output: Chapter in Book/Report/Conference proceedingConference paperAcademicpeer-review

Abstract

Quantile regression is a popular tool for modelling the conditional quantiles of a response as a function of one (or more) independent variables. Quantile curves for different probabilities often cross. In theory this is not possible, but it is frequently encountered in small data sets. We propose to estimate all quantile curves simultaneously with a surface, a so-called quantile sheet. This sheet is fitted with a sum of tensor products of B-splines with two penalties to control for smoothness.
Original languageEnglish
Title of host publicationBook of Abstracts of the Joint Meeting of y-BIS and jSPE
EditorsP. Canas Rodrigues, F. Marques, L. Ramon
Place of PublicationLisbon
PublisherUniversidade Nova de Lisboa
Pages186-187
ISBN (Print)9789728893316
Publication statusPublished - 2012
EventJoint Meeting of the y-BIS and jSPE -
Duration: 23 Jul 201226 Jul 2012

Conference

ConferenceJoint Meeting of the y-BIS and jSPE
Period23/07/1226/07/12

Fingerprint

Dive into the research topics of 'Quantile Sheets'. Together they form a unique fingerprint.

Cite this