Query-Aware Anonymization in Location-Based Service

Query privacy protection is an important part of the privacy protection in LBS. Currently, query privacy protection mainly consider the two typical attack models: query sampling attack and query homogeneity attack. The two attack model can be prevented by ensuring that cloaking regions satisfy ksharing property or l-diversity property, however, when the overall POI (point of interests)distribution is non-uniform, attackers may still infer a specific user´s query content with a high probability, which we refer to as skew attack. To address this issue, we modify the t-closeness model which is widely used in the data privacy protection, then we propose the QueryAwareCloak algorithm which is query-aware when generating cloaking regions. Finally, we verify the validation and effectiveness of the model and algorithm by experiments.