Some advances on likelihood-based inference arising from the Peaks Over Threshold method
Abstract: The Peaks Over Threshold (POT) method is the most popular statistical and "longevous" method for the analysis of univariate extremes. Despite this, some basic results relative to the corresponding likelihood-function are still unknown today. After discussing them briefly, we explain what their implications are for the inference and prediction of extreme events stemming from the ordinary to a conditional POT method.
Seminar organized by Prof. Hashorva