A Common Rationale for Global Sensitivity Measures and Their Estimation

Emanuele Borgonovo*, Gordon B Hazen, Elmar Plischke

*Corresponding author for this work

Research output: Contribution to journalArticle

28 Citations (Scopus)

Abstract

Measures of sensitivity and uncertainty have become an integral part of risk analysis. Many such measures have a conditional probabilistic structure, for which a straightforward Monte Carlo estimation procedure has a double-loop form. Recently, a more efficient single-loop procedure has been introduced, and consistency of this procedure has been demonstrated separately for particular measures, such as those based on variance, density, and information value. In this work, we give a unified proof of single-loop consistency that applies to any measure satisfying a common rationale. This proof is not only more general but invokes less restrictive assumptions than heretofore in the literature, allowing for the presence of correlations among model inputs and of categorical variables. We examine numerical convergence of such an estimator under a variety of sensitivity measures. We also examine its application to a published medical case study.

Original languageEnglish (US)
Pages (from-to)1871-1895
Number of pages25
JournalRisk Analysis
Volume36
Issue number10
DOIs
StatePublished - Oct 1 2016

Fingerprint

Risk analysis
Uncertainty

Keywords

  • Global sensitivity measures
  • Monte Carlo simulation
  • probabilistic sensitivity analysis
  • risk analysis
  • uncertainty analysis

ASJC Scopus subject areas

  • Safety, Risk, Reliability and Quality
  • Physiology (medical)

Cite this

Borgonovo, Emanuele ; Hazen, Gordon B ; Plischke, Elmar. / A Common Rationale for Global Sensitivity Measures and Their Estimation. In: Risk Analysis. 2016 ; Vol. 36, No. 10. pp. 1871-1895.
@article{11067afb023a48b3af64285a64512a29,
title = "A Common Rationale for Global Sensitivity Measures and Their Estimation",
abstract = "Measures of sensitivity and uncertainty have become an integral part of risk analysis. Many such measures have a conditional probabilistic structure, for which a straightforward Monte Carlo estimation procedure has a double-loop form. Recently, a more efficient single-loop procedure has been introduced, and consistency of this procedure has been demonstrated separately for particular measures, such as those based on variance, density, and information value. In this work, we give a unified proof of single-loop consistency that applies to any measure satisfying a common rationale. This proof is not only more general but invokes less restrictive assumptions than heretofore in the literature, allowing for the presence of correlations among model inputs and of categorical variables. We examine numerical convergence of such an estimator under a variety of sensitivity measures. We also examine its application to a published medical case study.",
keywords = "Global sensitivity measures, Monte Carlo simulation, probabilistic sensitivity analysis, risk analysis, uncertainty analysis",
author = "Emanuele Borgonovo and Hazen, {Gordon B} and Elmar Plischke",
year = "2016",
month = "10",
day = "1",
doi = "10.1111/risa.12555",
language = "English (US)",
volume = "36",
pages = "1871--1895",
journal = "Risk Analysis",
issn = "0272-4332",
publisher = "Wiley-Blackwell",
number = "10",

}

A Common Rationale for Global Sensitivity Measures and Their Estimation. / Borgonovo, Emanuele; Hazen, Gordon B; Plischke, Elmar.

In: Risk Analysis, Vol. 36, No. 10, 01.10.2016, p. 1871-1895.

Research output: Contribution to journalArticle

TY - JOUR

T1 - A Common Rationale for Global Sensitivity Measures and Their Estimation

AU - Borgonovo, Emanuele

AU - Hazen, Gordon B

AU - Plischke, Elmar

PY - 2016/10/1

Y1 - 2016/10/1

N2 - Measures of sensitivity and uncertainty have become an integral part of risk analysis. Many such measures have a conditional probabilistic structure, for which a straightforward Monte Carlo estimation procedure has a double-loop form. Recently, a more efficient single-loop procedure has been introduced, and consistency of this procedure has been demonstrated separately for particular measures, such as those based on variance, density, and information value. In this work, we give a unified proof of single-loop consistency that applies to any measure satisfying a common rationale. This proof is not only more general but invokes less restrictive assumptions than heretofore in the literature, allowing for the presence of correlations among model inputs and of categorical variables. We examine numerical convergence of such an estimator under a variety of sensitivity measures. We also examine its application to a published medical case study.

AB - Measures of sensitivity and uncertainty have become an integral part of risk analysis. Many such measures have a conditional probabilistic structure, for which a straightforward Monte Carlo estimation procedure has a double-loop form. Recently, a more efficient single-loop procedure has been introduced, and consistency of this procedure has been demonstrated separately for particular measures, such as those based on variance, density, and information value. In this work, we give a unified proof of single-loop consistency that applies to any measure satisfying a common rationale. This proof is not only more general but invokes less restrictive assumptions than heretofore in the literature, allowing for the presence of correlations among model inputs and of categorical variables. We examine numerical convergence of such an estimator under a variety of sensitivity measures. We also examine its application to a published medical case study.

KW - Global sensitivity measures

KW - Monte Carlo simulation

KW - probabilistic sensitivity analysis

KW - risk analysis

KW - uncertainty analysis

UR - http://www.scopus.com/inward/record.url?scp=84996540277&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84996540277&partnerID=8YFLogxK

U2 - 10.1111/risa.12555

DO - 10.1111/risa.12555

M3 - Article

C2 - 26857789

AN - SCOPUS:84996540277

VL - 36

SP - 1871

EP - 1895

JO - Risk Analysis

JF - Risk Analysis

SN - 0272-4332

IS - 10

ER -