Computation of sobol indices in global sensitivity analysis from small data sets by probabilistic learning on manifolds
peer reviewed Global sensitivity analysis provides insight into how sources of uncertainty contribute to uncertainty in predictions of computational models. Global sensitivity indices, also called variance-based sensitivity indices and Sobol indices, are most often computed with Monte Carlo methods....
Published in: | International Journal for Uncertainty Quantification |
---|---|
Main Authors: | , , |
Format: | Article in Journal/Newspaper |
Language: | English |
Published: |
Begell House Inc.
2021
|
Subjects: | |
Online Access: | https://orbi.uliege.be/handle/2268/295234 https://doi.org/10.1615/Int.J.UncertaintyQuantification.2020032674 |
id |
ftorbi:oai:orbi.ulg.ac.be:2268/295234 |
---|---|
record_format |
openpolar |
spelling |
ftorbi:oai:orbi.ulg.ac.be:2268/295234 2024-04-21T07:49:02+00:00 Computation of sobol indices in global sensitivity analysis from small data sets by probabilistic learning on manifolds Arnst, Maarten Soize, Christian Bulthuis, Kevin 2021 https://orbi.uliege.be/handle/2268/295234 https://doi.org/10.1615/Int.J.UncertaintyQuantification.2020032674 en eng Begell House Inc. http://www.dl.begellhouse.com/download/article/6958d1086605b6e3/IJUQ1102(1)-32674.pdf urn:issn:2152-5080 urn:issn:2152-5099 https://orbi.uliege.be/handle/2268/295234 info:hdl:2268/295234 doi:10.1615/Int.J.UncertaintyQuantification.2020032674 scopus-id:2-s2.0-85102479731 restricted access http://purl.org/coar/access_right/c_16ec info:eu-repo/semantics/restrictedAccess International Journal for Uncertainty Quantification, 11 (2), 1 - 23 (2021) Global sensitivity analysis Probabilistic learning on manifolds Small data Sobol index Statistics and Probability Modeling and Simulation Discrete Mathematics and Combinatorics Control and Optimization Engineering computing & technology Ingénierie informatique & technologie journal article http://purl.org/coar/resource_type/c_6501 info:eu-repo/semantics/article peer reviewed 2021 ftorbi https://doi.org/10.1615/Int.J.UncertaintyQuantification.2020032674 2024-03-27T14:56:48Z peer reviewed Global sensitivity analysis provides insight into how sources of uncertainty contribute to uncertainty in predictions of computational models. Global sensitivity indices, also called variance-based sensitivity indices and Sobol indices, are most often computed with Monte Carlo methods. However, when the computational model is computationally expensive and only a small number of samples can be generated, that is, in so-called small-data settings, usual Monte Carlo estimates may lack sufficient accuracy. As a means of improving accuracy in such small-data settings, we explore the use of probabilistic learning. The objective of the probabilistic learning is to learn from the available samples a probabilistic model that can be used to generate additional samples, from which Monte Carlo estimates of the global sensitivity indices are then deduced. We demonstrate the interest of such a probabilistic learning method by applying it in an illustration concerned with forecasting the contribution of the Antarctic ice sheet to sea level rise. Article in Journal/Newspaper Antarc* Antarctic Ice Sheet University of Liège: ORBi (Open Repository and Bibliography) International Journal for Uncertainty Quantification 11 2 1 23 |
institution |
Open Polar |
collection |
University of Liège: ORBi (Open Repository and Bibliography) |
op_collection_id |
ftorbi |
language |
English |
topic |
Global sensitivity analysis Probabilistic learning on manifolds Small data Sobol index Statistics and Probability Modeling and Simulation Discrete Mathematics and Combinatorics Control and Optimization Engineering computing & technology Ingénierie informatique & technologie |
spellingShingle |
Global sensitivity analysis Probabilistic learning on manifolds Small data Sobol index Statistics and Probability Modeling and Simulation Discrete Mathematics and Combinatorics Control and Optimization Engineering computing & technology Ingénierie informatique & technologie Arnst, Maarten Soize, Christian Bulthuis, Kevin Computation of sobol indices in global sensitivity analysis from small data sets by probabilistic learning on manifolds |
topic_facet |
Global sensitivity analysis Probabilistic learning on manifolds Small data Sobol index Statistics and Probability Modeling and Simulation Discrete Mathematics and Combinatorics Control and Optimization Engineering computing & technology Ingénierie informatique & technologie |
description |
peer reviewed Global sensitivity analysis provides insight into how sources of uncertainty contribute to uncertainty in predictions of computational models. Global sensitivity indices, also called variance-based sensitivity indices and Sobol indices, are most often computed with Monte Carlo methods. However, when the computational model is computationally expensive and only a small number of samples can be generated, that is, in so-called small-data settings, usual Monte Carlo estimates may lack sufficient accuracy. As a means of improving accuracy in such small-data settings, we explore the use of probabilistic learning. The objective of the probabilistic learning is to learn from the available samples a probabilistic model that can be used to generate additional samples, from which Monte Carlo estimates of the global sensitivity indices are then deduced. We demonstrate the interest of such a probabilistic learning method by applying it in an illustration concerned with forecasting the contribution of the Antarctic ice sheet to sea level rise. |
format |
Article in Journal/Newspaper |
author |
Arnst, Maarten Soize, Christian Bulthuis, Kevin |
author_facet |
Arnst, Maarten Soize, Christian Bulthuis, Kevin |
author_sort |
Arnst, Maarten |
title |
Computation of sobol indices in global sensitivity analysis from small data sets by probabilistic learning on manifolds |
title_short |
Computation of sobol indices in global sensitivity analysis from small data sets by probabilistic learning on manifolds |
title_full |
Computation of sobol indices in global sensitivity analysis from small data sets by probabilistic learning on manifolds |
title_fullStr |
Computation of sobol indices in global sensitivity analysis from small data sets by probabilistic learning on manifolds |
title_full_unstemmed |
Computation of sobol indices in global sensitivity analysis from small data sets by probabilistic learning on manifolds |
title_sort |
computation of sobol indices in global sensitivity analysis from small data sets by probabilistic learning on manifolds |
publisher |
Begell House Inc. |
publishDate |
2021 |
url |
https://orbi.uliege.be/handle/2268/295234 https://doi.org/10.1615/Int.J.UncertaintyQuantification.2020032674 |
genre |
Antarc* Antarctic Ice Sheet |
genre_facet |
Antarc* Antarctic Ice Sheet |
op_source |
International Journal for Uncertainty Quantification, 11 (2), 1 - 23 (2021) |
op_relation |
http://www.dl.begellhouse.com/download/article/6958d1086605b6e3/IJUQ1102(1)-32674.pdf urn:issn:2152-5080 urn:issn:2152-5099 https://orbi.uliege.be/handle/2268/295234 info:hdl:2268/295234 doi:10.1615/Int.J.UncertaintyQuantification.2020032674 scopus-id:2-s2.0-85102479731 |
op_rights |
restricted access http://purl.org/coar/access_right/c_16ec info:eu-repo/semantics/restrictedAccess |
op_doi |
https://doi.org/10.1615/Int.J.UncertaintyQuantification.2020032674 |
container_title |
International Journal for Uncertainty Quantification |
container_volume |
11 |
container_issue |
2 |
container_start_page |
1 |
op_container_end_page |
23 |
_version_ |
1796952806043680768 |