基于支持向量机的大学生助学贷款个人信用评价  

Personal credit evaluation of college student loans with support vector machines

作　　者：肖智[1] 王明恺[1] 谢林林[1] 

机构地区：[1]重庆大学经济与工商管理学院,重庆400030

出　　处：《清华大学学报（自然科学版）》2006年第z1期1120-1124,共5页Journal of Tsinghua University(Science and Technology)

基　　金：重庆市自然基金资助项目(CSTC.2004BB2183)

摘　　要：目前国内对大学生助学贷款个人信用的研究定性分析居多,很少运用定量的方法建立分析预测模型。该文在分析传统的信用评价模型优缺点的基础上,发现支持向量机方法(SVM)在评价贷款大学生个人信用应用时具有一定的优越性,试探性地运用支持向量机方法建立大学生助学贷款个人信用评价分析模型。通过实证分析获得了较高的预测准确率,并将分析结果与AHP、BPNN方法进行了比较,体现了SVM方法的相对优越性。因此,用支持向量机方法来评价贷款大学生个人信用是可行的、有效的。At present,there are many domestic studies on credit analysis of student loans,but these studies are qualitative researches basically,seldom using quantitative methods to set up analysis models.This paper discovers that it is advantageous to evaluate personal credit of student loans with support vector machines(SVM) based on the comparative analysis of advantage and disadvantage of traditional models in credit evaluation,and probingly uses the method of support vector machines to set up credit evaluation and analysis model of college student loans.Finally,this paper obtains high correct rate of classification through an example.The analysis results are compared with those from AHP and BPNN methods to show the relative advantage of SVM.Thus,the results show that it is feasible and efficient to evaluate personal credit of student loans with SVM.

关 键 词：支持向量机 大学生 助学贷款 个人信用 信用评价 

分 类 号：TP301[自动化与计算机技术—计算机系统结构] F832.4[自动化与计算机技术—计算机科学与技术]