| 1. |
- Hegre, Håvard, 1964-, et al.
(författare)
-
Lessons From an Escalation Prediction Competition
- 2022
-
Ingår i: International Interactions. - : Taylor & Francis. - 0305-0629 .- 1547-7444. ; 48:4, s. 521-554
-
Tidskriftsartikel (refereegranskat)abstract
- Recent research on the forecasting of violence has mostly focused on predicting the presence or absence of conflict in a given location, while much less attention has been paid to predicting changes in violence. We organized a prediction competition to forecast changes in state-based violence both for the true future and for a test partition. We received contributions from 15 international teams. The models leverage new insight on the targeted problem, insisting on methodological advances, new data and features, and innovative frameworks which contribute to the research frontiers from various perspectives. This article introduces the competition, presents the main innovations fostered by the teams and discusses ways to further expand and improve upon this wisdom of the crowd. We show that an optimal modeling approach builds on a good number of the presented contributions and new evaluation metrics are needed to capture substantial models' improvements and reward unique insights. La investigacion reciente sobre la prevision de la violencia se ha centrado principalmente en predecir la presencia o ausencia de conflictos en un determinado lugar, mientras que se ha prestado mucha menos atencion a predecir los cambios en la violencia. Organizamos una competencia de prediccion para predecir los cambios en la violencia estatal tanto para el futuro cierto como para una division del analisis. Recibimos aportes de quince equipos internacionales. Los modelos aprovechan las nuevas ideas sobre el problema especifico insistiendo en los avances metodologicos, los nuevos datos y caracteristicas, asi como en los marcos innovadores que contribuyen a las fronteras de la investigacion desde diversas perspectivas. Este articulo presenta la competencia y las principales innovaciones que los equipos fomentan, y analiza maneras de expandirse y mejorar aun mas a partir de esta sabiduria del publico. Mostramos que un enfoque de modelacion optimo se crea a partir de un buen numero de aportes presentados y que se necesitan nuevas metricas de evaluacion para capturar las mejoras considerables de los modelos y para premiar las ideas unicas. Les recherches recentes sur la prevision de la violence se sont principalement concentrees sur la prediction de la presence ou de l'absence de conflit dans un lieu donne, alors que beaucoup moins d'attention a ete accordee a la prediction des evolutions de la violence. Nous avons organise un concours de predictions dont l'objectif etait de prevoir les evolutions de la violence etatique a la fois pour le futur reel et pour une partition test. Nous avons recu des contributions de 15 equipes internationales. Les modeles concernes tirent profit de nouveaux renseignements sur le probleme cible en insistant sur les progres methodologiques, sur de nouvelles donnees et caracteristiques et sur des cadres innovants contribuant a elargir les frontieres des recherches de divers points de vue. Cet article presente le concours et les principales innovations proposees par les equipes et aborde les moyens d'etendre et d'ameliorer cette sagesse de la foule. Nous montrons qu'une approche optimale de la modelisation repose sur bon nombre des contributions presentees et que de nouvelles metriques d'evaluation sont necessaires pour saisir les ameliorations substantielles des modeles et recompenser les idees uniques.
|
|
| 2. |
- Hegre, Håvard, 1964-, et al.
(författare)
-
ViEWS : A political violence early-warning system
- 2019
-
Ingår i: Journal of Peace Research. - : SAGE Publications. - 0022-3433 .- 1460-3578. ; 56:2, s. 155-174
-
Tidskriftsartikel (refereegranskat)abstract
- This article presents ViEWS – a political violence early-warning system that seeks to be maximally transparent, publicly available, and have uniform coverage, and sketches the methodological innovations required to achieve these objectives. ViEWS produces monthly forecasts at the country and subnational level for 36 months into the future and all three UCDP types of organized violence: state-based conflict, non-state conflict, and one-sided violence in Africa. The article presents the methodology and data behind these forecasts, evaluates their predictive performance, provides selected forecasts for October 2018 through October 2021, and indicates future extensions. ViEWS is built as an ensemble of constituent models designed to optimize its predictions. Each of these represents a theme that the conflict research literature suggests is relevant, or implements a specific statistical/machine-learning approach. Current forecasts indicate a persistence of conflict in regions in Africa with a recent history of political violence but also alert to new conflicts such as in Southern Cameroon and Northern Mozambique. The subsequent evaluation additionally shows that ViEWS is able to accurately capture the long-term behavior of established political violence, as well as diffusion processes such as the spread of violence in Cameroon. The performance demonstrated here indicates that ViEWS can be a useful complement to non-public conflict-warning systems, and also serves as a reference against which future improvements can be evaluated.
|
|
| 3. |
- Hegre, Håvard, 1964-, et al.
(författare)
-
ViEWS(2020) : Revising and evaluating the ViEWS political Violence Early-Warning System
- 2021
-
Ingår i: Journal of Peace Research. - : Sage Publications. - 0022-3433 .- 1460-3578. ; 58:3, s. 599-611
-
Tidskriftsartikel (refereegranskat)abstract
- This article presents an update to the ViEWS political Violence Early-Warning System. This update introduces (1) a new infrastructure for training, evaluating, and weighting models that allows us to more optimally combine constituent models into ensembles, and (2) a number of new forecasting models that contribute to improve overall performance, in particular with respect to effectively classifying high- and low-risk cases. Our improved evaluation procedures allow us to develop models that specialize in either the immediate or the more distant future. We also present a formal, 'retrospective' evaluation of how well ViEWS has done since we started publishing our forecasts from July 2018 up to December 2019. Our metrics show that ViEWS is performing well when compared to previous out-of-sample forecasts for the 2015-17 period. Finally, we present our new forecasts for the January 2020-December 2022 period. We continue to predict a near-constant situation of conflict in Nigeria, Somalia, and DRC, but see some signs of decreased risk in Cameroon and Mozambique.
|
|
| 4. |
- Vesco, Paola, 1990-, et al.
(författare)
-
United they stand : findings from an escalation prediction competition
- 2022
-
Ingår i: International Interactions. - : Taylor & Francis. - 0305-0629 .- 1547-7444. ; 48:4, s. 860-896
-
Tidskriftsartikel (refereegranskat)abstract
- This article presents results and lessons learned from a prediction competition by ViEWS to improve collective scientific knowledge on forecasting (de-)escalation on the African continent. The competition call asked participants to forecast changes in state-based violence for the true future (October 2020 – March 2021) as well as for a held-out test partition. An external scoring committee, independent from both the organizers and participants, was formed to evaluate the models based on both qualitative and quanti- tative criteria, including performance, novelty, uniqueness and replicability. All models contributed to advance the research frontier by providing novel methodological or theo- retical insight, including new data, or adopting innovative model specifications. While we discuss several facets of the competition that could be improved moving forward, the collection passes an important test. When we build a simple ensemble prediction model – which draws on the unique insights of each contribution to differing degrees – we can measure an improvement in the prediction from the group, over and above what the average individual model can achieve. This wisdom of the crowd effect suggests that future competitions that build on both the successes and failures of ours, can contribute to scientific knowledge by incentivising diverse contributions as well as focusing a group’s attention on a common problem.
|
|
