Differentially Private Geo-Social Ranking

With the popularity of Geo-Social networks, the combination of social and spatial data facilitates Geo-Social data analyzing. One kind of such analyses is Geo-Social ranking. Given a query location q, the Geo-Social ranking problem is to rank the users in the Geo-Social network, based on their adjacency to q and their influential on their friends. However, the Geo-Social network contains massive privacy information of individuals, such as the relationships between users and the location information of users. Leakage of such information may cause serious impact on users. In this paper, we investigate the problem of privacy preserving Geo-Social ranking, which finishes the ranking task while protecting the social and spatial privacy of users. We solve the problem based on the notion of differential privacy. We define the differential privacy target in Geo-Social ranking and design a ranking algorithm that satisfies differential privacy. We evaluate our approach with real-world dataset to show the efficiency of our algorithm.