Multi-fidelity regression using a non-parametric relationship

Federico Zertuche 1 Celine Helbert 2 Anestis Antoniadis 3
1 MOISE - Modelling, Observations, Identification for Environmental Sciences
Inria Grenoble - Rhône-Alpes, LJK - Laboratoire Jean Kuntzmann, INPG - Institut National Polytechnique de Grenoble
2 PSPM - Probabilités, statistique, physique mathématique
ICJ - Institut Camille Jordan [Villeurbanne]
3 SAM - Statistique Apprentissage Machine
LJK - Laboratoire Jean Kuntzmann
Abstract : When the precision of the output of a heavy computer code can be tuned, it is possible to incorporate responses with different levels of fidelity to enhance the prediction of output of the most accurate simulation. This is usually done by adding several imprecise responses instead of a few precise ones. The main example for this type of computer experiments are the numerical solutions of differential equations. This problem has been studied by many authors, most notably by LeGratiet (2012) and Kennedy and O'Hagan (2001). In the present work, we propose a new approach that is different from the existing ones and based on a non-parametric relationship between two consecutive levels of fidelity.
Document type :
Conference papers
Complete list of metadatas

Cited literature [3 references]  Display  Hide  Download
Contributor : Céline Helbert <>
Submitted on : Wednesday, December 17, 2014 - 8:54:17 PM
Last modification on : Thursday, November 21, 2019 - 2:16:32 AM
Long-term archiving on: Monday, March 23, 2015 - 4:08:21 PM


Files produced by the author(s)


  • HAL Id : hal-01096661, version 1


Federico Zertuche, Celine Helbert, Anestis Antoniadis. Multi-fidelity regression using a non-parametric relationship. MASCOT 2014 - Méthodes d'Analyse Statistique pour les COdes et Traitements numériques, Apr 2014, Zurich, Switzerland. ⟨hal-01096661⟩



Record views


Files downloads