计算机工程与应用 ›› 2021, Vol. 57 ›› Issue (13): 239-245.DOI: 10.3778/j.issn.1002-8331.2004-0347

• 工程与应用 • 上一篇    下一篇

考虑社会网络关系的P2P借贷项目违约风险预测

游运,万常选,江腾蛟   

  1. 1.江西财经大学 信息管理学院,南昌 330013
    2.东华理工大学 理学院,南昌 330013
    3.江西财经大学 数据与知识工程江西省高校重点实验室,南昌 330013
  • 出版日期:2021-07-01 发布日期:2021-06-29

Project Default Risk Prediction Considering Social Network in P2P Lending

YOU Yun, WAN Changxuan, JIANG Tengjiao   

  1. 1.School of Information Technology, Jiangxi University of Finance and Economics, Nanchang 330013, China
    2.School of Science, East China University of Technology, Nanchang 330013, China
    3.Jiangxi Key Laboratory of Data and Knowledge Engineering, Jiangxi University of Finance and Economics, Nanchang 330013, China
  • Online:2021-07-01 Published:2021-06-29

摘要:

针对P2P(Peer to Peer)借贷项目违约风险预测中财务信息不完全或质量较低、预测准确率不高等问题,提出了一种考虑平台社会网络关系的P2P借贷项目违约风险预测的方法。通过对P2P借贷平台社会网络相关信息进行分析,从社会资本的结构维度、关系维度和认知维度发掘其中具有风险预测价值的关键特征,即社会网络风险特征,并将这些特征作为预测指标用于违约风险预测,依据多种非线性预测方法分别构建基于传统财务指标预测模型和引入社会网络风险特征后的混合指标预测模型,并对模型的预测结果进行了对比分析。实验结果表明,P2P借贷社会网络关系中蕴含着与借贷项目违约风险显著相关的特征,通过对这些特征进行有效挖掘并将其合理引入P2P借贷项目违约风险预测模型,有助于提高借贷项目违约风险预测效果,为投资者的投资风险规避及P2P借贷市场风险管理提供支持。

关键词: P2P借贷, 社会网络, 违约风险, 风险预测

Abstract:

In view of the fact that financial information is incomplete or of low quality and prediction accuracy of project default risk is low in P2P lending, the paper proposes a method to predict default risk of projects considering social network. By analyzing the social network information of P2P lending platforms, the key characteristics which have risk prediction value, namely social network risk characteristics, have been excavated according to structure dimension, relationship dimension and cognitive dimension of social capital and then take as indicator variables of default risk prediction. Two prediction models of pure financial indicators and mixed indicators with social network risk characteristics are constructed based on a variety of traditional nonlinear methods. The prediction results of prediction models are examined to analyze the value of social network risk characteristics in default risk prediction. Experimental results demonstrate that there are some characteristics significantly associated with default risk of projects in social network. These characteristics have been effectively mined and then reasonably introduced into prediction model of project default risk in P2P lending, which can help improve prediction effect of project default risk, and then provide support for investment risk avoidance of investors and risk management in P2P lending market.

Key words: P2P lending, social network, default risk, risk prediction