基于数据挖掘的信用卡申请客户信用评价问题研究  

Research on Credit Evaluation of Credit Card Application based on data Mining

作　　者：郭龙飞 严广乐[1] 

机构地区：[1]上海理工大学管理学院

出　　处：《经济研究导刊》2018年第5期110-113,共4页Economic Research Guide

基　　金："上海高原学科建设项目管理科学与工程"(10-17-303-004)

摘　　要：为了控制和防范信贷风险,银行必须对申请人进行有效的信用评估。通过传统的Logistic回归与随机森林模型,分别建立对应的信用评分模型,来比较两种模型的优缺点以达到最佳的预测效果,从而有效减小银行信贷业务的风险,更好地实现银行的利润最大化。In order to control and prevent the credit risk,the bank must carry on the effective credit evaluation to the applicant.Through thetraditional Logistic regression and the stochastic forest model.The corresponding credit scoring models are established to compare the advantagesand disadvantages of the two models in order to achieve the best prediction effect,so as to effectively reduce the risk of bank creditbusiness,and better realize the profit maximization of the bank.

关 键 词：信用等级划分 数据拟合 LOGISTIC回归 随机森林 

分 类 号：F830.48[经济管理—金融学]