CFE 2019: Start Registration
View Submission - CMStatistics
Title: Vine models for stationary time series Authors:  Thomas Nagler - Leiden University (Germany) [presenting]
Aleksey Min - Technische Universitaet Muenchen (Germany)
Abstract: Vine copulas can be used to simultaneously model inter-serial and cross-sectional dependence in stationary time series. We introduce a class of vine structures that is particularly suited for such applications. The class is characterized in a way that stationarity is guaranteed if pair-copulas are invariant to shifts in the time index (which is a necessary condition in any case). The class includes several structures that were already proposed in the literature.