Inference for marked Hawkes processes

Marked Hawkes processes are point process models incorporating a certain branching and clustering structure, and are useful for modelling many kinds of data, for example times of earthquakes. Efficient and reliable methods for statistical inference are important for the practical use of such models. In this project, Markov chain Monte Carlo algorithms for Bayesian inference exploiting the branching structure are investigated. Also other practical aspects of inference for Hawkes processes, such as various kinds of missing data, are explored.

Supported by Danish Natural Science Research Council