Non-parametric methods for global sensitivity analysis of model output with dependent inputs

被引：78

作者：

Mara, Thierry A. ^{[1
]}

Tarantola, Stefano ^{[2
]}

Annoni, Paola ^{[3
]}

机构：

[1] Univ Reunion Isl, PIMENT, Dept Phys, UFR ST, 15 Ave Rene Cassin, F-97715 St Denis, Reunion, France

[2] Commiss European Communities, Joint Res Ctr, Inst Energy & Transport, I-21020 Ispra, VA, Italy

[3] Commiss European Communities, Econ Anal Unit, Directorate Gen Reg & Urban Policy, B-1049 Brussels, Belgium

来源：

ENVIRONMENTAL MODELLING & SOFTWARE | 2015年 / 72卷

关键词：

Dependent inputs; Rosenblatt transformation; Variance-based sensitivity indices; Dependent contributions; Independent contributions; Iman & Conover sampling procedure; Radionuclide migration; MATHEMATICAL-MODELS; NONLINEAR MODELS; INDEXES; UNCERTAINTY; VARIABLES;

D O I：

10.1016/j.envsoft.2015.07.010

中图分类号：

TP39 [计算机的应用];

学科分类号：

081203 ; 0835 ;

摘要：

This paper addresses the issue of performing global sensitivity analysis of model output with dependent inputs. First, we define variance-based sensitivity indices that allow for distinguishing the independent contributions of the inputs to the response variance from their mutual dependent contributions. Then, two sampling strategies are proposed for their non-parametric, numerical estimation. This approach allows us to estimate the sensitivity indices not only for individual inputs but also for groups of inputs. After testing the accuracy of the non-parametric method on some analytical test functions, the approach is employed to assess the importance of dependent inputs on a computer model for the migration of radioactive substances in the geosphere. (C) 2015 Elsevier Ltd. All rights reserved.

引用

页码：173 / 183

页数：11

共 28 条

[1] Model emulation and moment-independent sensitivity analysis: An application to environmental modelling [J].

Borgonovo, E. ;

Castaings, W. ;

Tarantola, S. .

ENVIRONMENTAL MODELLING & SOFTWARE, 2012, 34 :105-115

[2] Measuring uncertainty importance: Investigation and comparison of alternative approaches [J].

Borgonovo, Emanuele .

RISK ANALYSIS, 2006, 26 (05) :1349-1361

[3] Sampling strategies in density-based sensitivity analysis [J].

Castaings, W. ;

Borgonovo, E. ;

Morris, M. D. ;

Tarantola, S. .

ENVIRONMENTAL MODELLING & SOFTWARE, 2012, 38 :13-26

[4] Local Polynomial Estimation for Sensitivity Analysis on Models With Correlated Inputs [J].

Da Veiga, Sebastien ;

Wahl, Francois ;

Gamboa, Fabrice .

TECHNOMETRICS, 2009, 51 (04) :452-463

[5] Importance measures in global sensitivity analysis of nonlinear models [J].

Homma, T ;

Saltelli, A .

RELIABILITY ENGINEERING & SYSTEM SAFETY, 1996, 52 (01) :1-17

[6] A DISTRIBUTION-FREE APPROACH TO INDUCING RANK CORRELATION AMONG INPUT VARIABLES [J].

IMAN, RL ;

CONOVER, WJ .

COMMUNICATIONS IN STATISTICS PART B-SIMULATION AND COMPUTATION, 1982, 11 (03) :311-334

[7] Estimation of global sensitivity indices for models with dependent variables [J].

Kucherenko, S. ;

Tarantola, S. ;

Annoni, P. .

COMPUTER PHYSICS COMMUNICATIONS, 2012, 183 (04) :937-946

[8] Global Sensitivity Analysis for Systems with Independent and/or Correlated Inputs [J].

Li, Genyuan ;

Rabitz, Herschel ;

Yelvington, Paul E. ;

Oluwole, Oluwayemisi O. ;

Bacon, Fred ;

Kolb, Charles E. ;

Schoendorf, Jacqueline .

JOURNAL OF PHYSICAL CHEMISTRY A, 2010, 114 (19) :6022-6032

[9] Nataf transformation based point estimate method [J].

Li HongShuang ;

Lue ZhenZhou ;

Yuan XiuKai .

CHINESE SCIENCE BULLETIN, 2008, 53 (17) :2586-2592

[10]

Liu P.L., 1986, Probabilistic Engineering Mechanics, V1, P105, DOI [DOI 10.1016/0266-8920(86)90033-0, 10.1016/0266-8920(86)90033-0]

← 1 2 3 →