k-匿名改进模型下的LCSS-TA轨迹匿名算法

LCSS-TA trajectory anonymity algorithm based on improved k-anonymity model

传统的欧几里德距离度量函数计算轨迹相似性时,要求轨迹的每个位置点都要有对应点。由于噪声点的存在,导致轨迹距离出现较大偏差,降低轨迹相似性,增加轨迹的信息损失。针对这一问题,结合LCSS(最长公共子序列)距离函数和(k,δ)-匿名模型设计了LCSS-TA(最长公共子序列轨迹匿名)算法。该算法通过将轨迹位置点之间的距离映射成0或1来减小噪声点可能导致的较大距离。在合成数据集和含噪声的数据集下的实验结果表明,提出的算法在满足轨迹k-匿名隐私保护的基础上,可以有效降低噪声干扰,减少轨迹的信息损失。

In traditional trajectory similarity calculation based on the Euclidean distance metric function, position of each point in the trajectory are required to have a corresponding point. The existence of noises could lead to track a larger distance deviation, reduce the trajectory similarity, increase trajectory information loss. In order to solve this problem, this paper designed LCSS-TA（ longest common subsequences trajectory anonymity）algorithm combining with LCSS（ longest common subsequences） distance function and （k, δ）-anonymity model. The algorithm could decrease the greater distance of the noises might to lead by mapping the distance between trajectory locations points to 0 or 1. In synthetic data set and data set with noise, the experiment results show that the algorithm can reduce noises interference and decrease the trajectory information loss on the basis of meeting with k-anonymity privacy protection.