摘要
构建了一种基于交互感知的动态自适应信任评估模型,将历史交互窗口和可信推荐数引入到了总体信任评估中,克服了传统模型对交互证据感知能力不足的问题。提出了基于满意度迭代的直接信任积累方法,并采用实体稳定度实现了激励和惩罚2种迭代策略,有效抑制了恶意伪装实体的作弊行为。给出了一种基于直接和间接相结合的综合推荐信任聚合方法,通过引入实体熟悉度和评分相似度解决了传统模型推荐准确度低和不可靠的问题。实验结果表明,与已有模型相比,该模型有效地提高了信任评估的准确性,并具有更强的抵御串谋实体协同作弊的能力。
A dynamic adaptive trust evaluation model was established based on interaction-aware. The historical interac- tion window and trustworthy recommendation number was introduced in overall trust evaluation method, which over- comes the shortage of traditional models that lack the capacity to interaction-aware. The direct trust accumulation method based on interaction satisfaction degree iterative calculation was proposed, which achieved the incentive and penalty it- erative strategy based on entity stability factor, and effectively inhibits malicious entities with camouflage. A synthetical recommendation trust aggregating method based on combination of direct and indirect recommendation trust was given, which solved the accuracy low and unreliable problems of traditional recommendation methods by introducing entity fa- miliarity factor and scoring similarity factor. Simulation results show that, compared to the existing trust model, the model can effectively improve the accuracy of trust evaluation, and can provide a better capacity of resisting collusive entities.
出处
《通信学报》
EI
CSCD
北大核心
2012年第10期60-70,共11页
Journal on Communications
基金
国家自然科学基金资助项目(61272125)
河北省自然科学基金资助项目(F2011203234)
河北高等学校科学技术研究重点基金资助项目(ZH2011115)~~
关键词
网络计算
信任
信任评估
交互感知
交互满意度
internet computing
trust
trust evaluation
interaction-aware
interaction satisfaction degree
