基于Logistic模型的个人消费信贷风险评估

Measuring the Risk of Individual Consumption Loan:Toward a Logistic Model

本文随机选取8家商业银行500组个人消费信贷样本,每组设计7个特征指标,首先运用SPSS16.0方法对各组指标进行描述性分析,然后对指标进行显著性检验、多重共线性检验,最后运用L o g i s t i c模型对各特征指标进行回归分析,建立银行个人信用计分模型,并随机选取20组样本数据对该模型进行了测试分析,结果表明,本文建立的个人信用评分模型能很好地评估个人消费信贷风险,可为银行开展个人消费信贷业务提供较好的指导。

This study randomly selected 500 samples of individual consumption loan from eight commercial banks in China and designed 7 feature indexes for each sample. First, SPSS16.0 was applied to implement descriptive analysis for each example. Second, the significance and multiple co-linearity tests were carried out and a logistic model was adopted to complete the regression analysis. Finally, a personal credit scoring model was established and 20 examples were used to test the model. Results show that the personal credit scoring model can accurately measure individual consumption loan risks. So it is argued that the model can provide better guidance for banks to develop consumer loan business.