Enhancing privacy using location semantics in location based services

Abstract

The development of LBS bring great convenience to people's lives, but also presents new challenges to privacy protection. Many of the existing methods are inadequate because they do not consider the situation that the attackers have back-ground knowledge, cannot withstand the background knowledge attack well. In this paper, we introduce the concept of (θ, δ)-diversity, we consider the scenario that attackers have the greatest background knowledge. The Earth Mover Distance is used to calculate the semantic similarity between the real location and the other locations, the submitted data contains not only locations which have similar semantic with the real location, but also locations belong to different categories. To prevent excessive number of locations which have similar semantic with the real location, we use a proportional way to ensure privacy. Our method can well resist the background knowledge attack. Experiments show that our method is very practical.