基于机器学习方法的个人信用风险评估研究

Research on Personal Credit Risk Assessment Based on Machine Learning Method

互联网金融在国内的兴起,使得个人信贷风险成为许多企业关注的热点。本文通过对LendingClub网站中个人信用贷款数据集的探索,构建基于xgboost和logistic regression组合算法xgboost-LR模型、随机森林和支持向量机算法三种机器学习方法,对个人信用风险进行全面的评估。通过实证数据分析,其中新构建的xgboost-LR算法评价效果最好,能够更加准确地预测个人信用风险。

With the rise of Internet finance in China,personal credit risk has become the focus of many enterprises.Based on the exploration of personal credit loan data set in LendingClub website,this paper constructed three machine learning methods based on XGBoost and Logistics regression combination algorithm xGboost-LR model,random forest and support vector machine algorithm to comprehensively evaluate personal credit risk.Through empirical data analysis,the newly constructed XGBoost-LR algorithm has the best evaluation effect and can predict personal credit risk more accurately.