|
Activity Number:
|
80
- Advancement in Spatial and Spatiotemporal Point Process
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Sunday, July 29, 2018 : 4:00 PM to 5:50 PM
|
|
Sponsor:
|
Section on Statistics and the Environment
|
|
Abstract #330767
|
Presentation
|
|
Title:
|
A Computationally Tractable Estimation Procedure for Self-Exciting Spatio-Temporal Point Process Models
|
|
Author(s):
|
James Molyneux* and Frederic Paik Schoenberg
|
|
Companies:
|
UCLA Statistics and UCLA
|
|
Keywords:
|
Self-exciting point process;
Spatial-temporal model fitting;
Statistical computation;
Stoyan-Grabarnik;
Inverse conditional intensity
|
|
Abstract:
|
Fitting procedures for self-exciting spatio-temporal point process models are typically carried out by maximizing the log-likelihood of the conditional intensity function. This method involves numerically approximating an integral term which is computationally intensive and can take an unreasonable amount of time or computing power to fit even moderately sized data sets. We show that using a new estimation procedure, adapted from the Stoyan-Grabarnik diagnostic, avoids the need to approximate the integral term allowing for faster model fitting while producing comparable estimates.
|
Authors who are presenting talks have a * after their name.
