CMStatistics 2021: Start Registration
View Submission - CMStatistics
Title: Bayesian non-asymptotic extreme value hierarchical models Authors:  Enrico Zorzetto - Princeton University (United States)
Antonio Canale - University of Padua (Italy) [presenting]
Marco Marani - University of Padova (Italy)
Abstract: A general Bayesian hierarchical model is introduced for estimating the probability distribution of extreme values of intermittent random sequences. Our approach avoids the asymptotic assumption typical of the traditional extreme value theory, and accounts for the possible underlying variability in the distribution of event magnitudes and occurrences, which are described through a latent temporal process. Focusing on daily rainfall extremes, the structure of the proposed model lends itself to incorporating a prior geo-physical understanding of the rainfall process. Empirical performance is illustrated showing less tendency to overfitting and better out-of-sample predictions. Spatio-temporal extensions of the model are also discussed.