Gaussian process quantile regression using expectation propagation

Alexios Boukouvalas, Remi Barillec, Dan Cornford

Research output: Unpublished contribution to conferenceUnpublished Conference Paperpeer-review

Abstract

Direct quantile regression involves estimating a given quantile of a response variable as a function of input variables. We present a new framework for direct quantile regression where a Gaussian process model is learned, minimising the expected tilted loss function. The integration required in learning is not analytically tractable so to speed up the learning we employ the Expectation Propagation algorithm. We describe how this work relates to other quantile regression methods and apply the method on both synthetic and real data sets. The method is shown to be competitive with state of the art methods whilst allowing for the leverage of the full Gaussian process probabilistic framework.
Original languageEnglish
Publication statusPublished - 27 Jun 2012
Event29th International Conference on Machine Learning - Ediburgh, United Kingdom
Duration: 26 Jun 20121 Jul 2012

Conference

Conference29th International Conference on Machine Learning
Abbreviated titleICML 2012
Country/TerritoryUnited Kingdom
CityEdiburgh
Period26/06/121/07/12

Keywords

  • Gaussian process
  • probablisitc modelling
  • decision theory

Fingerprint

Dive into the research topics of 'Gaussian process quantile regression using expectation propagation'. Together they form a unique fingerprint.

Cite this