This session was part of the 2021 Emerging Technologies in Peacebuilding and Prevention Virtual Practitioners Workshop hosted by the NYU Center on International Cooperation on December 1-2, 2021.
Session title: “Early-Warning Systems for Political Violence: How Should We Define Our Prediction Targets?”
Speaker: VIEWS Director Håvard Hegre
Early warning of armed conflict is necessary to facilitate early action. Stakeholders seeking to prevent a humanitarian disaster, or preparing to assist in one, will be better able to do so if they have an early indication of where major conflict is likely to occur, as well as what the likely intensity and duration of this fighting will be. A number of quantitative early-warning systems have recently been developed to this end and are running with regular updates of their risk assessments. To date, all of these systems – including ViEWS – have however been limited to binary (conflict/no conflict) forecasts, many of them further restricted to single-period forecasts for high-intensity violence only.
Hoping to inspire humanitarian actors and peace-builders to make better use of predictive analytics in their work, this lightning talk presents an alternative way forward: the `continuous multi-period incidence target’, or CMPI for short. It introduce this new prediction target and elaborate on its many uses for the international community. The discussion is based on the the lessons learnt from developing the political Violence Early-Warning System (ViEWS, https://views.pcr.uu.se) at Uppsala University and Peace Research Institute Oslo.