HJDM  >> Vol. 5 No. 2 (April 2015)

    基于财务数据的小额信贷决策模型
    The Study of Jot Loan Decision Model Based on Financial Information

  • 全文下载: PDF(351KB) HTML   XML   PP.25-31   DOI: 10.12677/HJDM.2015.52004  
  • 下载量: 1,516  浏览量: 5,391   科研立项经费支持

作者:  

胡淙海,费 宇:云南财经大学,统计与数学学院,云南 昆明;
高正捷:华盛顿大学,数学学院,美国 西雅图

关键词:
小额信贷Logistic回归最小误判代价准则Jot Credit Logistic Regression Minimum Misclassification Cost Criterion

摘要:

本文利用Logistic回归分析不同贷款申请企业的还款概率,再通过最小误判代价准则确定可接受的还款概率,在此基础上构建最小误判代价(minimum misclassification cost)模型及改进的最小误判代价(advanced minimum misclassification cost)模型,通过对XY小额信贷公司的历史贷款数据进行分析,可以发现两个模型相对XY小额信贷公司采用的贷款收入比率(rate of debt and income)模型,有了较大的改进,在降低风险提高收益的同时,还能针对不同风险水平对贷款行为进行一定地调整,并且改进的最小误判代价模型还能对贷款金额进行一定的调整。

In this paper, we use Logistic Regression to analyze the repayment probability of different enter-prises, and then we determine appropriate threshold repayment probability by the minimum cost criterion. On this basis, we build the minimum miscalculation cost model and advanced minimum misclassification cost model. After analyzing the XY jot credit company’s history data, comparing with the rate of debt and income model which is being used by XY Company, we find that these two models have more advantages. They can not only reduce the risk of increase in revenue, but also make adjustment according to different levels of risk behavior; besides, the advanced minimum misclassification cost model can adjust the loan amounts.

文章引用:
胡淙海, 费宇, 高正捷. 基于财务数据的小额信贷决策模型[J]. 数据挖掘, 2015, 5(2): 25-31. http://dx.doi.org/10.12677/HJDM.2015.52004