2017/06/11 by Massil Achab, Emmanuel Bacry, Achab, Massil +5
Mathematics · #Point processes and geometric inequalities
paper · pdf · doi:10.48550/arxiv.1706.03411
We introduce a new non parametric method that allows for a direct, fast and\nefficient estimation of the matrix of kernel norms of a multivariate Hawkes\nprocess, also called branching ratio matrix. We demonstrate the capabilities of\nthis method by applying it to high-frequency order book data from the EUREX\nexchange. We show that it is able to uncover (or recover) various relationships\nbetween all the first level order book events associated with some asset when\nmapped to a 12-dimensional process. We then scale up the model so as to account\nfor events on two assets simultaneously and we discuss the joint high-frequency\ndynamics.\n