基于投影寻踪动态聚类模型的p2p网贷风险评价体系构建及实例分析

Constructing p2p Lending Risk Evaluation System and Case Study Based on Projection Pursuit Dynamic Clustering Model

由于市场监管等一系列政策的不完善,导致如今p2p网贷在给用户带来便利的同时也存在巨大的风险。为防范该风险,本文根据平台风险指数1个一级指标、平台成交量等4个二级指标和平均预期收益率等14个三级指标构成的评价指标体系和采集到的样本数据应用投影寻踪动态聚类(PPDC)对100家网贷平台进行实证评估研究,建模结果表明:PPDC模型与投影寻踪聚类(PPC)模型的结果基本一致,排名与网贷之家排名结果的一致性好,且不受人为主观因素的影响,又能求得平台风险和评价指标权重的大小及其排序,在p2p网贷风险评价中能够取得良好的效果,是分析p2p风险指标的一种新方法。

Due to the imperfection of a series of policies such as market supervision, P2P lending has brought convenience to users, but also exists great risks. To guard against this risk,this paper evaluates 100 online lending platforms by using projection pursuit dynamic clustering (PPDC) based on an evaluating indicator system of one first-level indicator including platform risk index、four second-level indicators including platform trading volumes、 fourteen third-level indicators including average expected returns and collected samples data.Modeling results show that the results of PPDC model and projection pursuit clustering (PPC) model are basically the same and the ranking is consistent with the ranking results of online lending homes, and is not affected by subjective factors .Simultaneously,it can also find the platform risk and the weight of the evaluating indicator and their ranking. Because it can achieve good results in p2p lending risk assessment, it is a new method to analyze p2p risk indicators.