Stepping into the Future : the next Generation of Crisis Forecasting Models

Content Provider	Semantic Scholar
Author	Ward, Michael D. Dial Metternich, Nils W. Carrington, Christopher Dorff, Cassy L. Gallop, Max Hollenbach, Florian Max Benjamin Schultz, Anna Charlotte Weschle, Simon
Copyright Year	2012
Abstract	Developing political forecasting models is not only relevant for scientific advancement, but also increases the ability of political scientists to inform public policy decisions. Taking this perspective seriously, the International Crisis Early Warning System (ICEWS) was developed under a DARPA initiative to provide predictions of international crisis, domestic crisis, rebellion, insurgency, and ethnic violence (Events of Interest/EOIs) in about two-dozen countries in the US PACOM Area of Responsibility. As part of a larger project coordinated by Lockheed Martin Advanced Technology Labs, a team at Duke University created a series of geographically informed statistical models for these EOIs. The generated predictions have been highly accurate, with few false negative and positive categorizations. Predictions are made at the monthly level for three months periods into the future. The major variables to generate the predictions include 1) event data culled from FACTIVA reports, 2) structural political characteristics of states, 3) economic characteristics of states, and 4) contextual features of each country. These later characteristics take into account the social-spatial context of each individual country, thereby allowing the models to escape the limitations of treating each country as independent from the influence of events and forces in nearby countries. For each EOI we present a separate prediction model, which captures the unique dynamics of each outcome. Each of these models has a high degree of accuracy in reproducing historical data measured monthly over the past 10 years, and is approximately equally accurate in making three-month forecasts
File Format	PDF HTM / HTML
Alternate Webpage(s)	http://www.guillaumenicaise.com/wp-content/uploads/2013/10/Crisis-Forecasting-Models.pdf
Language	English
Access Restriction	Open
Content Type	Text
Resource Type	Article