Cargando…

Models with higher effective dimensions tend to produce more uncertain estimates

Mathematical models are getting increasingly detailed to better predict phenomena or gain more accurate insights into the dynamics of a system of interest, even when there are no validation or training data available. Here, we show through ANOVA and statistical theory that this practice promotes fuz...

Descripción completa

Detalles Bibliográficos
Autores principales: Puy, Arnald, Beneventano, Pierfrancesco, Levin, Simon A., Lo Piano, Samuele, Portaluri, Tommaso, Saltelli, Andrea
Formato: Online Artículo Texto
Lenguaje:English
Publicado: American Association for the Advancement of Science 2022
Materias:
Acceso en línea:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9581491/
https://www.ncbi.nlm.nih.gov/pubmed/36260678
http://dx.doi.org/10.1126/sciadv.abn9450
_version_ 1784812637908893696
author Puy, Arnald
Beneventano, Pierfrancesco
Levin, Simon A.
Lo Piano, Samuele
Portaluri, Tommaso
Saltelli, Andrea
author_facet Puy, Arnald
Beneventano, Pierfrancesco
Levin, Simon A.
Lo Piano, Samuele
Portaluri, Tommaso
Saltelli, Andrea
author_sort Puy, Arnald
collection PubMed
description Mathematical models are getting increasingly detailed to better predict phenomena or gain more accurate insights into the dynamics of a system of interest, even when there are no validation or training data available. Here, we show through ANOVA and statistical theory that this practice promotes fuzzier estimates because it generally increases the model’s effective dimensions, i.e., the number of influential parameters and the weight of high-order interactions. By tracking the evolution of the effective dimensions and the output uncertainty at each model upgrade stage, modelers can better ponder whether the addition of detail truly matches the model’s purpose and the quality of the data fed into it.
format Online
Article
Text
id pubmed-9581491
institution National Center for Biotechnology Information
language English
publishDate 2022
publisher American Association for the Advancement of Science
record_format MEDLINE/PubMed
spelling pubmed-95814912022-10-26 Models with higher effective dimensions tend to produce more uncertain estimates Puy, Arnald Beneventano, Pierfrancesco Levin, Simon A. Lo Piano, Samuele Portaluri, Tommaso Saltelli, Andrea Sci Adv Social and Interdisciplinary Sciences Mathematical models are getting increasingly detailed to better predict phenomena or gain more accurate insights into the dynamics of a system of interest, even when there are no validation or training data available. Here, we show through ANOVA and statistical theory that this practice promotes fuzzier estimates because it generally increases the model’s effective dimensions, i.e., the number of influential parameters and the weight of high-order interactions. By tracking the evolution of the effective dimensions and the output uncertainty at each model upgrade stage, modelers can better ponder whether the addition of detail truly matches the model’s purpose and the quality of the data fed into it. American Association for the Advancement of Science 2022-10-19 /pmc/articles/PMC9581491/ /pubmed/36260678 http://dx.doi.org/10.1126/sciadv.abn9450 Text en Copyright © 2022 The Authors, some rights reserved; exclusive licensee American Association for the Advancement of Science. No claim to original U.S. Government Works. Distributed under a Creative Commons Attribution License 4.0 (CC BY). https://creativecommons.org/licenses/by/4.0/This is an open-access article distributed under the terms of the Creative Commons Attribution license (https://creativecommons.org/licenses/by/4.0/) , which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
spellingShingle Social and Interdisciplinary Sciences
Puy, Arnald
Beneventano, Pierfrancesco
Levin, Simon A.
Lo Piano, Samuele
Portaluri, Tommaso
Saltelli, Andrea
Models with higher effective dimensions tend to produce more uncertain estimates
title Models with higher effective dimensions tend to produce more uncertain estimates
title_full Models with higher effective dimensions tend to produce more uncertain estimates
title_fullStr Models with higher effective dimensions tend to produce more uncertain estimates
title_full_unstemmed Models with higher effective dimensions tend to produce more uncertain estimates
title_short Models with higher effective dimensions tend to produce more uncertain estimates
title_sort models with higher effective dimensions tend to produce more uncertain estimates
topic Social and Interdisciplinary Sciences
url https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9581491/
https://www.ncbi.nlm.nih.gov/pubmed/36260678
http://dx.doi.org/10.1126/sciadv.abn9450
work_keys_str_mv AT puyarnald modelswithhighereffectivedimensionstendtoproducemoreuncertainestimates
AT beneventanopierfrancesco modelswithhighereffectivedimensionstendtoproducemoreuncertainestimates
AT levinsimona modelswithhighereffectivedimensionstendtoproducemoreuncertainestimates
AT lopianosamuele modelswithhighereffectivedimensionstendtoproducemoreuncertainestimates
AT portaluritommaso modelswithhighereffectivedimensionstendtoproducemoreuncertainestimates
AT saltelliandrea modelswithhighereffectivedimensionstendtoproducemoreuncertainestimates