Nonnegativity-enforced Gaussian process regression

Andrew Pensoneault; Xiu Yang; Xueyu Zhu

doi:10.1016/j.taml.2020.01.036

Back

Nonnegativity-enforced Gaussian process regression

Journal article

Open access

Peer reviewed

Nonnegativity-enforced Gaussian process regression

Andrew Pensoneault, Xiu Yang and Xueyu Zhu

Theoretical and applied mechanics letters, Vol.10(3), pp.182-187

03/2020

DOI: 10.1016/j.taml.2020.01.036

Files and links (1)

url

https://doi.org/10.1016/j.taml.2020.01.036View

Published (Version of record) Open Access

Abstract

•This method imposes non-negativity in Gaussian process regression in a probabilistic way.•This method significantly narrows the confident interval in the prediction. Gaussian process (GP) regression is a flexible non-parametric approach to approximate complex models. In many cases, these models correspond to processes with bounded physical properties. Standard GP regression typically results in a proxy model which is unbounded for all temporal or spacial points, and thus leaves the possibility of taking on infeasible values. We propose an approach to enforce the physical constraints in a probabilistic way under the GP regression framework. In addition, this new approach reduces the variance in the resulting GP model.

Constrained optimization

Gaussian process regression

Details

Title: Subtitle: Nonnegativity-enforced Gaussian process regression
Creators: Andrew Pensoneault - University of Iowa
Xiu Yang - Lehigh University
Xueyu Zhu - University of Iowa
Resource Type: Journal article
Publication Details: Theoretical and applied mechanics letters, Vol.10(3), pp.182-187
DOI: 10.1016/j.taml.2020.01.036
ISSN: 2095-0349
eISSN: 2095-0349
Publisher: Elsevier Ltd
Language: English
Date published: 03/2020
Academic Unit: IIHR--Hydroscience and Engineering; Mathematics
Record Identifier: 9984240867402771

Metrics

42 Record Views

29 Times Cited - Web of Science