Traditionally Hawkes processes are used to model time–continuous pointprocesses with history dependence . Here we propose an extended model where theself–effects are of both excitatory and inhibitory type and follow a GaussianProcess . Efficient approximate Bayesian inference is achieved via dataaugmentation, and we describe a mean–field variational inference approach tolearn the model parameters . To demonstrate the flexibility of the model we apply our methodology on data from three different domains and compare it topreviously reported results. To demonstrate our methodology, we compare it to the previous results of previous work on the Hawkes process to demonstrate its flexibility and learning when data is scarce. To

Author(s) : Noa Malem-Shinitski, Cesar Ojeda, Manfred Opper

Links : PDF - Abstract

Code :
Coursera

Keywords : model - hawkes - process - demonstrate - compare -

Leave a Reply

Your email address will not be published. Required fields are marked *