Bayesian Variable Selection in Generalized Extreme Value Regression: Modeling Annual Maximum Temperature

In many applications, interest focuses on assessing relationships between covariates and the extremes of distribution a continuous response. For example, in climate studies, usual approach to assess change has been based analysis annual maximum data. Using generalized extreme value (GEV) distribution, we can model trends temperature using high number available atmospheric covariates. However, there is typically uncertainty which candidate should be included. Bayesian methods for variable selection are very useful identify important such currently limited moderately dimensional GEV regression. We propose method stochastic search (SSVS) algorithm proposed posterior computation. The applied series three Spanish stations.