Personalized Geo-Social Group Queries in Location-Based Social Networks

Geo-social group query, one of the most important issues in LBSNs, combines both location and social factors to generate useful computational results, which is attracting increasing interests from both industrial and academic communities. In this paper, we propose a new type of queries, personalized geo-social group (PGSG) queries, which aim to retrieve both a user group and a venue. Specifically, a PGSG query intends to find a group-venue pattern (consisting of a venue and a group of users with size h), where each user in the group is socially connected with at least c other users in the group and the maximum distance of all the users in the group to the venue is minimized. To tackle the problem of the PGSG query, we propose GVPS, a novel search algorithm to find the optimal user group and venue simultaneously. Moreover, we extend the PGSG query to top-k personalized geo-social group (TkPGSG) query. Instead of finding the optimal solution in the PGSG query, the TkPGSG query is to return multiple feasibility solutions to guarantee the diversity. We propose an advanced search algorithm TkPH to address the TkPGSG query. Comprehensive experimental results demonstrate the efficiency and effectiveness of our proposed approaches in processing the PGSG query and the TkPGSG query on large real-world datasets.