基于PCA-RBF神经网络的商业银行财务风险评价研究

Evaluation of Commercial Bank Financial Risk Based on the Method of Principal Component Analysis-RBF Neural Network

为了有效地防范、化解商业银行的财务风险,本文设计了基于主成分分析和RBF神经网络相结合的财务风险综合评价方法。本文首先利用因子分析法对各评级指标进行提炼,从而减少了RBF神经网络训练数据,简化了网络结构,然后选择提炼后的部分商业银行数据利用RBF神经网络进行了训练,最后选择几家商业银行的数据对已经训练好的RBF-ANN系统进行了测试。本文的研究为商业银行财务风险评价、财务风险监控和财务风险控制提供了新的思路和方法。

In order to prevent the financial risk of commercial bank effectively, this paper designs an intelligent financial risk evaluation method based on the principal component analysis and RBF neural network. Firstly, this paper uses method of fac- tor analysis to refine the evaluation index, hence reduces the RBF neural network training data, simplifies network structure. Secondly, it trains RBF neural network with part of the sample refined data.. It tests the trained RBF-ANN system with data of the sample commercial bank. Thus a new thought and method has been provided for the commercial bank financial risk evalua- tion, financial risk monitor and financial risk control.