网络大数据信用风险评分能有效预测信贷违约风险吗?

Could Online Credit Score Predict the Risk of Default Effectively?

利用互联网个人小额消费信贷的大样本微观数据,分析借款人互联网信用风险评分与其贷款违约风险的关系。研究结果显示,基于互联网大数据的个人信用风险评分系统,无论是公司自主研发的信用评分卡,还是权威第三方研发的欺诈评分卡,均能够预测网络借款人的违约风险,网络信用风险高的借款人逾期违约率、逾期未结清率更高,需要更多次催收才能最终结清借款。除此之外,借款金额、借款次数等借款特征以及性别、年龄和户籍属性等借款人特征的作用仍不容忽视。利用人工智能和机器学习等先进技术手段更深入地分析借款人信息,完善基于互联网大数据的风控体系,对于降低网络借贷市场的风险至关重要。

Using the large sample micro-data of personal small-scale consumer credit on the Internet,this paper analyzes the relationship between the borrower's Internet credit risk score and its loan default risk.The research results show that the personal credit risk scoring system based on Internet big data can effectively predict the default risk of online borrowers.Borrowers with high network credit risk have overdue default rate and overdue default rate,which requires more collections.Finally settle the loan.In addition,the characteristics of borrowing such as the amount of borrowing,the number of borrowings,and the characteristics of borrowers such as gender,age and household registration attributes cannot be ignored.Using advanced techniques such as artificial intelligence and machine learning to analyze borrower information more deeply and improve the network big data risk control system is essential to reduce the risk of the online lending market.