PrivateCheckIn:一种移动社交网络中的轨迹隐私保护方法

PrivateCheckIn:Trajectory Privacy-Preserving for Check-In Services in MSNS

移动设备的发展及无线网络的普及促使移动社交网络的出现及发展.签到服务作为移动社交网络中的主流应用,存在着严重的轨迹隐私泄露风险.文中针对签到服务中假名用户的轨迹隐私泄露问题,提出了一种轨迹隐私保护方法PrivateCheckIn.该方法设计了一种签到序列缓存机制,通过为缓存的签到序列建立前缀树、对前缀树进行剪枝及重构形成k-匿名前缀树,遍历k-匿名前缀树得到k-匿名签到序列,达到了轨迹k-匿名的隐私保护效果.文中证明了PrivateCheckIn方法既能保护假名用户的轨迹隐私,又确保损失签到位置最少,有效地保证了用户体验.通过构建前缀树的方式获取轨迹k-匿名集降低了计算代价.最后,文中在真实数据集上与(k,δ)-anonymity方法进行了充分的对比实验,验证了PrivateCheckIn方法的准确性与有效性.

With the development of mobile devices and wireless networks,mobile social network services（MSNS） arise and develop fast.Check-in service as one of the most popular services in MSNS,has serious personal privacy leakage threats.In this paper,we propose a trajectory privacy-preserving method called PrivateCheckIn,which can protect trajectory privacy for pseudonym users in check-in services.At first,we buffer the check-in sequences of pseudonym users,and then we build prefix trees for buffered check-in sequences,prune and re-construct prefix trees to get the k-anonymized version.At last,we traverse the k-anonymized prefix tree to get k-anonymized check-in sequences,which can achieve a privacy guarantee of k-anonymity.We prove in this paper,PrivateCheckIn guarantees the number of lost check-in locations is minimized while satisfying users＇ privacy requirements.PrivateCheckIn also reduces the cost of finding trajectory k-anonymity set.At last,we run a set of comparative experiments with（k,δ）-anonymity on real-world datasets,the results show accuracy and effectiveness of PrivateCheckIn.