CAJ | 학술논문

针对 P2P网络中存在的安全问题，本文提出一种基于隐Markov模型(Hidden Markov Model, HMM)的P2P信任模型，采用HMM对节点行为进行建模，在此模型的基础上，利用节点交互结果作为观测历史，由前向-后向算法计算得到节点行为概率分布，作为节点的信任值评估。考虑到信任评估实时性的需求，文章还提出了一种基于滑动窗口和遗忘因子的模型更新算法。
침대 P2P망락중존재적안전문제，본문제출일충기우은Markov모형(Hidden Markov Model, HMM)적P2P신임모형，채용HMM대절점행위진행건모，재차모형적기출상，이용절점교호결과작위관측역사，유전향-후향산법계산득도절점행위개솔분포，작위절점적신임치평고。고필도신임평고실시성적수구，문장환제출료일충기우활동창구화유망인자적모형경신산법。
Aiming at the P2P networks security issues, this paper proposes the HMM-based P2P trust model, uses HMM to model the behavior of the peer. On the basis of this model, peer interaction result is used as the observation history. The probability distribution of peer behavior evaluated as trust can be calculated by HMM forward-backward algorithm. In consideration of the real time demand of trust evaluation, we proposes a model updating algorithm based on sliding window and forgetting factor.