Constructing and comparing user mobility profiles for location-based services

With the increasing availability of location-acquisition technologies, we have better access to collections of large spatio-temporal datasets. This brings new opportunities to location-based services (LBS), especially when knowledge of users' movement behaviour (i.e., mobility profiles) can be extracted from such datasets. For instance, in social networks, friends can be recommended according to similarity scores between user mobility profiles. In this paper, we propose a new approach to construct users' mobility profiles and calculate the mobility similarities between users. We model mobility profiles as traces of places that users frequently visit and use frequent sequential pattern mining technologies to extract them. To compare users' mobility profiles, we first discuss the weakness of a similarity measurement in the literature and then propose our new measurement. We evaluate our work using a real-life dataset published by Microsoft Research Asia and the experimental results show that our approach outperforms the existing works on different aspects.