国际货币基金组织：2024利用卫星数据改进东加勒比社会援助目标报告（英文版）

Prioritizing populations most in need of social assistance is an important policy decision. In the Eastern Caribbean, social assistance targeting is constrained by limited data and the need for rapid support in times of large economic and natural disaster shocks. We leverage recent advances in machine learning and satellite imagery processing to propose an implementable strategy in the face of these constraints. We show that local well-being can be predicted with high accuracy in the Eastern