分布式系统推荐信任模型研究

Research on distributed system recommendation trust model

针对分布式网络中节点请求推荐信任时需要选取推荐节点的问题,提出一种能够准确评估推荐节点推荐能力信任度的信任模型DSRT(Distributed System Recommendation Trust)。该模型在区分服务质量信任度和推荐能力信任度的基础上,研究了影响推荐能力信任度的节点相关性、频繁度和风险等因素,其中,相关性不仅考虑了节点的服务关心相似度,还考虑了节点的评价能力相似度,频繁度和风险的计算均考虑了请求节点和其他交互节点的不同。模拟实验表明,该模型能够准确选取推荐节点,提高节点交互的满意率。

In distributed network, a node requesting recommendation needs to select right recommendation node. This paper proposes a recommendation trust model. This model distinguishes service quality trust from recommended capacity trust, includs nodes’correlations, frequencies and risk. In these elements, correlations are measured by service care similarity and assessment capacity similarity, and frequencies and risk are calculated considering different nodes. Experimental results demonstrate that the model can select recommendation nodes accurately and improve the satisfaction rate of the nodes interactions.