【摘要】 为解决动态社会网络发布中敏感边的隐私保护问题,针对攻击者将目标节点在不同时刻的节点度作为背景知识的应用场景,提出了一种新的基于动态网络的敏感边的隐私保护方法,它的思想是:首先通过k-分组和(k,Δd)-匿名发布隐私保护方法来确保目标节点不能被唯一识别,被攻击识别的概率不超过1/k;其次根据泄露概率对边进行保护,确保敏感边泄露的概率不超过用户给定参数u。理论分析和实验证明,所提出的方法可以抵御攻击者对敏感边的攻击,能有效地保护社会网络中用户的隐私信息,同时保证了动态社会网络发布的质量。更多还原

【Abstract】 In order to solve the issues of privacy preservation of sensitive edges in dynamic social networks data publication,we proposed a novel technique about the privacy preservation of sensitive edges based on dynamic social networks.The atta-cker uses the degrees of target nodes at different times as their background knowledge.Firstly,by using kgrouping and(k,Δd)-anonymous,it can be sure that the target nodes can not be uniquely identified by privacy attackers.The probability of being uniquely identified is no more than 1/k.Secondly,this method can ensure that the leakage probability of sensitive edges will not exceed the user defined parameter u.Theoretical analysis and experiments show that the method presented in this paper can resist sensitive edges identification attacks.It can not only protect the users privacy information effectively but also ensure the utility of published data in dynamic social networks.更多还原