An M-estimator of spatial tail dependence

John Einmahl, Anna Kiriliouk, A. Krajina, J.J.J. Segers

Research output: Contribution to journalArticleScientificpeer-review

Abstract

Tail dependence models for distributions attracted to a max-stable law are fitted by using observations above a high threshold.To cope with spatial, high dimensional data, a rankbased M-estimator is proposed relying on bivariate margins only. A data-driven weight matrix is used to minimize the asymptotic variance. Empirical process arguments show that the estimator is consistent and asymptotically normal. Its finite sample performance is assessed in simulation experiments involving popular max-stable processes perturbed with additive noise. An analysis of wind speed data from the Netherlands illustrates the method.
Original languageEnglish
Pages (from-to)275-298
Number of pages23
JournalJournal of the Royal Statistical Society, Series B
Volume78
Issue number1
DOIs
Publication statusPublished - Jan 2016

Fingerprint

Tail Dependence
Spatial Dependence
Stable Laws
M-estimator
Stable Process
Empirical Process
Asymptotic Variance
Wind Speed
Additive Noise
High-dimensional Data
Data-driven
Margin
Simulation Experiment
Minimise
Estimator
Model
Observation
Asymptotic variance
Finite sample
Simulation experiment

Keywords

  • Brown-Resnick process
  • exceedances
  • multivariate extremes
  • ranks
  • spatial statistics
  • stable tail dependence function

Cite this

Einmahl, John ; Kiriliouk, Anna ; Krajina, A. ; Segers, J.J.J. / An M-estimator of spatial tail dependence. In: Journal of the Royal Statistical Society, Series B. 2016 ; Vol. 78, No. 1. pp. 275-298.
@article{51403ee5285e419ba3697999e49bac6d,
title = "An M-estimator of spatial tail dependence",
abstract = "Tail dependence models for distributions attracted to a max-stable law are fitted by using observations above a high threshold.To cope with spatial, high dimensional data, a rankbased M-estimator is proposed relying on bivariate margins only. A data-driven weight matrix is used to minimize the asymptotic variance. Empirical process arguments show that the estimator is consistent and asymptotically normal. Its finite sample performance is assessed in simulation experiments involving popular max-stable processes perturbed with additive noise. An analysis of wind speed data from the Netherlands illustrates the method.",
keywords = "Brown-Resnick process, exceedances, multivariate extremes, ranks, spatial statistics, stable tail dependence function",
author = "John Einmahl and Anna Kiriliouk and A. Krajina and J.J.J. Segers",
year = "2016",
month = "1",
doi = "10.1111/rssb.12114",
language = "English",
volume = "78",
pages = "275--298",
journal = "Journal of the Royal Statistical Society, Series B",
issn = "1369-7412",
publisher = "Wiley-Blackwell",
number = "1",

}

An M-estimator of spatial tail dependence. / Einmahl, John; Kiriliouk, Anna; Krajina, A.; Segers, J.J.J.

In: Journal of the Royal Statistical Society, Series B, Vol. 78, No. 1, 01.2016, p. 275-298.

Research output: Contribution to journalArticleScientificpeer-review

TY - JOUR

T1 - An M-estimator of spatial tail dependence

AU - Einmahl, John

AU - Kiriliouk, Anna

AU - Krajina, A.

AU - Segers, J.J.J.

PY - 2016/1

Y1 - 2016/1

N2 - Tail dependence models for distributions attracted to a max-stable law are fitted by using observations above a high threshold.To cope with spatial, high dimensional data, a rankbased M-estimator is proposed relying on bivariate margins only. A data-driven weight matrix is used to minimize the asymptotic variance. Empirical process arguments show that the estimator is consistent and asymptotically normal. Its finite sample performance is assessed in simulation experiments involving popular max-stable processes perturbed with additive noise. An analysis of wind speed data from the Netherlands illustrates the method.

AB - Tail dependence models for distributions attracted to a max-stable law are fitted by using observations above a high threshold.To cope with spatial, high dimensional data, a rankbased M-estimator is proposed relying on bivariate margins only. A data-driven weight matrix is used to minimize the asymptotic variance. Empirical process arguments show that the estimator is consistent and asymptotically normal. Its finite sample performance is assessed in simulation experiments involving popular max-stable processes perturbed with additive noise. An analysis of wind speed data from the Netherlands illustrates the method.

KW - Brown-Resnick process

KW - exceedances

KW - multivariate extremes

KW - ranks

KW - spatial statistics

KW - stable tail dependence function

U2 - 10.1111/rssb.12114

DO - 10.1111/rssb.12114

M3 - Article

VL - 78

SP - 275

EP - 298

JO - Journal of the Royal Statistical Society, Series B

JF - Journal of the Royal Statistical Society, Series B

SN - 1369-7412

IS - 1

ER -