基于k-means和SVM的银行个人信用评估应用  

Application of bank personal credit rating model based on k-means and SVM

作　　者：戴德宝[1] 倪立平 薛铭[1] 

机构地区：[1]上海大学管理学院,上海200444

出　　处：《江苏科技大学学报（自然科学版）》2017年第6期836-842,共7页Journal of Jiangsu University of Science and Technology:Natural Science Edition

基　　金：上海市教委科研创新项目(13YS015)

摘　　要：现有的个人信用评估模型大多数是解决信用二分类问题,而信用等级划分需要进一步研究.文中以SVM为基础,结合k-means聚类方法,假设具有相似特征的客户拥有相同信用情况,提出了k-means和SVM结合的个人信用评估模型,使之不仅能对个人进行二分类划分,且能将客户划分为不同的信用等级.实验结果表明,与其他模型相比,提出的模型二分类精度较高,并且能得到个人的信用等级,具有较高的实用价值.Most of the existing personal credit rating models are built to solve two-classification problems. And the credit rating division needs further research. This paper is based on the SVM with the clustering method,kmeans. Assuming that the customers who have the similar characteristics have the same credit rating,the personal credit rating model based on k-means and SVM is proposed and it can divide the customers not only into two classifications but also into different credit ratings. The experimental results show that compared with other models,two classification accuracy of the proposed model is higher and it can get a personal credit rating,which has a higher practical value.

关 键 词：信用评估 支持向量机 K-MEANS 信用等级 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]