P2P电子商务信誉模型研究

发布时间：2018-04-13 00:31

本文选题：P2P电子商务 + 信誉模型　；参考：《重庆大学》2010年硕士论文

【摘要】： 随着对P2P网络技术研究的逐渐深入,P2P作为一种分布式网络技术在许多领域的得到了广泛的应用,其中也包括在电子商务领域中的应用。然而P2P网络技术存在的固有安全性问题已成为制约P2P电子商务发展的关键因素。信誉模型的研究为解决P2P电子商务中的信誉保障问题提供了一种方案。信誉模型的主要功能是完成对P2P电子商务中用户信誉度的评估,为用户之间的交易提供安全的信任关系。而目前主要地将信誉模型没有提出一套完整的信誉管理机制和信誉评估机制有机结合的信誉模型,多数信誉模型只侧重解决其中一个方面。针对上述问题,本文以P2P电子商务为研究背景,从信誉管理机制和信誉评估机制两个方面构建一种基于簇域的信誉模型。论文的主要工作包括以下几个方面: ①分析当前主要的P2P信誉模型。在学习和掌握P2P网络技术的特点和网络拓扑的基础上,分析了信任特征和信任模型结构,对已有的信誉模型进行了分类,总结了当前主要的信誉模型的优缺点,为本文构建基于簇域的P2P电子商务信誉模型奠定基础。 ②针对分布式P2P电子商务系统的信誉管理机制,提出一种基于簇域的信誉管理模型。在分析现有集中式和分布式网络在信誉管理系统的优缺点后,提出了一种基于簇域的信誉管理网络拓扑结构。利用簇域子网层和索引网络层完成了P2P电子商务网络中的商品发布和搜索流程,实现电子商务的在线交易。利用拓扑结构中的信誉代理节点实现分布式信誉数据的高效存取,为信誉评估模型实现奠定基础。 ③针对分布式P2P电子商务系统的信誉评估机制,提出一种基于簇域的信誉评估模型P2P-CRep。针对现有信誉评估机制中普遍存在的问题,本文的P2P-CRep信誉模型作了进一步地改进。具体改进措施包括:针对电子商务中交易的多因素性和层次性,利用模糊综合评判法对交易进行评价;直接信任度、推荐信誉和综合信誉的计算中引入了较多的影响因素;根据节点之间的评估误差和簇域相似度计算推荐节点的推荐可信度;根据平均推荐可信度和平均交易量两个因素计算置信因子,再利用置信因子法综合计算节点的信誉度;针对电子商务网络中的累积小交易信誉行骗和共谋欺诈恶意行为,在P2P-CRep信誉评估算法中引入了对这两种恶意行为的抵御机制。 ④模型分析和仿真实验。通过数学分析,证明了基于簇域的信誉管理模型网络具有一定的网络动态性和扩展性,且具有较高网络路由性能。通过仿真实验,验证了P2P-CRep信誉模型能够提高电子商务网络中交易的成功率,且对共谋欺诈和累积小交易信誉行骗的恶意行为具有有效的抵御能力。最后在总结本文工作的基础上,提出了若干有待进一步深入研究和探讨的问题。
[Abstract]:As a distributed network technology, P2P has been widely used in many fields, including in the field of electronic commerce.However, the inherent security problem in P2P network technology has become a key factor restricting the development of P2P e-commerce.The research of reputation model provides a scheme to solve the problem of reputation protection in P2P e-commerce.The main function of the reputation model is to evaluate the reputation of users in P2P e-commerce, and to provide a secure trust relationship for transactions between users.At present, the reputation model has not been put forward a complete reputation management mechanism and reputation evaluation mechanism, most reputation models only focus on one aspect.Aiming at the above problems, this paper builds a reputation model based on cluster domain from two aspects of reputation management mechanism and reputation evaluation mechanism, taking P2P e-commerce as the research background.The main work of the thesis includes the following aspects:1. Analyze the main P2P reputation models.On the basis of learning and mastering the characteristics of P2P network technology and network topology, this paper analyzes the trust characteristics and trust model structure, classifies the existing reputation models, and summarizes the advantages and disadvantages of the current main reputation models.It lays a foundation for constructing the reputation model of P2P e-commerce based on cluster domain.2. Aiming at the reputation management mechanism of distributed P2P electronic commerce system, a reputation management model based on cluster domain is proposed.After analyzing the advantages and disadvantages of existing centralized and distributed networks in reputation management systems, a cluster domain-based reputation management network topology is proposed.By using cluster subnet layer and index network layer, the commodity publishing and searching process in P2P e-commerce network is completed, and the online transaction of electronic commerce is realized.The efficient access of distributed reputation data is realized by using the reputation proxy node in the topology, which lays a foundation for the implementation of reputation evaluation model.3. Aiming at the reputation evaluation mechanism of distributed P2P e-commerce system, a cluster based reputation evaluation model P2P-CRepP is proposed.Aiming at the common problems in the existing reputation evaluation mechanism, the P2P-CRep reputation model of this paper is further improved.The specific improvement measures include: according to the multi-factor and hierarchical nature of the transaction in electronic commerce, the fuzzy comprehensive evaluation method is used to evaluate the transaction, the direct trust, the recommended reputation and the comprehensive reputation are introduced into the calculation of the more influential factors.According to the evaluation error and cluster similarity between nodes, the recommended credibility of the recommended node is calculated, the confidence factor is calculated according to the two factors of average recommendation credibility and average transaction volume, and the credibility of the node is calculated comprehensively by using the confidence factor method.Aiming at the malicious behavior of accumulated small transaction reputation fraud and conspiracy fraud in the electronic commerce network, this paper introduces a mechanism to resist these two malicious acts in the P2P-CRep reputation evaluation algorithm.4 Model analysis and simulation experiment.Through mathematical analysis, it is proved that the reputation management model network based on cluster domain has certain network dynamic and expansibility, and has high network routing performance.The simulation results show that the P2P-CRep reputation model can improve the success rate of transaction in electronic commerce network and can resist the malicious behavior of collusion fraud and accumulated reputation fraud of small transaction effectively.Finally, on the basis of summarizing the work of this paper, some problems that need to be further studied and discussed are put forward.
【学位授予单位】：重庆大学
【学位级别】：硕士
【学位授予年份】：2010
【分类号】：TP393.02

【相似文献】