Title: Comparison of classes of generalized Hill estimators
Authors: Frederico Caeiro - NOVA.ID.FCT - Universidade Nova de Lisboa (Portugal) [presenting]
Ivette Gomes - FCiencias.ID, Universidade de Lisboa and CEAUL (Portugal)
Ivanilda Cabral - Universidade de Cabo Verde (Cape Verde)
Abstract: The focus is on the estimation of the extreme value index, the primary parameter of extreme events. For heavy tails, classical extreme value index estimators, such as the Hill estimator, have usually a strong bias. Consequently those estimators are quite sensitive to the number of top order statistics used in the estimation. To improve the mean squared error of the aforementioned estimators, many alternative estimators have appeared in the literature. We analyse several generalizations of the Hill estimator. The aim is to study their non degenerate asymptotic behaviour and to compare them altogether.