Shapley effects for sensitivity analysis with correlated inputs: comparisons with Sobol' indices, numerical estimation and applications

TL;DR

This paper compares Shapley effects and Sobol' indices for sensitivity analysis with correlated inputs, demonstrating the advantages of Shapley effects in interpretability and proposing efficient estimation methods using metamodels.

Contribution

It provides new analytical insights into how Shapley effects behave with correlated Gaussian inputs and compares them to Sobol' indices, highlighting their interpretability benefits.

Findings

Shapley effects effectively handle correlated inputs in sensitivity analysis.

Metamodel-based estimation reduces computational costs for complex models.

Shapley effects offer clearer interpretation than Sobol' indices with dependent inputs.

Abstract

The global sensitivity analysis of a numerical model aims to quantify, by means of sensitivity indices estimate, the contributions of each uncertain input variable to the model output uncertainty. The so-called Sobol' indices, which are based on the functional variance analysis, present a difficult interpretation in the presence of statistical dependence between inputs. The Shapley effect was recently introduced to overcome this problem as they allocate the mutual contribution (due to correlation and interaction) of a group of inputs to each individual input within the group.In this paper, using several new analytical results, we study the effects of linear correlation between some Gaussian input variables on Shapley effects, and compare these effects to classical first-order and total Sobol' indices.This illustrates the interest, in terms of sensitivity analysis setting and interpretation, of the Shapley effects in the case of dependent inputs. For the practical issue of computationally demanding computer models, we show that the substitution of the original model by a metamodel (here, kriging) makes it possible to estimate these indices with precision at a reasonable computational cost.