基于支持向量机的P2P网络借贷投资决策分析

发布时间：2018-01-01 13:42

本文关键词：基于支持向量机的P2P网络借贷投资决策分析　出处：《中国科技论文》2017年05期 　论文类型：期刊论文

【摘要】：为了获得最优贷款投资组合,从投资人的角度考虑P2P网络借贷的收益和风险,以及如何分配投资人的现金。首先,基于支持向量机(support vector machine,SVM)理论建立了贷款违约概率预测模型,得到每笔贷款的违约概率;进而,根据违约概率之间的相似距离和核回归构建贷款收益预测模型,进行贷款收益和风险的评估;然后,利用投资组合优化模型,辅助投资人制定合适的投资策略;最后,选取Lending Club平台数据对提出的模型进行验证。研究结果表明:基于SVM投资决策模型的贷款收益预测准确度,投资组合收益率均比逻辑回归模型更优。
[Abstract]:In order to obtain the optimal loan portfolio, consider the income and risk of P2P network loan from the investor's point of view, and how to distribute the investor's cash. Based on support vector machine (SVM) theory, a prediction model of loan default probability is established, and the default probability of each loan is obtained. Then, according to the similar distance between default probability and nuclear regression, the prediction model of loan income is constructed to evaluate the loan income and risk. Then, using portfolio optimization model to assist investors to formulate appropriate investment strategy; Finally, the Lending Club platform data are selected to verify the proposed model. The research results show that the prediction accuracy of loan income based on SVM investment decision model. Portfolio returns are better than logical regression models.
【作者单位】：武汉理工大学理学院;
【基金】：国家自然科学基金资助项目(81671633)
【分类号】：F830.5
【正文快照】： P2P借贷是1种基于网络的个人到个人的借贷。P2P借贷包含2个重要主体:借款人和投资人。借款人在网络上发布贷款申请,出借人则根据个人兴趣选择合适的贷款申请进行投标。相比于传统借贷模式,P2P借贷中的投资网络是双向的,每笔贷款可以获得多个投资人投资,而每个投资人也可以投资

【参考文献】