This chapter considers how to treat spoken and unspoken assumptions and how to use uncertainty quantification and global sensitivity analysis to make them transparent. As part of this, the broad role of assumptions (or hypotheses) in scientific modelling are analysed, and investigations are made into how they affect alternative elements of a model. We address models of data (machine learning) and models of phenomena (simulators). We then discuss the impact of varying assumptions on the output of a mathematical model, highlighting the role of uncertainty quantification and sensitivity analysis. We single out four main sensitivity analysis goals emerging from the literature.

Mind the assumptions: quantify uncertainty and assess sensitivity

Borgonovo, Emanuele
2023

Abstract

This chapter considers how to treat spoken and unspoken assumptions and how to use uncertainty quantification and global sensitivity analysis to make them transparent. As part of this, the broad role of assumptions (or hypotheses) in scientific modelling are analysed, and investigations are made into how they affect alternative elements of a model. We address models of data (machine learning) and models of phenomena (simulators). We then discuss the impact of varying assumptions on the output of a mathematical model, highlighting the role of uncertainty quantification and sensitivity analysis. We single out four main sensitivity analysis goals emerging from the literature.
2023
9780198872412
9780191983597
Saltelli, Andrea; Di Fiore, Monica
The politics of modelling: numbers between science and policy
Borgonovo, Emanuele
File in questo prodotto:
File Dimensione Formato  
ModelsAssumptions.pdf

accesso aperto

Descrizione: File Preprint
Tipologia: Documento in Pre-print (Pre-print document)
Licenza: NON PUBBLICO - Accesso privato/ristretto
Dimensione 1.04 MB
Formato Adobe PDF
1.04 MB Adobe PDF Visualizza/Apri

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11565/4062079
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 1
  • ???jsp.display-item.citation.isi??? ND
social impact