通过R型聚类分析筛选指标,建立了商业银行经营风险预警指标体系.综合运用主成分分析和模糊综合评判方法,建立了基于加权平均模糊综合评判的商业银行经营风险预警模型.以曾在1999年7月发生严重危机的汕头市商业银行为例对风险预警模型进行了实例分析和验证.文章的特色与创新一是运用R型聚类分析剔除了相关性强的指标;二是采用主成分分析法的最大方差旋转来确定指标权重,三是采用梯形隶属度函数反映不同程度风险发生的可能性;四是指标体系的建立综合地考虑了资产风险、流动性风险、资本风险和盈利风险.
A new index system is set up after cluster analysis of R type. And an early warning model of commercial banks' operating risk is set up based on the fuzzy comprehensive appraisal and the principal component analysis. The early warning model is verified by using Shantou Commercial Bank as an example in which a serious crisis took place in July of 1999. The main contributions of this paper lies in four aspects. Firstly, R cluster analysis is used to delete the corelations among indexes. Secondly, the maximum variance of the principal analysis method is used to get the weight of the index. Thirdly, the trapezoid membership function is used to reflect possibility of the different levels of risk. Fourthly, asset risk, liquidity risk, capital risk, and profitability risk are considered in the index system.