Minimal Gelbrich Distance to Uncorrelation - Archive ouverte HAL
Journal Articles IEEE Control Systems Letters Year : 2023

Minimal Gelbrich Distance to Uncorrelation

Abstract

This letter reports new properties of the Wasserstein/Gelbrich distance and associated ambiguity sets to analyze the correlation between two scalar random variables. A simple closed expression is derived for the Gelbrich distance between two bidimensional random distributions. Moreover, the minimum disturbance in the Gelbrich metric required to reach uncorrelation between two random variables is obtained. This allows us to determine the robustness of the Pearson coefficient within an ambiguity set. A numerical example showcases the potential use of the obtained results in the field of variable selection.
Fichier principal
Vignette du fichier
version_finale_LCSS_24_Minimal_Distance_to_Uncorrelation.pdf (409.63 Ko) Télécharger le fichier
Origin Files produced by the author(s)

Dates and versions

hal-04379498 , version 1 (31-05-2024)
hal-04379498 , version 2 (29-07-2024)

Identifiers

Cite

Matthieu Borelle, Teodoro Alamo, Cristina Stoica Maniu, Sylvain Bertrand, Eduardo Camacho. Minimal Gelbrich Distance to Uncorrelation. IEEE Control Systems Letters, 2023, 8, pp.61-66. ⟨10.1109/LCSYS.2023.3343990⟩. ⟨hal-04379498v2⟩
152 View
58 Download

Altmetric

Share

More