在各种信任管理模型中,共谋欺骗是最棘手的恶意行为之一.由于信任值的主观特性,目前对共谋欺骗还没有较好的判定方法.本文利用聚类方法,提出一种集中信任模型下的共谋检测算法.为检验算法性能和参数设置,文中设计了相关的仿真实验,并进行了量化分析.实验结果表明,在选取合适参数的情况下,算法具有较好的检测效果.
Collusion is one of the most intractable malicious behaviors in the trust management model.Due to the subjective property of trust value,there is no good method to determine collusion action yet.This paper proposes a collusion detection algorithm for concentrative trust management model by clustering method.A correlative simulation program is designed to test the performance of the algorithm and settings of parameters.We also have a quantitative analysis for it.The simulation results indicate that the method is effective when choosing proper parameters.