Data-Driven State Fragility Index Measurement Through Classification Methods

FRONTIERS IN PHYSICS(2022)

引用 0|浏览16
暂无评分
摘要
As environmental changes cause a series of complex issues and unstable situation, exploring the impact of environmental changes is essential for national stability, which is helpful for early warning and provides guidance solutions for a country. The existing mainstream metric of national stability is the Fragile States Index, which includes many indicators such as abstract concepts and qualitative indicators by experts. In addition, these indicators may have preferences and bias because some data sources come from unreliable platforms; it may not reflect the real situation for the current status of countries. In this article, we propose a method based on ensemble learning, named CR, which can be obtained by quantifiable indicators to reflect national stability. Compared with the current mainstream methods, our proposed CR method highlights quantitative factors and reduces qualitative factors, which is an advantage of simplicity and interoperability. The extensive experimental results show a significant improvement over the SOTA methods (7.13% improvement in accuracy, 2.02% improvement in correlation).
更多
查看译文
关键词
data-driven, quantifiable, ranking, state for fragility, classification
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要