Local dependence functions for extreme value distributions
Kotz & Nadarajah (2002) introduced a measure of local dependence which is a localized version of the Pearson's correlation coefficient. In this paper we provide detailed analyses (both algebraic and numerical) of the form of the measure for the class of bivariate extreme value distributions. We consider, in particular, five families of bivariate extreme value distributions. We also discuss two applications of the new measure. In the first application we introduce an overall measure of correlation and produce evidence to suggest that it is superior than the usual Pearson's correlation coefficient. The second application introduces two new concepts for ordering of bivariate dependence.
Year of publication: |
2003
|
---|---|
Authors: | Nadarajah, Saralees ; Mitov, Kosto ; Kotz, Samuel |
Published in: |
Journal of Applied Statistics. - Taylor & Francis Journals, ISSN 0266-4763. - Vol. 30.2003, 10, p. 1081-1100
|
Publisher: |
Taylor & Francis Journals |
Saved in:
Saved in favorites
Similar items by person
-
Reliability For A Bivariate Gamma Distribution
Nadarajah, Saralees, (2005)
-
Ratio of Logistic and Bessel Random Variables
Nadarajah, Saralees, (2005)
-
The modified Weibull distribution for asset returns
Nadarajah, Saralees, (2006)
- More ...