基于BP神经网络组和DS证据理论的信用风险评估算法  

An Assessment Algorithm of Credit Risk Based on BP Neural Network Group and DS Evidential Theory

作　　者：郭英见[1] 吴冲[1] 于天军[1] 

机构地区：[1]哈尔滨工业大学管理学院,哈尔滨150001

出　　处：《合肥工业大学学报（社会科学版）》2009年第5期42-45,共4页Journal of Hefei University of Technology(Social Sciences)

基　　金：国家自然科学基金资助项目(70773029)

摘　　要：结合BP神经网络和DS证据理论,将其有效地结合应用于商业银行的信用评估中。该方法通过对信用风险的输入数据特征进行分类,建立BP网络组,对网络组的输出,建立对于各类信用度的基本概率分配函数,最后利用DS证据理论融合,从而实现信用风险的最终决策。通过实际案例,验证了算法的可行性和有效性。This paper applies the combination of BP neural network and DS evidential theory to the credit assessmen of commercial banks. This method firstly classifies some input data characteristics of credit risk, then BP neural networks groups are set up; the basic probability assignment functions are obtained according to the network group＇s outputs; finally a fusion is achieved with the employment of DS evidential theory; and the final decision will be made. The paper proves the feasibility and validity of the algorithm with the practical cases of credit assessment.

关 键 词：BP神经网络组 信用风险 DS证据理论 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程] TP274[自动化与计算机技术—控制科学与工程]