Seminar by Juliana Schulz

Multivariate count models based on comonotonic shocks

Multi-dimensional count data frequently occur in many different fields of study, including risk management, insurance, environmental sciences and many more. In analyzing multivariate data, it is imperative that the underlying modelling assumptions adequately reflect both the marginal behavior as well as the dependence between components. In this work, we focus on constructing flexible multivariate count models based on convolutions of Poisson random vectors. In particular, the model formulation allows for different degrees of dependence by incorporating comonotonic shock vectors in the construction.

The general model framework will be presented, and various estimation techniques will be discussed. Several simulation studies will also be presented, along with a real data application involving extreme rainfall events.