基于查询概率和时间分布的假语义位置选择算法

Dummy Semantic-location Selection Algorithm Based on Query Probability and Time Distribution

为了解决传统的假位置隐私保护方法未充分考虑访问时间和位置语义信息导致的隐私泄露问题,给出了基于查询概率和时间分布的假语义位置选择算法。利用查询概率构建了候选假语义位置集,通过语义位置间的余弦相似度对该集合中的语义位置进行了排序。根据用户设定的K值、敏感语义位置类型和语义安全阈值依次选择假语义位置,直至满足用户的隐私要求。通过实验将本文给出的算法和MMDS等其他3种算法进行了对比,得出以下结论:本文给出的算法在位置保护程度上约提升30%,在访问时间不可区分程度上约提升50%。

To solve the problem of privacy leakage caused by access time and location semantic information in traditional dummy location privacy protection methods, a dummy semantic location selection algorithm based on query probability and time distribution is proposed. The candidate pseudo semantic position set is constructed by using the query probability, and the semantic positions in the set are sorted by the cosine similarity between the semantic positions. According to the K value, sensitive semantic location type and semantic security threshold set by the user, the false semantic location is selected in turn until the user’s privacy requirements are met. The algorithm is compared with other three algorithms such as MMDS through experiments, and the following conclusions are drawn: the algorithm proposed in this paper improves the location protection by about 30% and the access time indistinguishable by about 50%.