CFE 2019: Start Registration
View Submission - CMStatistics
B1925
Title: Bayesian semiparametric regression for heavy-tailed responses Authors:  Junho Lee - University of Edinburgh (United Kingdom) [presenting]
Miguel de Carvalho - CEAUL (Centro de Estatistica e Aplicacoes), Universidade de Lisboa (Portugal)
Abstract: Statistical modelling of extreme events-such as hurricane Dorian-is of the utmost importance for a variety of fields of research. We will propose a Bayesian semiparametric regression model for heavy-tailed responses. The proposed Bayesian smoothing method is motivated by the need of examining how the magnitude of extreme values may change along with a covariate. The model is built using a Pareto-type specification for the tail of the response, with covariates being modelled via generalized additive model (GAM) and Bayesian P-splines. Finally, details on computational implementation will be discussed over the talk, along with a set of illustrations on simulated and real data. We will cover the main results from a Monte Carlo experiment.