基于ANFIS和Elman网络的信用评价研究  

Research of Corporate Credit Scoring Based on ANFIS and Elman Neural Networks

作　　者：梁樑[1] 吴德胜[1] 王志强[1] 熊立[1] 王国华[1] 

机构地区：[1]中国科学技术大学商学院,安徽合肥230026

出　　处：《管理工程学报》2005年第1期69-73,共5页Journal of Industrial Engineering and Engineering Management

基　　金：国家863基金项目(2002AA41361)

摘　　要：BP神经网络用作信用等级分类可取得较好的效果,但在过分要求输出信用分值时效果不佳。针对该缺陷,本文采用自适应神经网络(ANFIS)和Elman网络研究公司信用评分。文中提出了一套甄选方法准则,用于建立适合我国企业的信用评分指标体系;然后依据该指标体系建立了基于Elman网络和ANFIS的信用评估模型;采用V foldCross validation技巧,利用样本公司实际指标数据对该模型的评分效果进行了实证研究。Although strong power is given by BP neural networks to discriminate credit grades,it behaves bad when scoring credit conditions.Then the Adaptive Neural Fuzzy Inference System and Elman neural networks are proposed to score corporate credit conditions in this study.Some methods and guide lines are put forward,so that some distinguished indexes can be chosen.Thus credit scoring models applied to Chinese corporations are set up by Adaptive Neural Fuzzy Inference System(ANFIS) and BP neural networks.Finally,experiential research of the models' scoring power is carried out by use of actual data by V-fold Cross-validation technique.

关 键 词：信用评分 自适应神经模糊推理 ELMAN网络 V-fold Cross-validation技巧 主成分分析 

分 类 号：F272.5[经济管理—企业管理]