摘要
本文以供给侧结构性改革期间潜在的系统性金融风险为研究对象,通过FSI方法处理2002. 01～2016.12期间金融市场数据的结果作为样本集,运用4种核函数SVM模型,Logit回归,DDA以及BPNN模型来构建预警模型,并采用F1-Score和AUC对预警模型四个时期预测结果进行对比分析。实证结果表明,多项式核函数SVM预警模型不仅拥有优越的学习和预测能力,同时能够提前捕捉到供给侧结构性改革期间的系统性金融风险信号,进而能为金融风险管理部门防范系统性金融风险,保障供给侧结构性改革顺利推进提供有力的模型工具。
This paper studies the underlying systemic financial risks during the period of structural reform of the supply front, adopting FSI measures to deal with the finaneial market data during period 2002.01 ~ 2016.12 and establish SVM warning model with four kernel funetion and Logit regression, DDA and BPNN warning model. Subsequently, the thesis analyzes their predieting performanees during four predieting periods through F1-Seore and AUC evaluation. The results indieate that SVM warning model with polynomial kernel funetion outperforms other models and has exeellent aeeuraey to prediet the systemie finaneial risks during the period of struetural reform of the supply front. Therefore, the finaneial risk management seetor ean adopt this model as an effeetive deviee whieh ean prevent the oeeurrenee of systemie finaneial risks and simultaneously make struetural reform of the supply front sueeessful proeeeding.
作者
淳伟德
肖杨
CHUN Wei-de;XIAO Yang(School of Business,Chengdu University of Technology,Chengdu 610059,China)
出处
《预测》
CSSCI
北大核心
2018年第5期36-42,共7页
Forecasting
基金
国家社会科学基金资助项目(17BJY188)
国家自然科学基金资助项目(71771032)
教育部人文社会科学青年基金资助项目(17YJC790168)
四川省软科学计划资助项目(2016ZR0137
2017JY0158
2017ZR0204
2017ZR0205)
