基于MDF-Nearest算法的轨迹隐私保护方法

传统轨迹匿名方法在匿名集生成时没有考虑用户多种特征属性，在信息攻击下无法有效保护真实位置；在轨迹形成方面因没有将余弦角度和轨迹间距离作为形成的依据，导致某些虚假轨迹无法有效保护真实轨迹。为改善以上问题，构建一种依据用户多重特征信息构建的匿名集以保证匿名有效性；采用协作用户的真实轨迹并计算相似性，从而生成虚假轨迹相似性高的MDF-Nearest算法。实验结果表明，该方法随着k值的变大与生成轨迹数量的增多，隐私保护效果逐渐改进；与传统k匿名方法相比，该算法时间开销降低41.7%，而隐私保护程度可提高至97.1%。因此该方法能以较低的时间开销，提供质量可靠的位置服务，保护用户信息。

Trajectory Privacy Protection Method Based on MTF-Nearest+Algorithm

The traditional trajectory anonymity method does not take the user’s multiple feature attributes into account when generating the anonymous set， which leads to inefficient protection of the real position under the information attack. What’s more， the angle and the inter-track distance are not formed in the trajectory formation， and thus real trajectory cannot be effectively protected. This paper proposes an anonymous set based on the user’s multiple feature information to ensure the validity of the anonymity. The real trajectory of the collaborative user is used and the similarity is calculated to generate a false trajectory with high similarity. The experimental results show that with the value of k and the number of generated trajectories increases， the effect of privacy protection is better. Compared with the traditional k-anonymous method， the time cost of the algorithm is reduced by about 41.7%， while the degree of privacy protection is increased to about 97.1%. The method can provide high quality location services and protect user information with low time overhead.