论文部分内容阅读
近年来,互联网金融在国内发展迅速,各种类型的小额网贷平台更是层出不穷,但是随着平台跑路,欺诈等各种问题的涌现,暴露出对小额网贷平台监管的不足。本文以C5.0决策树算法为核心建立起风险监控模型,首先利用大数据技术抓取平台数据、公开信息数据和征信机构数据,然后对数据分析,筛选关键指标体系,建立了一种风控模型,最后评估和验证模型及算法的有效性,并通过对模型的参数进行调整,提出以保证整体错误率的前提下,尽可能的降低α错误率的评价标准。通过实验发现,该风控模型对问题平台有着很好的预测能力。
In recent years, Internet finance has been developing rapidly in China. Various types of micro-credit platform are emerging incessantly. However, with the emergence of various problems such as platform running and fraud, the lack of supervision on the micro-credit platform is exposed. Based on C5.0 decision tree algorithm, this paper establishes a risk monitoring model. Firstly, using big data technology to capture platform data, public information data and credit rating agency data, and then analyze and screen the key index system and establish a wind Finally, the validity of the model and algorithm is evaluated and verified. And by adjusting the parameters of the model, an evaluation criterion is proposed to reduce the α-error rate as much as possible while ensuring the overall error rate. The experiment shows that the wind-control model has a good predictive ability for the problem platform.