摘要
2018年6月起发生的P2P网络借贷行业大规模平台倒闭事件引发社会各界关注,识别P2P网络借贷平台的信用风险成为热点问题。构建P2P平台信用风险指标体系,选取当时具有代表性的正常和倒闭平台的指标数据引入二元Logistic回归模型,识别出平台收益率、分散度、流动性和平台业务成就因子对P2P平台产生信用风险有显著的影响作用,最终通过训练BP神经网络得出4个因子识别平台信用风险准确率达到90%。以此为出借人甄选平台提供建议,为网贷行业健康发展提供参考。
Since June 2018, the large-scale closure of online lending industry has attracted attention from all walks of life, and the identification of credit risks of online P2P platforms has become a hot issue. This paper constructs the P2P platform credit risk index system, selecting the representative platform data into the binary Logistic regression model,and identifies that the rate of return, dispersion,liquidity and business achievement factors have significant impact on the credit risk of P2P platforms.Finally, by training BP neural network, the accuracy rate of 4 factors in identifying platform credit risk reached 90%. It provides suggestions for the lender select platform and references for the healthy development of the online P2P industry.
出处
《金融理论与实践》
北大核心
2019年第10期51-58,共8页
Financial Theory and Practice
基金
天津市社科规划项目(TJGL15-010)
