Spatial prediction of armed conflicts from the perspective of political geography using bivariate frequency ratio method (FR) in East African States.

Armed conflicts, as significant human phenomena, profoundly impact populations and reflect a state's capacity to fulfill its responsibilities. These conflicts arise from various causes, necessitating robust predictive models to understand their spatial distribution. This study employs the Bivariate Frequency Ratio (FR) method to spatially predict the occurrence of armed conflicts across the East African States, drawing on 42 political geography-related criteria. The development of the predictive model involved classifying the region into five conflict-prone categories influenced by critical political geography factors. Geospatial datasets, curated in a GIS environment, were sourced from approved online portals. The findings indicate that Burundi exhibits the highest vulnerability to armed conflict, followed closely by Rwanda, Uganda, and Somalia. Ethiopia and South Sudan show a moderate risk, while predictions for Zimbabwe, Zambia, and Mozambique suggest lower likelihoods of conflict. The model's accuracy was validated using the Receiver Operating Characteristic (ROC) curve, demonstrating its effectiveness. Furthermore, the model's applicability extends to other regions, offering a valuable tool for global conflict prediction.