Local Polynomial Estimation for Sensitivity Analysis on Models With Correlated Inputs

TL;DR

This paper introduces local polynomial methods for estimating sensitivity indices in models with correlated inputs, providing new estimators with strong theoretical properties and practical application to a kinetic model.

Contribution

It proposes two novel local polynomial estimators for sensitivity analysis with correlated inputs, supported by theoretical analysis and real-world application.

Findings

Both estimators have desirable theoretical properties.

The methods effectively analyze sensitivity in correlated input models.

Application to a kinetic model demonstrates practical utility.

Abstract

Sensitivity indices when the inputs of a model are not independent are estimated by local polynomial techniques. Two original estimators based on local polynomial smoothers are proposed. Both have good theoretical properties which are exhibited and also illustrated through analytical examples. They are used to carry out a sensitivity analysis on a real case of a kinetic model with correlated parameters.