%0 Journal Article
%T P2P平台满标影响因素研究—以人人贷为例<br>Research on the Factors Influencing Successful Targets on P2P Platform—A Case Study of Ren Ren Dai
%A 刘梦玲
%J Hans Journal of Data Mining
%P 54-59
%@ 2163-1468
%D 2016
%I Hans Publishing
%R 10.12677/HJDM.2016.61007
%X <div style=\"text-align:justify;\">
	<span style=\"line-height:1.5;\">本文选取以人人贷为例的P2P平台上5718条标的信息记录，以能反映标的状况的变量作为因变量，以借出信用分、借款金额等5个变量作为自变量，建立Logistic回归模型以帮助借款人对借款成功(满标)概率进行预测，最终模型预测的正确率达到93.7%。此外，以标的状况为因变量的决策树分类的结果显示误判率为0.05，综合两种方法最终得出信用等级、借出信用分数和以往逾期次数对标的状态影响最大，而标的类型的影响不显著。<br />
In this paper, 5718 target recordings on Ren Ren Dai (a Peer-to-Peer lending platform) are selected for research. We take the variable which can reflect the target status as dependent variable, and lend credit scores, loan amount, etc. as independent variables. In order to help borrowers predict the success ratio, we build the Logistic model; the accuracy rate of the model is verified to be 93.7%. In addition, the results of the decision tree classification show a failure rate of 0.05. Finally, a comprehensive of the two methods comes to a conclusion that: credit rating, lending credit score and overdue times have the greatest impacts on the target status, and the variable target type has no significant effect.</span><span style=\"line-height:1.5;\"></span> 
</div>
%K P2P，Logistic模型，决策树，误判率<br>Peer-to-Peer
%K Logistic Model
%K Decision Tree
%K Failure Rate
%U http://www.hanspub.org/journal/PaperInformation.aspx?PaperID=16829