P2P电子商务中基于声誉的关系信任模型

发布时间：2018-05-24 00:35

本文选题：对等网络 + 电子商务　；参考：《广西大学学报(自然科学版)》2013年05期

【摘要】：P2P电子商务迅速得到发展,但如何在众多的商品中选择自己满意的商品并不是一件容易的事。为了满足用户多样性的要求,提出了一种基于声誉的关系信任模型,用户在Agent的协助下以商品的信任值排序后选择自己满意的商品。商品的信任值包含启动信任值和经验信任值两部分,启动信任值由商品的声明属性与同类商品的实际属性和信任值对比获得,经验信任值由交易经历和用户评价获得,在计算经验信任值时,用动态时间窗口减少计算量,在全局范围评估评价可信度以降低恶意评价的影响。测试和计算结果表明启动信任值能让无评价或仅有很少评价的优质新商品能与信任度很高的商品合理竞争,比另外两种信任启动方法更适合电子商务系统;关系信任模型对恶意评价有较强的抵抗能力。
[Abstract]:P2P e-commerce is developing rapidly, but it is not easy to choose satisfied goods among many products. In order to meet the requirements of user diversity, a reputation based relational trust model is proposed. With the help of Agent, users choose their own satisfied products after ranking them by the trust value of their products. The trust value of a commodity consists of two parts: the starting trust value and the empirical trust value. The initial trust value is obtained by comparing the declared attribute of the commodity with the actual property and trust value of the same commodity, and the empirical trust value is obtained by the transaction experience and the user evaluation. In order to reduce the impact of malicious evaluation, the dynamic time window is used to reduce the amount of computation and evaluate the reliability of the evaluation in the global scope when calculating the empirical trust value. The test and calculation results show that the starting trust value can make the high quality new goods with no evaluation or little evaluation compete reasonably with the products with high trust level, which is more suitable for the electronic commerce system than the other two trust initiation methods. The relationship trust model has strong resistance to malicious evaluation.
【作者单位】：贵州师范大学数学与计算机科学学院;中南大学信息科学与工程学院;
【基金】：贵州省科学技术基金资助项目(黔科合J字[2010]2102号)
【分类号】：TP393.08

【参考文献】