Online Program Home
My Program

Abstract Details

Activity Number: 3 - Recent Developments in Network Testing
Type: Invited
Date/Time: Sunday, July 28, 2019 : 2:00 PM to 3:50 PM
Sponsor: IMS
Abstract #300495 Presentation
Title: Change Point Detection for Self-Exciting Point Processes
Author(s): Daren Wang* and Rebecca Willett
Companies: University of Chicago and University of Chicago
Keywords: Self-Exciting Point Processes; High dimensional change point

We consider the problem of change point detection for Self-Exciting Point Processes (SEPP). SEPPs are commonly used to model data from biological neural networks, crime event data across multiple regions, and a variety of other scientific studies. In these scenarios, we observe a collection of discrete events at each time t, and we assume that the distribution of future events only depends on past events. When the data-generating mechanism of the SEPP is stable over time, the estimation of the self-excitation parameters is well-studied. In this talk, we consider the setting in which the self-excitation parameters are unstable and change over time in a piecewise constant manner. Within a general high-dimensional change point framework, we will develop algorithms that accurately can estimate the locations of the change points. Our approach is based on a novel variant of the fused LASSO combined with multivariate binary segmentation of the regression coefficients. We provide sharp theoretical bounds as well as real data examples to justify our findings.

Authors who are presenting talks have a * after their name.

Back to the full JSM 2019 program