B1634
Title: Flexible Hawkes Excitation Kernels
Authors: Lekha Patel - Sandia National Laboratories (United States) [presenting]
Abstract: The Hawkes Processes is a popular type of self-exciting point process that has found application in the modeling of financial stock markets, earthquakes, and social media cascades. However, understanding the underlying Hawkes source and pattern of excitation is important for many real-world applications, such as criminal behavior. In this work, we develop a novel Bayesian non-parametric model for a Hawkes process whose excitation kernel is flexibly modeled via a Gaussian Process prior, thereby allowing for different levels of self-excitation. The utility of this model is presented for modeling and predicting extreme terror attacks in Afghanistan.
