A1400
Title: Estimation of the extreme value index using generalized probability weighted moments
Authors: Ayana Mateus - NOVA.ID.FCT - Universidade Nova de Lisboa (Portugal) [presenting]
Frederico Caeiro - NOVA.ID.FCT - Universidade Nova de Lisboa (Portugal)
Abstract: In the field of statistics of extremes, precise estimation of the extreme value index is essential for accurate tail inference. This parameter enables the estimation of other tail-related parameters, such as extreme quantiles, which provide critical information for decision-makers in industries ranging from insurance and finance to environmental management and engineering. Pareto-type models are examined within a semi-parametric framework, introducing a new class of estimators based on a generalized probability-weighted moments. The asymptotic distribution of the proposed class of estimators is established, and illustrations based on simulated values are provided.
