Constructing And Comparing User Mobility Profiles For Location-Based Services

Along with the increasing availability of location-acquisition technologies (GPS, GSM networks, etc.) we have better access to collections of large spatio-temporal datasets. This brings us new opportunities in location-based services. First, using that large amount of information of spatio-temporal data, we can discover useful knowledge about users’ movement behaviors and generate their mobility profiles, which will be useful for researching anonymous trajectories or other novel applications in location-based services. Second, using those location-acquisition technologies, users can record their location histories. Such real-world location histories allow for the discovery of users’points of interests in places over time, which helps us to understand the relationships between users based on their location data. As a requirement for recommendation systems, we need to calculate a similarity score between each pair of users to recommend each user a list of potential friends.In this master’s thesis, we propose a novel method to generate users’ mobility profiles and calculate users’ similarity between each other. As part of the mobility profile generation procedure, we first detect users’Regions of Interest and use frequent sequential pattern (FSP) mining technology to mine users’ frequent trajectory patterns. In the similarity computation procedure, we first define a similarity function between users’ frequent trajectory patterns and then calculate the similarity between users based on their patterns’ similarity. We evaluate our work based on the GPS data which is collected in the Geolife project (Microsoft Research Asia) by178users in a period of over four years (from April2007to October2011). As a result, our method outperforms some state of the art methods in certain areas.