Open access
Date
2010-03Type
- Journal Article
ETH Bibliography
yes
Altmetrics
Abstract
This paper explores the joint extreme-value behavior of discontinuous random variables. It is shown that as in the continuous case, the latter is characterized by the weak limit of the normalized componentwise maxima and the convergence of any compatible copula. Illustrations are provided and an extension to the case of triangular arrays is considered which sheds new light on recent work of Coles and Pauli (Stat Probab Lett 54:373–379, 2001) and Mitov and Nadarajah (Extremes 8:357–370, 2005). This leads to considerations on the meaning of the bivariate upper tail dependence coefficient of Joe (Comput Stat Data Anal 16:279–297, 1993) in the discontinuous case. Show more
Permanent link
https://doi.org/10.3929/ethz-b-000020735Publication status
publishedExternal links
Journal / series
ExtremesVolume
Pages / Article No.
Publisher
SpringerSubject
Copula; Discrete distribution; Joint extremes; Maximum; Upper tail dependence; Triangular arrayOrganisational unit
02204 - RiskLab / RiskLab
Notes
It was possible to publish this article open access thanks to a Swiss National Licence with the publisherMore
Show all metadata
ETH Bibliography
yes
Altmetrics