摘要
以人人贷平台作为研究对象,通过编写程序抓取实际交易数据,构建了基于二元logistic回归方法的P2P(Peerto-Peer)网络借贷决策模型,以研究影响投资决策的因素。该模型对13684条样本进行了是否满标的预测,其综合预测率为97%。实证结果表明,P2P平台上的借款信息对投资决策行为及是否能成功借贷的影响十分显著——利率、年龄、婚姻状态、学历、信用等级以及逾期次数与满标成功率呈正相关关系;金额、期限、申请借贷笔数与满标成功率呈负相关关系。
This paper selected＂Renrendai＂platform as the research object,and wrote a program to fetch the actual transaction data,to build an online P2 Plending decision model which based on binary logistic regression method for studying the influence factors on investment decision.This model was used to predict whether the 13684 samples would be fully funded,and its overall prediction rate was as high as 97%.The empirical results show that the information displaying on P2 Pnetwork lending platform has significant effects on investment decision-making behavior and ultimate lending success rate——Interest rates,age,marital status and education,credit rating and the number of overdue presented positive correlation with full scale probability;Amount,term,the number of loan application are negatively related with full scale probability.
出处
《科技和产业》
2016年第12期163-168,共6页
Science Technology and Industry
