The extremal dependence structure of a regularly varying random vector X is fully described by its limiting spectral measure. In this paper, we investigate how to recover characteristics of the measure, such as extremal coefficients, from the extremal behaviour of convex combinations of components of X. Our considerations result in a class of new estimators of moments of the corresponding combinations for the spectral vector. We show asymptotic normality by means of a functional limit theorem and, focusing on the estimation of extremal coefficients, we verify that the minimal asymptotic variance can be achieved by a plug-in estimator using subsampling bootstrap. We illustrate the benefits of our approach on simulated and real data.
%0 Journal Article
%1 10.1214/24-AOS2387
%A Oesting, Marco
%A Wintenberger, Olivier
%D 2024
%I Institute of Mathematical Statistics
%J The Annals of Statistics
%K exc2075 myown ourwork peerreviewed pn5 pn5-10
%N 6
%P 2529 -- 2556
%R 10.1214/24-AOS2387
%T Estimation of the spectral measure from convex combinations of regularly varying random vectors
%U https://doi.org/10.1214/24-AOS2387
%V 52
%X The extremal dependence structure of a regularly varying random vector X is fully described by its limiting spectral measure. In this paper, we investigate how to recover characteristics of the measure, such as extremal coefficients, from the extremal behaviour of convex combinations of components of X. Our considerations result in a class of new estimators of moments of the corresponding combinations for the spectral vector. We show asymptotic normality by means of a functional limit theorem and, focusing on the estimation of extremal coefficients, we verify that the minimal asymptotic variance can be achieved by a plug-in estimator using subsampling bootstrap. We illustrate the benefits of our approach on simulated and real data.
@article{10.1214/24-AOS2387,
abstract = {The extremal dependence structure of a regularly varying random vector X is fully described by its limiting spectral measure. In this paper, we investigate how to recover characteristics of the measure, such as extremal coefficients, from the extremal behaviour of convex combinations of components of X. Our considerations result in a class of new estimators of moments of the corresponding combinations for the spectral vector. We show asymptotic normality by means of a functional limit theorem and, focusing on the estimation of extremal coefficients, we verify that the minimal asymptotic variance can be achieved by a plug-in estimator using subsampling bootstrap. We illustrate the benefits of our approach on simulated and real data.},
added-at = {2024-12-30T16:48:12.000+0100},
author = {Oesting, Marco and Wintenberger, Olivier},
biburl = {https://puma.ub.uni-stuttgart.de/bibtex/27d5360d7c737e3ce987e385ac79cc245/marcooesting},
doi = {10.1214/24-AOS2387},
interhash = {1917717d09db562e04dede966123a64b},
intrahash = {7d5360d7c737e3ce987e385ac79cc245},
journal = {The Annals of Statistics},
keywords = {exc2075 myown ourwork peerreviewed pn5 pn5-10},
number = 6,
pages = {2529 -- 2556},
publisher = {Institute of Mathematical Statistics},
timestamp = {2024-12-30T16:48:12.000+0100},
title = {{Estimation of the spectral measure from convex combinations of regularly varying random vectors}},
url = {https://doi.org/10.1214/24-AOS2387},
volume = 52,
year = 2024
}
