Self-Exciting Point Process Models of Civilian Deaths in Iraq

Abstract

Our goal in this paper is to analyze temporal patterns of civilian death reports in Iraq. For this purpose we employ a branching point process model similar to those used in earthquake analysis. Here the rate of events is partitioned into the sum of a Poisson background rate and a self-exciting component in which events trigger an increase in the rate of the process. More speci fically, each event generated by the process in turn generates a sequence of o ffspring events according to a Poisson distribution. Whereas the background rate is typically assumed to be stationary for seismic activity, such an assumption is not valid in the context of civilian deaths in Iraq. We propose three simple adjustments to account for background rate variation and compare the eff ectiveness of each model using Iraq Body Count data from 2003 to 2007. Our results indicate that branching point processes are well suited for modeling the temporal dynamics of violence in Iraq.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Jan 01, 2010
Accession Number
ADA547009

Entities

People

  • Andrea Bertozzi
  • Erik Lewis
  • George Mohler
  • P. J. Brantingham

Organizations

  • University of California, Los Angeles

Tags

Communities of Interest

  • Materials and Manufacturing Processes

DTIC Thesaurus Topics

  • Consistency
  • Crime
  • Criminals
  • Data Analysis
  • Data Sets
  • Death
  • Earthquakes
  • Excitation
  • Mathematical Models
  • Maximum Likelihood Estimation
  • Models
  • Monte Carlo Method
  • New York
  • Statistical Analysis
  • Statistics
  • Step Functions
  • Violence

Readers

  • Distributed Systems and Data Platform Development
  • Graph Algorithms and Convex Optimization.
  • Statistical inference.