Time Series Procedures to Improve Extreme Quantile Estimation

Clara Cordeiro, Dora P. Gomes, M. Manuela Neves

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

1 Citation (Scopus)

Abstract

Although extreme events can occur rarely, they may have significant social and economic impacts. To assess the risk of extreme events, it is important to study the extreme quantiles of the distribution. The accurate semi-parametric estimation of high quantiles depends strongly on the estimation of some crucial parameters that appear in extreme value theory. Procedures that combine extreme value theory and time series modelling have revealed themselves as a nice compromise to capture extreme events. Here we study the estimation of extreme quantiles after adequate time series modelling. Using the R software, our approach will be applied to the daily mean flow discharge rate values of two rivers in Portugal.
Original languageEnglish
Title of host publicationStatistical Modelling and Risk Analysis
Subtitle of host publicationSelected contributions from ICRA9, Perugia, Italy, May 25-27, 2022
EditorsChristos P. Kitsos, Teresa A. Oliveira, Francesca Pierri, Marialuisa Restaino
Place of PublicationCham
PublisherSpringer
Pages69-80
Number of pages12
ISBN (Electronic)978-3-031-39864-3
ISBN (Print)978-3-031-39863-6
DOIs
Publication statusPublished - 2023
Event9th International Conference on Risk Analysis, ICRA9 2022 - Perugia, Italy
Duration: 25 May 202227 May 2022

Publication series

NameSpringer Proceedings in Mathematics and Statistics
PublisherSpringer
Volume430
ISSN (Print)2194-1009
ISSN (Electronic)2194-1017

Conference

Conference9th International Conference on Risk Analysis, ICRA9 2022
Country/TerritoryItaly
CityPerugia
Period25/05/2227/05/22

Keywords

  • Extreme value theory
  • High quantiles
  • Semiparametric estimation
  • Time series

Fingerprint

Dive into the research topics of 'Time Series Procedures to Improve Extreme Quantile Estimation'. Together they form a unique fingerprint.

Cite this