摘要
新冠疫情对经济和社会的影响逐步显现,银行业正承受着前所未有的压力。作为经济金融领域的重要一环,银行业对确保不发生系统性金融风险有着重要现实意义。应及时识别风险系数较高的银行,做到提前干预,避免风险传导。本文利用R型聚类模型选取了银行风险评价相关基础指标,利用Q型聚类模型实现了对25家银行风险等级的分类,并进行了模型灵敏度分析,不同函数方法下误差分析显示:R型聚类误差接近于零,Q型聚类中5家银行的风险等级分类结果存在一定误差。以此为基础,通过机器学习算法中的支持向量机模型(SVM)实现了对4家待分类银行的风险确认,并完成了灵敏度及误差分析,模型为及时甄别风险银行提供了参考。
The impact of COVID-19 on the economy and society is gradually emerging,and the banking industry is under unprecedented pressure.As an important part of the economic and financial field,the banking industry has important practical significance to ensure that systemic financial risks do not occur.Banks with higher risk coefficients should be identified in time to intervene in advance to avoid risk transmission.The R-type clustering model is used to select the relevant basic indicators of bank risk evaluation,the Q-type clustering model is used to classify the risk levels of 25 banks,and the model sensitivity analysis is performed.The error analysis under different function methods shows:R-type clustering the class error is close to zero,and there is a certain error in the risk classification results of the five banks in the Q-type cluster.Based on this,the support vector machine model(SVM)in the machine learning algorithm is used to realize the risk confirmation of the four banks to be classified,and the sensitivity and error analysis are completed.The model provides a reference for timely screening of risky banks.
作者
张凤林
Zhang Fenglin(TianJin Branch of the People's Bank of China)
出处
《金融发展评论》
2021年第6期1-13,共13页
Financial Development Review
关键词
银行业
风险分类识别
聚类分析
机器学习
支持向量机
Banking
Risk Classification and Recognition
Clustering Analysis
Machine Learning
Support Vector Machine