摘要
针对标准SVM模型在信贷评估中单纯将客户划分为违约或者未违约的不足,提出了利用基于后验概率的SVM用于住房信贷评估的方法.利用商业银行的住房信贷数据进行的实证研究表明,基于后验概率的SVM模型通过将标准SVM的决策值转化为后验概率输出,能够对住房信贷客户的违约概率进行估计,对于商业银行根据客户的违约概率制定相应的信贷政策以及设计相应的住房信贷产品更具有实践意义.
Aim at the deficiency of standard SVM used in credit evaluation merely to classify the clients into default or non-default group, this paper presents a method for individual mortgage loan by using posterior probability SVM model. Using individual mortgage loan data to check the model＇ s applicability, the results indicate that posterior probability SVM can output posterior probability by transforming the decision value of standard SVM to estimate the default probability of the borrowers, which is more significant for commercial banks to make the credit policy and design the relevant mortgage type according to the estimated default probability.
出处
《哈尔滨商业大学学报（自然科学版）》
CAS
2008年第3期377-381,384,共6页
Journal of Harbin University of Commerce：Natural Sciences Edition
基金
哈尔滨工业大学技术.政策.管理(TPM)
国家哲学社科创新基地资助项目(htcsr06t06)
关键词
支持向量机
后验概率
住房贷款
个人信用
信贷评估
support vector machine
posterior probability
individual mortgage loan
personal credit
credit evaluation
