基于差分隐私的LBS群组最近邻查询-计算机科学.PDFVIP

第 44 卷第6A 期 Vo l. 44 No. 6A 计算机科学 2017 年 6 月 CO岛1PUTER SCIENCE June 2017 基于差分隐私的 LBS 群组最近邻查询 马银方张琳 (南京邮电大学计算机学院 南京 210003) 摘 要 针对当前基于位置服务CLB凹的群纽最近邻查询中出现的隐私保护问题，提出了一种新的基于差分隐私保 护的 LBS 群纽最近邻查询方法，该方法满足了差分隐私性质并引入了区域不可区分这一新的理念。基于分类及聚 类给出了 LBS 群纽构建方法并提供了群纽隐私预算分配机制。提出了 LBS 群纽用户位置扰乱算法(GPOL) ，将群纽 最近邻查询转换为群纽质心的最近邻查询，并将其应用到整个隐私保护框架中。实验结采表明该方法能够有效地抵 御现有的交叉攻击和组合攻击。 关键词 位置服务，群最近邻查询，差分隐私，区域不可分 中图法分类号 TP393 文献标识码 A LBS Group Nearest Neighbor Query Method Based on Differential Privacy MA Yin-fang ZHANG Lin (College of Computer ,Nanjing University of Posts and Telecommunications ,Nanjing 210003 ,China) Abstract For the privacy issues caused by the group nearest neighbor query scenarios formed by multi-user collabora tion ,a new LBS group nearest neighbor query method based on privacy protection was proposed ,which introduces the gecγindistinguishability concept and satisfies with differential privacy property. LBS group construction mechanism based on classification and clustering was given and the group privacy budget allocation mechanism was studied. LBS Group-users position obfuscation by laplace (GPOL) algorithm was also introduced ,which has applied the group cen troid nearest neighbor query instead of group nearest neighbor query into the entire privacy protection framework. Ex perimental results show that this method can effectively resist the existing cross attacks and combo attacks. Keywords Location services ,Group nearest neighbor query ,Differential privacy ,Geo-indistinguishability