Back To Index Previous Article Next Article Full Text


Statistica Sinica 20 (2010), 365-378





A PRACTICAL WAY FOR ESTIMATING TAIL

DEPENDENCE FUNCTIONS


Liang Peng


Georgia Institute of Technology


Abstract: Estimating tail dependence functions is important for applications of multivariate extreme value theory, and only a fraction of the upper order statistics are involved in the estimation. How to choose the sample fraction or threshold is of importance in practice. Motivated by the recent methodologies on threshold selection for a tail index in Guillou and Hall (2001) and Peng (2009a), we apply the idea in Peng (2009a) to obtain a data-driven method for choosing the threshold in estimating a tail dependence function. Further we propose a simple bias-reduction estimator, and the combination of the bias-reduction estimator with the threshold selection procedure gives a satisfactory way of estimating a tail dependence function. This is supported by a simulation study. Moreover, a sub-sample bootstrap method is proposed to construct a confidence interval for a tail dependence function.



Key words and phrases: Bivariate extremes, bootstrap, tail dependence, threshold.

Back To Index Previous Article Next Article Full Text