B0989
Title: Markov-switching Hawkes processes for high-frequency trade data
Authors: Ioane Muni Toke - CentraleSupelec (France) [presenting]
Abstract: The focus is on a multidimensional point process defined by multiple Hawkes-like intensities and a switching mechanism based on a hidden Markov chain. Previous works in such a setting assume constant intensities between consecutive events. We extend the model to general piecewise-constant Hawkes excitation kernels and develop an expectation-maximization algorithm for the statistical inference of the parameters of the multiple Hawkes intensities as well as the state transition probabilities. The numerical convergence of the estimators is extensively tested on simulated data. The model is finally applied to equity transaction data on multiple financial markets in an attempt to identify meaningful market states.
