hetGP: Heteroskedastic Gaussian Process Modeling and Sequential Design in R

Simple item page
dc.contributor.authorBinois, Mickaelen
dc.contributor.authorGramacy, Robert B.en
dc.date.accessioned2022-04-14T16:57:29Zen
dc.date.available2022-04-14T16:57:29Zen
dc.date.issued2021-07en
dc.description.abstractAn increasing number of time-consuming simulators exhibit a complex noise structure that depends on the inputs. For conducting studies with limited budgets of evaluations, new surrogate methods are required in order to simultaneously model the mean and variance fields. To this end, we present the hetGP package, implementing many recent advances in Gaussian process modeling with input-dependent noise. First, we describe a simple, yet efficient, joint modeling framework that relies on replication for both speed and accuracy. Then we tackle the issue of data acquisition leveraging replication and exploration in a sequential manner for various goals, such as for obtaining a globally accurate model, for optimization, or for contour finding. Reproducible illustrations are provided throughout.en
dc.description.notesThe work of MB is supported by the U.S. Department of Energy, Office of Science, Office of Advanced Scientific Computing Research, under Contract No. DE-AC02-06CH11357. RBG gratefully acknowledges funding from a DOE LAB 17-1697 via subaward from Argonne National Laboratory for SciDAC/DOE Office of Science ASCR and High Energy Physics, and partial support from National Science Foundation grants DMS-1849794, DMS-1821258 and DMS-1621746. Many thanks to D. Austin Cole for comments on early drafts and to Gail Pieper for her useful language editing. Finally, we thank the reviewing team for the helpful suggestions and comments.en
dc.description.sponsorshipU.S. Department of Energy, Office of Science, Office of Advanced Scientific Computing ResearchUnited States Department of Energy (DOE) [DE-AC02-06CH11357]; DOE via Argonne National LaboratoryUnited States Department of Energy (DOE) [LAB 17-1697]; National Science FoundationNational Science Foundation (NSF) [DMS-1849794, DMS-1821258, DMS-1621746]en
dc.description.versionPublished versionen
dc.format.mimetypeapplication/pdfen
dc.identifier.doihttps://doi.org/10.18637/jss.v098.i13en
dc.identifier.issn1548-7660en
dc.identifier.issue13en
dc.identifier.urihttp://hdl.handle.net/10919/109665en
dc.identifier.volume98en
dc.language.isoenen
dc.rightsCreative Commons Attribution 4.0 Internationalen
dc.rights.urihttp://creativecommons.org/licenses/by/4.0/en
dc.subjectinput-dependent noiseen
dc.subjectlevel-set estimationen
dc.subjectoptimizationen
dc.subjectreplicationen
dc.subjectstochastic krigingen
dc.titlehetGP: Heteroskedastic Gaussian Process Modeling and Sequential Design in Ren
dc.title.serialJournal of Statistical Softwareen
dc.typeArticle - Refereeden
dc.type.dcmitypeTexten

Files

Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
v98i13.pdf
Size:
1.25 MB
Format:
Adobe Portable Document Format
Description:
Published version
Download

Collections

Scholarly Works, Statistics