【摘要】 针对离线场景下利用轨迹数据挖掘用户移动模式时会泄露用户敏感位置隐私问题,文章利用兴趣点的地理空间分布,提出一种基于语义空间匿名的轨迹移动模式挖掘算法来抵御攻击者地图匹配攻击或语义推断攻击,同时实现用户移动模式的挖掘。该算法首先利用网格划分技术对城区进行均匀网格划分产生网格区域;然后利用网格区域中兴趣点的位置分布和语义差异度对轨迹停留点进行空间匿名以满足(k,l)隐私模型;最后借鉴经典模式挖掘Prefix Span算法思想对匿名轨迹数据集进行频繁移动模式的挖掘。理论分析和仿真实验验证了算法的安全性和有效性,与现有空间匿名的轨迹移动模式挖掘算法MCSPP相比,该算法不仅降低平均信息损失度,同时挖掘的频繁移动模式空间语义解释性更高。更多还原

【Abstract】 Aiming at the mining user movement patterns using trajectory data in offline scenario to leaks the privacy problems of the user sensitive location,using the geographic spatial distribution of points of interest,an anonymous trajectory-based moving pattern mining algorithm based on semantic space is proposed to defend against attacker map matching attacks or semantic inference attacks while implementing user mobility patterns mining.The algorithm first uses grid division technology to divide the urban area into uniform grids to generate grid areas.Then use the location distribution and semantic difference degree of the interest points in the grid area to spatially annotate the trajectory stay points to satisfy the(k,l) privacy model.Finally,the idea of mining Prefix Span algorithm based on classical model mining is used to mine frequent moving patterns of anonymous trajectory datasets.Theoretical analysis and simulation experiments verify the security and effectiveness of the algorithm.Compared with MCSPP,an existing space-based anonymous trajectory moving pattern mining algorithm,this algorithm not only reduces the average information loss degree,but also has a higher spatial interpretation of the frequent movement patterns of mining.更多还原