Location prediction in location-based social networks

Abstract

Developments in mobile devices and wireless networks have led to the increasing popularity of location-based socialnetworks. These networks allow users to explore new places , share their location, videos and photos and make friends. They give information about the mobility of users, which can be used to improve the networks. This paper studies the problem of predi cting the next check-in of users of loca tion-based social networks. For an accura te prediction, we first analyse the datasets tha t are obtained from the social networks, Foursquare and Gowalla. Then we obtain some features like place popularity, place popular time range , place distance to user’s home , user’s past visits , category preferences and friendships , which are used for predi ction and deeper understanding of the user behaviours . We use each fea ture indi vidually, and then in combination, using the new method. Finally, we compare the acqui red results and observe the improvement with the new method. Keywords: Loca tion predi ction, loca tion-based social network, check-in data.