CMStatistics 2017: Start Registration
View Submission - CMStatistics
B0696
Title: Regularly varying time series and max-stable processes Authors:  Anja Janssen - KTH Royal Institue of Technology (Sweden) [presenting]
Abstract: Some connections between stationary multivariate regularly varying time series and the class of max-stable processes in discrete time will be discussed. We focus on the extremal aspects of both processes which can be expressed in terms of a limiting process that is known as the (spectral) tail process, which describes the conditional behavior of the process before and after an extremal event at a given time. We explore how the stationarity of the processes is reflected in the extremal behavior and how it is for example possible to construct an underlying stationary max-stable process for a given spectral tail process.