摘要
采用支持向量机和核主成分分析法构建中国银行业系统性风险预警模型,将预警结果与BP神经网络模型和Logit回归模型的预警结果进行对比,并基于2008年1月~2017年9月的数据,采用SVM预警模型预测2009年1月~2018年9月中国银行业系统性风险水平。研究结果显示:与BP神经网络和Logit回归模型相比,SVM模型具有较高的预警正确率;在不同的阶段中国银行业系统性风险水平呈现出不同的变动趋势。建议中国政府部门和银行业警惕资本市场泡沫增长等隐性风险,不断完善银行业内部系统的风险防控机制,持续强化银行业宏观审慎监管。
The paper applies support vector machines( SVM) and kernel principal component analysis to construct a banking systemic risk prediction model in China,and compares the result with that of BP neural networks and Logit regression. Then,the paper predicts China' s banking systemic risk during January 2009 to September 2018 by using the SVM prediction model based on the data from January 2008 to September 2017. The research demonstrates that: by comparison,the SVM model has higher prediction accuracy than BP neural networks and Logit regression; the China' s banking systemic risk shows different trends at different stages. It is suggested that Chinese government and banking industry should be alert to various hidden risks such as bubble growth of capital market,continuously improve the prevention and control mechanism of financial risk and constantly strengthen the macro-prudential supervision in banking industry.
引文
[1]陈秋玲,薛玉春,肖璐.金融风险预警:评价指标、预警机制与实证研究[J].上海大学学报(社会科学版),2009(5):127-144.
[2]胡海青,张琅,张道宏.供应链金融视角下的中小企业信用风险评估研究---基于SVM与BP神经网络的比较研究[J].管理评论,2012(11):70-80.
[3]苏治,卢曼,李德轩.深度学习的金融实证应用:动态、贡献与展望[J].金融研究,2017(5):111-126.
[4]许传华,徐慧玲,杨雪莱.我国金融风险预警模型的建立与实证研究[J].经济问题,2012(2):83-86.
[5]肖斌卿,颜建晔,杨旸,等.金融安全预警系统的建模与实证研究:基于中国数据的检验[J].国际商务---对外经济贸易大学学报,2015(6):97-106.
[6]杨霞,吴林.我国银行业系统性风险预警研究[J].统计与决策,2015(10):147-150.
[7]AHN J J,OH K J,KIM T Y,et al.Usefulness of support vector machine to develop an early warning system for financial crisis[J].Expert Systems with Applications,2011(4):2966-2973.
[8]CORTES C,VAPNIK V.Support-vector networks[J].Machine Learning,1995(3):273-297.
[9]CAVALCANTE R C,BRASILEIRO R C,SOUZA V L F,et al.Computational intelligence and financial markets:a survey and future directions[J].Expert Systems with Applications,2016:194-211.
[10]FIORAMANTI M.Predicting sovereign debt crises using artificial neural networks:a comparative approach[J].Journal of Financial Stability,2008(2):149-164.
[11]FRANKEL J A,ROSE A K.Currency crashes in emerging markets:an empirical treatment[J].Journal of International Economics,1996(3):351-366.
[12]ITURRIAGA F J L,SANZ I P.Bankruptcy visualization and prediction using neural networks:a study of US commercial banks[J].Expert Systems with applications,2015(6):2857-2869.
[13]KAMINSKY G,LIZONDA S,REINHART C M.Leading indicators of currency crises[J].Staff Papers,1998(1):1-48.
[14]KIM T Y,HWANG C,LEE J.Korean economic condition indicator using a neural network trained on the 1997crisis[J].Journal of Data Science,2004(4):371-381.
[15]LI S,WANG M,HE J.Prediction of banking systemic risk based on support vector machine[J].Mathematical Problems in Engineering,2013.
[16]QIAN H,MAO Y,XIANG W,et al.Recognition of human activities using SVM multi-class classifier[J].Pattern Recognition Letters,2010(2):100-111.
[17]SACHS J,TORNELL A,VELASCO A.The Mexican peso crisis:sudden death or death foretold?[J].Journal of International Economics,1996,41(3-4):265-283.
[18]SHIN K S,LEE T S,KIM H.An application of support vector machines in bankruptcy prediction model[J].Expert Systems with Applications,2005(1):127-135.
[19]YU L,WANG S,LAI K K et al.A multiscale neural network learning paradigm for financial crisis forecasting[J].Neurocomputing,2010(4):716-725.