Michael Höhle: Spatial, Temporal and Spatio-temporal Statistical Modelling of Clustered Events

This talk is about the statistical analysis of point patterns which are realizations of stochastic processes exhibiting clustering behaviour. The underlying stochastic model could, e.g., be an inhomogeneous Poisson process, Hawkes' self-exciting process or Cox process. Data are now available in various temporal and spatial aggregated forms and the goal is to infer model parameters. To motivate such modelling I will illustrate application using infectious disease occurrence data, location of unexploded WWII bombs and the adjustment of reporting delays in routine reporting data.

The talk can be seen as the first class in an intermediate/advanced course about statistical inference for stochastic processes, but also aims at introducing some of my research in a not too technical way.