摘要
提出了一种基于不满意度的网络安全模型,主要功能是帮助用户在网络环境中正确地选择交易对象,屏蔽恶意节点,基于不满意度(degree of dissatisfaction,DoD)对交易节点进行分类控制。节点的不满意度定义为该节点属于恶意节点集的概率。a)使用粗糙集(rough set)模块与Bayesian学习器计算节点的不满意度,依据节点的交易历史记录计算节点的本地不满意度(local DoD,LDoD),依据反馈推荐意见计算推荐不满意度(recom-mended DoD,RDoD),基于不满意度将节点划分为可信任节点、陌生节点、恶意节点等不同的类型;b)基于推荐意见的信息熵(information entropy)计算其可信度,对反馈推荐意见进行综合。实验表明,与已有的安全模型相比,提出的安全管理模型对恶意节点具有更高的检测率,具有更满意的交易成功率。
This paper proposed a new security model based on the degree of dissatisfaction （DoD）. The main function was to help users in network environment to select correctly the transaction object and to avoid malicious node. This paper defined the degree of dissatisfaction as the probability that a node belonged to the malicious node set. This model used the rough set and Bayesian learner to calculate the DoD of the nodes, and based on historical transaction records to calculate the local DoD （LDoD） , based on feedbacks on recommendations to calculate the recommended DoD （ RDoD）, and divided nodes into trust nodes, strange nodes and malicious nodes. It calculated the credibility of feedbacks based on the information entropy, and used the credibility of feedbacks to integrate feedbacks. Compared with existing trust models, the experiment indicates that the mod- el can obtain higher examination rate over malicious nodes, with the higher transaction success rate.
出处
《计算机应用研究》
CSCD
北大核心
2013年第2期566-569,共4页
Application Research of Computers
基金
国家自然科学基金资助项目(60873071)
关键词
网络安全模型
不满意度
粗糙集
贝叶斯学习器
信息熵
仿真
network security model
degree of dissatisfaction
rough set
Bayesian learner
information entropy
simula-tion
