摘要
将最小总风险准则MOR与贝叶斯网络分类器相结合,提出了一种新型信用评估模型。在两个真实数据集上以MOR用10层交叉验证对贝叶斯网络信用评估模型进行了测试,并与最小错误概率准则MPE的贝叶斯网络分类器的结果进行了对比。结果表明,基于MOR的贝叶斯网络分类模型可以有效地减小信用评估风险。
This paper integrated NOR (minimum overall risk rule) into Bayesian network classifiers, and proposed new credit scoring models. According to MOR, they were tested using 10-fold cross validation with two real world data sets, and compared with Bayesian network classifier based on MPE. Results demonstrate that the Bayesian network elassifiers based on MOR are able to reduce effectively the eredit seoring risk.
出处
《计算机应用研究》
CSCD
北大核心
2009年第1期50-53,58,共5页
Application Research of Computers
基金
国家自然科学基金资助项目(70371026)
四川省教育厅科研项目(2006C082)
关键词
个人信用评估
最小总风险准则
最小错误概率准则
贝叶斯网络分类器
consumer credit scoring
minimum overall risk rule (MOR)
minimum probability of error rule ( MPE )
Bayesian network classifiers