A0226
Title: General graphical models for stable processes
Authors: Florian Brueck - University of Geneva (Switzerland) [presenting]
Abstract: General graphical models are introduced for multivariate Levy processes with stable margins. First, Huesler-Reiss Levy processes are defined in terms of a Levy measure, which is derived from the exponent measure of a Huesler-Reiss distribution. The Huesler-Reiss Levy processes can be equipped with arbitrary alpha-stable margins, and their conditional independence structure is encoded in a single d-times-d matrix. Furthermore, it is shown how the conditional independence structure of a Huesler-Reiss Levy process may be estimated, and a simulation scheme is provided to generate trajectories of the corresponding paths.
