基于数据挖掘的P2P网贷个人信用评价模型研究  

Research on the Personal Credit Evaluation Model of P2P Net Loan Based on Data Mining

作　　者：何嘉欣 张涛 陈旭岚 关悦 

机构地区：[1]广西科技大学,广西 柳州

出　　处：《建模与仿真》2021年第4期991-1002,共12页Modeling and Simulation

摘　　要：近年来,随着互联网迅猛发展,P2P网络借贷平台凭借自身门槛低、收益高、操作便捷的特点,进入大众的视野并吸引了较多的借款者与投资者。如何提高、完善P2P网贷平台的风险监控能力,进一步加强对P2P网贷平台的管理,特别是对借款者的信用评价,降低投资者的投资风险是对P2P网贷行业未来发展十分重要的问题。本文针对国外Lending Club官网中给出的数据,对数据先进行预处理,再进行特征选择,最后选择评价指标在研究方法上分别使用传统的分类方法:Logistic回归模型、支持向量机模型和决策树模型。分析结果表明,支持向量机模型对结果预测准确度最高,相比之下,逻辑回归和决策树模型的预测能力相比较差些。在支持向量机模型的基础上,对模型的参数进行调优改进,改进后的模型预测准确度更高,因此该模型适合应用到P2P网贷个人信用评价中。As the Internet develops rapidly in recent years, the P2P network lending platform, with its low threshold, high income and convenient operation, has entered the public view and attracted more borrowers and investors. How to improve and perfect the risk monitoring ability of P2P network loan platform, further strengthen the management of P2P network loan platform, especially the credit evaluation of borrowers, and reduce the investment risk of investors is very important to the future development of P2P network loan industry. This paper aims at the data given in the official website of foreign Lending Club, carries on the characteristic selection to the data advanced, finally selects the evaluation index to use the traditional classification method respectively in the research method: the Logistic regression model, the support vector machine model and the decision tree model. Analysis results show that support vector machine model has the highest prediction accuracy and is suitable for P2P network loan personal credit evaluation model. Based on the SVM model, the parameters of the model are tuning and improved, and the improved model prediction accuracy is higher, so the model is suitable for application to the P2P online loan personal credit evaluation.

关 键 词：P2P网贷 个人信用评价 逻辑回归 支持向量机 决策树 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]