Cover Image

Covariate-varying threshold selection method in non-stationary generalized pareto model

Non-stationary data usually exist in real life and influenced by covariates. The non-stationary extremes are usually modelled by setting a constant high threshold, u, where the threshold exceedances are modelled by Generalized Pareto distribution (GP). Covariates model is incorporated to the GP par...

Full description

Saved in:
Bibliographic Details
Main Authors: Shihabuddin, Afif, Ali, Norhaslinda, Adam, Mohd Bakri
Format: Article
Language:English
Published: Academy of Sciences Malaysia 2019
Online Access:http://psasir.upm.edu.my/id/eprint/81045/1/PARETO.pdf
http://psasir.upm.edu.my/id/eprint/81045/
https://www.akademisains.gov.my/asmsj/article/covariate-varying-threshold-selection-method-in-non-stationary-generalized-pareto-model/
Tags: Add Tag
No Tags, Be the first to tag this record!
id my.upm.eprints.81045
record_format eprints
spelling my.upm.eprints.810452021-04-29T00:56:24Z http://psasir.upm.edu.my/id/eprint/81045/ Covariate-varying threshold selection method in non-stationary generalized pareto model Shihabuddin, Afif Ali, Norhaslinda Adam, Mohd Bakri Non-stationary data usually exist in real life and influenced by covariates. The non-stationary extremes are usually modelled by setting a constant high threshold, u, where the threshold exceedances are modelled by Generalized Pareto distribution (GP). Covariates model is incorporated to the GP parameters to account for non-stationarity. However, the threshold, u, may be high enough for GP approximation on certain covariates but not on others, which in this case may violate the asymptotic basis of the GP model. In this paper, a covariate-varying threshold selection method based on regression tree is suggested and applied on simulated non-stationary data sets. The regression tree will be used to partition data sets into stationary groups with similar covariate condition. Thus, a constant high threshold can be fixed within each group. The tree-based threshold exceedances can then be modelled by stationary GP which is a simpler model compared to the non-stationary GP. Simulation study is done to demonstrate and assess the performance of this method compared to the conventional method. The results show that the proposed method is a reasonable complement to the conventional method. Academy of Sciences Malaysia 2019 Article PeerReviewed text en http://psasir.upm.edu.my/id/eprint/81045/1/PARETO.pdf Shihabuddin, Afif and Ali, Norhaslinda and Adam, Mohd Bakri (2019) Covariate-varying threshold selection method in non-stationary generalized pareto model. ASM Science Journal, 12 (spec. 3). pp. 285-297. ISSN 1823-6782 https://www.akademisains.gov.my/asmsj/article/covariate-varying-threshold-selection-method-in-non-stationary-generalized-pareto-model/
institution Universiti Putra Malaysia
building UPM Library
collection Institutional Repository
continent Asia
country Malaysia
content_provider Universiti Putra Malaysia
content_source UPM Institutional Repository
url_provider http://psasir.upm.edu.my/
language English
description Non-stationary data usually exist in real life and influenced by covariates. The non-stationary extremes are usually modelled by setting a constant high threshold, u, where the threshold exceedances are modelled by Generalized Pareto distribution (GP). Covariates model is incorporated to the GP parameters to account for non-stationarity. However, the threshold, u, may be high enough for GP approximation on certain covariates but not on others, which in this case may violate the asymptotic basis of the GP model. In this paper, a covariate-varying threshold selection method based on regression tree is suggested and applied on simulated non-stationary data sets. The regression tree will be used to partition data sets into stationary groups with similar covariate condition. Thus, a constant high threshold can be fixed within each group. The tree-based threshold exceedances can then be modelled by stationary GP which is a simpler model compared to the non-stationary GP. Simulation study is done to demonstrate and assess the performance of this method compared to the conventional method. The results show that the proposed method is a reasonable complement to the conventional method.
format Article
author Shihabuddin, Afif
Ali, Norhaslinda
Adam, Mohd Bakri
spellingShingle Shihabuddin, Afif
Ali, Norhaslinda
Adam, Mohd Bakri
Covariate-varying threshold selection method in non-stationary generalized pareto model
author_facet Shihabuddin, Afif
Ali, Norhaslinda
Adam, Mohd Bakri
author_sort Shihabuddin, Afif
title Covariate-varying threshold selection method in non-stationary generalized pareto model
title_short Covariate-varying threshold selection method in non-stationary generalized pareto model
title_full Covariate-varying threshold selection method in non-stationary generalized pareto model
title_fullStr Covariate-varying threshold selection method in non-stationary generalized pareto model
title_full_unstemmed Covariate-varying threshold selection method in non-stationary generalized pareto model
title_sort covariate-varying threshold selection method in non-stationary generalized pareto model
publisher Academy of Sciences Malaysia
publishDate 2019
url http://psasir.upm.edu.my/id/eprint/81045/1/PARETO.pdf
http://psasir.upm.edu.my/id/eprint/81045/
https://www.akademisains.gov.my/asmsj/article/covariate-varying-threshold-selection-method-in-non-stationary-generalized-pareto-model/
_version_ 1698698930555977728
score 13.211869

Similar Items