定量预测的风险来源与处理方法——以“高烈度政治动荡”预测研究项目的再分析为例
发布时间：2018-11-06 19:41
【摘要】：近年来我国国际关系学界对预测研究的兴趣迅速增长。国际关系预测有其较高的学术创新潜力和政策应用价值,应成为我国国际关系学科的一个重要的研究方向。科学预测基于对现有信息的分析,对未来尚未发生的事件进行判断,是风险较大的研究工作,对于包括国际关系在内的社会科学来说更是如此。本文在社会科学的普遍框架下,探讨了国际关系定量预测的风险来源及其处理方法,并对"高烈度政治动荡"预测研究项目(Political Instability Task Force)进行评述和再分析,修正原研究在风险控制和预测评估方面的缺陷和不足,运用新方法进行多模型平均预测来降低预测中最为棘手的模型风险。通过理论探讨和实例分析,本文旨在强调国际关系定量预测研究对预测风险进行过程控制和结果评估的重要性,并以此管窥现有大量用于处理和评估预测不确定性的定量工具,包括进行变量选择、模型比较、模型平均的多种方法。
[Abstract]:In recent years, there has been a rapid increase in interest in forecasting research in China's international relations community. The prediction of international relations has higher academic innovation potential and policy application value, which should be an important research direction of international relations in China. Based on the analysis of the existing information, scientific prediction is a risky research work, especially for the social sciences, including international relations, to judge the events that have not yet occurred in the future. Under the general framework of social science, this paper probes into the risk sources of quantitative prediction of international relations and its treatment methods, and reviews and reanalyses the prediction project (Political Instability Task Force) of "High intensity political turbulence". In order to reduce the most difficult model risk, the new method is used to reduce the most difficult model risk by correcting the defects and shortcomings of the original study in risk control and prediction evaluation. Through theoretical discussion and case analysis, this paper aims to emphasize the importance of quantitative forecasting research in international relations to process control and result assessment of forecasting risks, and to explore a large number of quantitative tools for dealing with and evaluating forecasting uncertainties. It includes variable selection, model comparison and model averaging.
【作者单位】： 清华大学国际关系学系;
【基金】：教育部国别与区域研究指向性课题“国际安全的大数据研究”的阶段性成果
【分类号】：D81
[Abstract]:In recent years, there has been a rapid increase in interest in forecasting research in China's international relations community. The prediction of international relations has higher academic innovation potential and policy application value, which should be an important research direction of international relations in China. Based on the analysis of the existing information, scientific prediction is a risky research work, especially for the social sciences, including international relations, to judge the events that have not yet occurred in the future. Under the general framework of social science, this paper probes into the risk sources of quantitative prediction of international relations and its treatment methods, and reviews and reanalyses the prediction project (Political Instability Task Force) of "High intensity political turbulence". In order to reduce the most difficult model risk, the new method is used to reduce the most difficult model risk by correcting the defects and shortcomings of the original study in risk control and prediction evaluation. Through theoretical discussion and case analysis, this paper aims to emphasize the importance of quantitative forecasting research in international relations to process control and result assessment of forecasting risks, and to explore a large number of quantitative tools for dealing with and evaluating forecasting uncertainties. It includes variable selection, model comparison and model averaging.
【作者单位】： 清华大学国际关系学系;
【基金】：教育部国别与区域研究指向性课题“国际安全的大数据研究”的阶段性成果
【分类号】：D81
【相似文献】
相关期刊论文 前10条
1 永强;国际关系与“人道主义干预”国际研讨会在京举行[J];国际问题研究;2000年06期
2 陈峰君;21世纪初期亚太国际关系基本考量[J];当代世界;2000年01期
3 贾平安,张树桐;当前国际关系的主要特征[J];重庆社会主义学院学报;2000年02期
4 熊伟;;国际关系网络化[J];世界知识;2000年17期
5 吴永强;《新挑战——国际关系中的“人道主义干预”》简介[J];国际问题研究;2001年01期
6 郭隆隆;新世纪国际关系中的软国力竞争[J];w挛胖芸，
本文编号：2315272
本文链接：https://www.wllwen.com/shekelunwen/waijiao/2315272.html
