P2P网贷用户信用评估

发布时间：2018-07-22 14:27

【摘要】：随着互联网金融的发展,P2P网贷、众筹融资等给人们带来便利的同时,也使人们面临着巨大的信用风险等一系列弊端,解决这些问题,为用户带来更安全、更便捷的交易环境成为研究的热点。在P2P网贷(Peer-to-Peer)中,借贷双方不可能像传统金融中那样彼此深度了解,信息不对称性尤为突出。解决这个问题的一个有效办法就是尽量准确评估借贷主体的信用,即建立有效的P2P用户征信体系。本文在研究了国内外学者对于P2P网贷用户的信用评判方法后,发现国内学者对通过用户信用评估指标来评判P2P网贷用户信用的研究方法甚少,而且其研究方法操作大多较复杂,信用评估值也不一定很好。本文提出了一种研究信用评估指标的信用评判方法——综合评判法,其求得的信用评估值是通过对用户的信用评估指标进行分层以及多次综合加权考虑后得到,不仅操作简单、省时,而且具有一定的可信度。并用融360的数据对其进行了实验验证。从本质上讲,当我们知道的借款人各方面的信息越多,就会对其信用进行更准确的评判,因此想到将信息融合技术运用到P2P网贷用户信用评估中。本文在近似认知计算系统这一理论信息融合思想的启发下,首次将信息融合技术运到P2P网贷中,提出了基于粒计算的信息融合方法。该方法融合了用户多方面信息,不仅避免了用户以虚假信息得到较高信用评估值的可能,而且能够对用户的信用进行全面、准确的评判。同样也用融360的数据以及收集到的用户的第三方数据(如各部门的执法数据等)对其进行实验验证。区块链是互联网金融的底层技术构架,具有去中心化,信息高度透明等特点,并且其本质上就是各方信任机制建设的一个完美的数学解决方案。本文在对其进行研究后,提出了将区块链技术运用到P2P网贷用户信用评估中的设想,初次建立了基于区块链技术的模型并就其对用户信用验证的底层验证机制进行了算法设计。或许在不久的将来能够在一定程度上减少用户信用值不真实的现象。
[Abstract]:With the development of Internet finance, P2P network loan, crowdfunding and so on bring convenience to people, at the same time, it also makes people face a series of malpractices such as huge credit risk, solve these problems, and bring more security to users. More convenient trading environment has become the focus of research. In P2P network loan (Peer-to-Peer), it is impossible for both parties to understand each other in depth as in traditional finance, especially in information asymmetry. An effective way to solve this problem is to evaluate the credit of the borrowers as accurately as possible, that is, to establish an effective P2P user credit system. After studying the domestic and foreign scholars' credit evaluation methods for P2P network loan users, it is found that domestic scholars have few research methods to judge the P2P network loan users' credit through the user credit evaluation index, and most of the research methods are complicated. Credit ratings are not necessarily good either. In this paper, a credit evaluation method, which is used to study the credit evaluation index, is put forward. The credit evaluation value is obtained by stratifying the credit evaluation index of the user and considering the credit evaluation index for many times. It is not only simple to operate. Save time, and have certain credibility. It is verified by the data of Rong 360. In essence, the more information we know about the borrower, the more accurate the credit evaluation will be, so we think of applying the information fusion technology to the credit evaluation of P2P network loan users. Inspired by the theoretical information fusion idea of approximate cognitive computing system, the information fusion technology is transported to P2P network loan for the first time in this paper, and a method of information fusion based on granular computing is proposed. This method not only avoids the possibility that users can obtain higher credit evaluation value by false information, but also can comprehensively and accurately judge the credit of users. It is also tested with 360 data and third-party data collected from users (such as law enforcement data from various departments). Block chain is the underlying technical framework of Internet finance, which has the characteristics of decentralization and high transparency of information. In essence, it is a perfect mathematical solution for the construction of trust mechanism of all parties. In this paper, we put forward the idea of applying the blockchain technology to the credit evaluation of the users in P2P network, and establish the model based on the block-chain technology for the first time, and design the algorithm for the underlying verification mechanism of the user's credit verification. Perhaps in the near future can reduce to some extent the user credit value untrue phenomenon.
【学位授予单位】：昆明理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：F832.4;F724.6

【参考文献】