Home > Research > Publications & Outputs > Cross-validatory extreme value threshold select...

Links

Text available via DOI:

View graph of relations

Cross-validatory extreme value threshold selection and uncertainty with application to ocean storm severity

Research output: Contribution to Journal/MagazineJournal articlepeer-review

Published

Standard

Cross-validatory extreme value threshold selection and uncertainty with application to ocean storm severity. / Northrop, P.J.; Attalides, N.; Jonathan, P.
In: Journal of the Royal Statistical Society: Series C (Applied Statistics), Vol. 66, No. 1, 01.01.2017, p. 93-120.

Research output: Contribution to Journal/MagazineJournal articlepeer-review

Harvard

Northrop, PJ, Attalides, N & Jonathan, P 2017, 'Cross-validatory extreme value threshold selection and uncertainty with application to ocean storm severity', Journal of the Royal Statistical Society: Series C (Applied Statistics), vol. 66, no. 1, pp. 93-120. https://doi.org/10.1111/rssc.12159

APA

Northrop, P. J., Attalides, N., & Jonathan, P. (2017). Cross-validatory extreme value threshold selection and uncertainty with application to ocean storm severity. Journal of the Royal Statistical Society: Series C (Applied Statistics), 66(1), 93-120. https://doi.org/10.1111/rssc.12159

Vancouver

Northrop PJ, Attalides N, Jonathan P. Cross-validatory extreme value threshold selection and uncertainty with application to ocean storm severity. Journal of the Royal Statistical Society: Series C (Applied Statistics). 2017 Jan 1;66(1):93-120. Epub 2016 May 20. doi: 10.1111/rssc.12159

Author

Northrop, P.J. ; Attalides, N. ; Jonathan, P. / Cross-validatory extreme value threshold selection and uncertainty with application to ocean storm severity. In: Journal of the Royal Statistical Society: Series C (Applied Statistics). 2017 ; Vol. 66, No. 1. pp. 93-120.

Bibtex

@article{bf8058a3a33c487e937af49e5b9faa47,
title = "Cross-validatory extreme value threshold selection and uncertainty with application to ocean storm severity",
abstract = "Design conditions for marine structures are typically informed by threshold-based extreme value analyses of oceanographic variables, in which excesses of a high threshold are modelled by a generalized Pareto distribution. Too low a threshold leads to bias from model misspecification, and raising the threshold increases the variance of estimators: a bias–variance trade-off. Many existing threshold selection methods do not address this trade-off directly but rather aim to select the lowest threshold above which the generalized Pareto model is judged to hold approximately. In the paper Bayesian cross-validation is used to address the trade-off by comparing thresholds based on predictive ability at extreme levels. Extremal inferences can be sensitive to the choice of a single threshold. We use Bayesian model averaging to combine inferences from many thresholds, thereby reducing sensitivity to the choice of a single threshold. The methodology is applied to significant wave height data sets from the northern North Sea and the Gulf of Mexico. {\textcopyright} 2016 Royal Statistical Society",
keywords = "Cross-validation, Extreme value theory, Generalized Pareto distribution, Predictive inference, Threshold",
author = "P.J. Northrop and N. Attalides and P. Jonathan",
year = "2017",
month = jan,
day = "1",
doi = "10.1111/rssc.12159",
language = "English",
volume = "66",
pages = "93--120",
journal = "Journal of the Royal Statistical Society: Series C (Applied Statistics)",
issn = "0035-9254",
publisher = "Wiley-Blackwell",
number = "1",

}

RIS

TY - JOUR

T1 - Cross-validatory extreme value threshold selection and uncertainty with application to ocean storm severity

AU - Northrop, P.J.

AU - Attalides, N.

AU - Jonathan, P.

PY - 2017/1/1

Y1 - 2017/1/1

N2 - Design conditions for marine structures are typically informed by threshold-based extreme value analyses of oceanographic variables, in which excesses of a high threshold are modelled by a generalized Pareto distribution. Too low a threshold leads to bias from model misspecification, and raising the threshold increases the variance of estimators: a bias–variance trade-off. Many existing threshold selection methods do not address this trade-off directly but rather aim to select the lowest threshold above which the generalized Pareto model is judged to hold approximately. In the paper Bayesian cross-validation is used to address the trade-off by comparing thresholds based on predictive ability at extreme levels. Extremal inferences can be sensitive to the choice of a single threshold. We use Bayesian model averaging to combine inferences from many thresholds, thereby reducing sensitivity to the choice of a single threshold. The methodology is applied to significant wave height data sets from the northern North Sea and the Gulf of Mexico. © 2016 Royal Statistical Society

AB - Design conditions for marine structures are typically informed by threshold-based extreme value analyses of oceanographic variables, in which excesses of a high threshold are modelled by a generalized Pareto distribution. Too low a threshold leads to bias from model misspecification, and raising the threshold increases the variance of estimators: a bias–variance trade-off. Many existing threshold selection methods do not address this trade-off directly but rather aim to select the lowest threshold above which the generalized Pareto model is judged to hold approximately. In the paper Bayesian cross-validation is used to address the trade-off by comparing thresholds based on predictive ability at extreme levels. Extremal inferences can be sensitive to the choice of a single threshold. We use Bayesian model averaging to combine inferences from many thresholds, thereby reducing sensitivity to the choice of a single threshold. The methodology is applied to significant wave height data sets from the northern North Sea and the Gulf of Mexico. © 2016 Royal Statistical Society

KW - Cross-validation

KW - Extreme value theory

KW - Generalized Pareto distribution

KW - Predictive inference

KW - Threshold

U2 - 10.1111/rssc.12159

DO - 10.1111/rssc.12159

M3 - Journal article

VL - 66

SP - 93

EP - 120

JO - Journal of the Royal Statistical Society: Series C (Applied Statistics)

JF - Journal of the Royal Statistical Society: Series C (Applied Statistics)

SN - 0035-9254

IS - 1

ER -