A Bayesian forecasting model of international conflict. (July 2020)

• Record Type: Journal Article
• Authors: Williford, George W; Atkinson, Douglas B
• Other Names: Baird Ryan G. guest-editor.
• Is Part Of: Journal of defense modeling and simulation. Volume 17:Number 3(2020:Jul.)
• Page Start: 235
• Page End: 242
• Publication Date: 2020-07
• Subjects: International conflict -- forecasting -- prediction -- Bayesian statistics; Military art and science -- Computer simulation -- Periodicals; 355.0011305
• DOI: 10.1177/1548512919827659
• Languages: English
• ISSNs: 1548-5129

Abstract

Scholars and practitioners in international relations have a strong interest in forecasting international conflict. However, due to the complexity of forecasting rare events, existing attempts to predict the onset of international conflict in a cross-national setting have generally had low rates of success. In this paper, we apply Bayesian methods to develop a forecasting model designed to predict the onset of international conflict at the yearly level. We find that this model performs substantially better at producing accurate predictions both in and out of sample.