基于RSBP的商业银行个人信用风险测评的实证研究

A Study on Personal Credit Evaluation in Chinese Commercial Banks Based on RSBP

介绍和分析了目前国内外个人信用风险评估研究的现状和不足,建立了一套包含工商、税务、法院等信息的评估指标体系。构建了基于RSBP的我国商业银行个人信用风险评估模型,利用粗糙集理论对属性进行约简,并把约简结果输入BP神经网络,从而缩短了网络训练时间,提高了网络收敛速度和测试精度。通过比较基于BP神经网络和RSBP的评价结果证明了该模型的优越性。

Aiming at the issue of personal credit risks in commercial banks in our country,this paper analyzes current researches on personal credit evaluation and establishes an evaluating index system which includes the information of industry and commerce,taxation,legal issues,etc.It presents an approach of BP neural network with rough sets theory for personal credit evaluating.Through attribute reduction based on variable precision with rough sets,the simulation results show that the training time is decreased,the convergence speed and testing accuracy are improved.Finally,the performance of RSBP is tested by comparing its evaluations with that of BP neural network.