摘要： 基于多数名声系统中2~3人合谋的现象，提出反名声合谋(ACR)简便算法，通过计算个体间的评分平均分及评分相似性等手段识别合谋个体，并进行相应的名声纠正。但该算法无法区分合谋小组和高信任度小组(HTG)，进而采用信任矩阵分别计算个体的组内和组间名声，并根据个体间关系决定信任判断采用的名声，使ACR兼容HTG。实例分析表明，2种方法对ACR都是有效的。

关键词: 合谋欺诈, 反名声合谋, 评分相似性, 信任矩阵

Abstract: Because conspiracy cheat in most Reputation system is often plotted by two or three peers, a simple and convenient method to Anti-Conspiracy of Reputation(ACR) is presented. It identifies the Conspiracy Peers(CPs) by calculating the rating similarity and the average rating between peers, and then corrects the peer’s reputation. But it can’t classify the conspiracy group and High Trusted Group(HTG). In order to overcome this deficiency, this paper presents a new method compatible with HTG, where a Trust Matrix(TM) is adopted, both reputations inside and outside HTG are calculated respectively, and the reputation which the trust judgment adopts is decided by the relationship of the peers. Example analysis shows two methods are helpful to ACR.

Key words: conspiracy cheat, Anti-Conspiracy of Reputation(ACR), rating similarity, Trust Matrix(TM)

中图分类号: 

• TP309