Estimating the spectral measure of an extreme value distribution (Q1275958): Difference between revisions

Let \((X_i, Y_i),\;i=1,\dots , n\) be i.i.d. random vectors with a bivariate distribution function \(F\in \mathcal D(G)\) (domain of attraction of a bivariate extreme value distribution \(G).\) The distribution function \(G\) is assumed to have non-degenerate marginal distributions, and it is characterized by extreme value indices \(\gamma _1, \gamma _2\) and by the spectral measure \(\varPhi .\) An estimator \(\hat \varPhi \) of \(\varPhi \) is constructed on the basis of the sample \((X_i, Y_i)\) and the consistency of \(\hat \varPhi \) (weak as well as strong) is proved. Moreover, the asymptotic normality of this estimator is established under additional conditions on the distribution function \(F\) and the extreme value indices \(\gamma _1, \gamma _2\).