A0321
Title: Copula models for time series extremes
Authors: Pavel Krupskiy - Melbourne University (Australia) [presenting]
Marc Genton - KAUST (Saudi Arabia)
Raphael Huser - King Abdullah University of Science and Technology (Saudi Arabia)
Abstract: Two approaches for building flexible models for time series extremes are considered. The first approach uses Cauchy convolution processes with different kernel functions, and the second approach assumes that there exists a common factor that affects all the realizations of the time series process. We consider some interesting special cases and discuss inference methods for these two classes of models. We apply these models to analyze a wind data set.
