Identifying Conditioning Factors and Predictors of Conflict Likelihood for Machine Learning Models: A Literature Review

In this research, we focused on armed conflicts and related violence. The study reviewed the use of machine learning to predict the likelihood of conflict escalation and the role of conditioning factors. The results showed that machine learning and predictive models could help identify conflict-prone locations and geospatial factors contributing to conflict escalation. The study found 46 relevant papers and emphasized the importance of considering unique predictors and conditioning factors for each conflict. It was found that the conflict susceptibility of a region is influenced principally by its socioeconomic conditions and its political/governance factors. We concluded that machine learning has the potential to be a valuable tool in conflict analysis and, therefore, it can be an asset in conflict mitigation and prevention, but the accuracy of the models depends on data quality and the careful selection of conditioning factors. Future research should aim to refine the methodology for more accurate prediction of the models.

Automated summary

This research focused on armed conflicts and related violence, reviewing the use of machine learning to predict conflict escalation and the role of conditioning factors. The results showed that machine learning can help identify conflict-prone locations and contributing geospatial factors. The study emphasized the importance of unique predictors and conditioning factors for each conflict and concluded that machine learning has the potential to be a valuable tool in conflict analysis and prevention.