Mining user patterns for location prediction in mobile social networks

Understanding human mobility dynamics is of an essential importance to today mobile applications, including context-aware advertising and city wide sensing applications. Recently, Location-based social networks (LBSNs) have attracted important researchers' efforts, to investigate spatial, temporal and social aspects of user patterns. LBSNs allow users to "check-in" at geographical locations and share this information with friends. In this paper, analysis of check-ins data provided by Foursquare, the online location-based social network, allows us to construct a set of features that capture: spatial, temporal and similarity characteristics of user mobility. We apply this knowledge to location prediction problem, and combine these features in supervised learning for future location prediction. We find that the supervised classifier based on the combination of multiple features offers reasonable accuracy.