Global sensitivity analysis and Wasserstein spaces

TL;DR

This paper develops new global sensitivity analysis methods for computer codes with outputs as distribution functions or stochastic models, using Wasserstein spaces and Fréchet means, with practical procedures and numerical illustrations.

Contribution

It introduces novel sensitivity indices based on Wasserstein metrics for distributional outputs and stochastic codes, extending traditional methods to these complex settings.

Findings

New sensitivity indices based on Wasserstein Fréchet means.

Procedures for second-level sensitivity analysis on input distributions.

Numerical studies demonstrating the methods' effectiveness.

Abstract

Sensitivity indices are commonly used to quantity the relative inuence of any specic group of input variables on the output of a computer code. In this paper, we focus both on computer codes the output of which is a cumulative distribution function and on stochastic computer codes. We propose a way to perform a global sensitivity analysis for these kinds of computer codes. In the rst setting, we dene two indices: the rst one is based on Wasserstein Fr{\'e}chet means while the second one is based on the Hoeding decomposition of the indicators of Wasserstein balls. Further, when dealing with the stochastic computer codes, we dene an ideal version of the stochastic computer code thats ts into the frame of the rst setting. Finally, we deduce a procedure to realize a second level global sensitivity analysis, namely when one is interested in the sensitivity related to the input distributions rather than in the sensitivity related to the inputs themselves. Several numerical studies are proposed as illustrations in the dierent settings.