Wednesday, April 5, 2017 (1:00 p.m. in Yost 306)

Title: Dependency measures and copulas for multidimensional infinitely divisible distributions

Speaker: Wesley Maddox (Case Western Reserve University)

Advisor: Wojbor Woyczynski (Professor, Case Western Reserve University)

Abstract: Multivariate Linnik distributions, despite their potential usefulness, have not been well-studied due to their difficulty in finding representations and estimators. Parameter estimation methods are applied to the problem of inference in multivariate Linnik distributions. Multivariate Linnik distributions are then applied to develop an elliptical Linnik copula, while dependency experiments are performed on this copula. A parametric inference method is also given for the elliptical Linnik copula. Finally, simulation and dependency is studied for a pathological bivariate distribution with stable marginals but which is not stable.