Subinterval sensitivity for high dimensional models

Tools

Ochnio, Dawid and de Angelis, Marco; (2025) Subinterval sensitivity for high dimensional models. In: Proc. of the 35th European Safety and Reliability & 33rd Society for Risk Analysis Europe Conference. Research Publishing Services, NOR. (In Press)

[thumbnail of Ochnio-de-Angelis-ESREL-SRAE-2025-Subinterval-sensitivity-for-high-dimensional-models] Text. Filename: Ochnio-de-Angelis-ESREL-SRAE-2025-Subinterval-sensitivity-for-high-dimensional-models.pdf
Accepted Author Manuscript
Restricted to Repository staff only until 1 January 2099.
Download (315kB) | Request a copy

Abstract

This paper introduces an interval-based non-probabilistic sensitivity analysis method, named subinterval sensitivity. A powerful, reliable and rigorous sensitivity analysis method, which is best suited to quantify the importance of inputs purely with respect to their mathematical model. The method has only recently and partially appeared in the literature, while its scalability to high-dimensional models is claimed here for the first time. We apply subinterval sensitivity to quantify and rank the importance of the parameters of a trained neural network model while drawing comparisons with the established Sobol' sensitivity analysis method. Sensitivities on the parameters of a trained neural network can shed light on overparametrization and explainability of the neural network surrogate model.

ORCID iDs

Ochnio, Dawid and de Angelis, Marco ORCID logoORCID: https://orcid.org/0000-0001-8851-023X;
CORE (COnnecting REpositories)