Abstract

The advance of GPS-enabled devices has facilitated people to log their location history with GPS trajectories. These trajectories imply to some extent an individual’s behaviors and interests related to their outdoor movements. Therefore, we can understand users and locations as well as the correlation between them based on these trajectories. By mining a user’s life pattern from their trajectories, we are able to automatically respond to the user’s unspoken needs. By mining multiple users’ trajectories, we can find out the top interesting locations, travel sequences and the travel experts in a given region. This information can enable generic travel recommendation and help people understand an unfamiliar city with minimal effort. By measuring the similarity between different users’ location histories, we could estimate the similarity between users and perform a personalized friend recommendation. Using such user similarity, a personalized location recommendation can be conducted in terms of the location history of a user and that of others. Overall, we can mine from the user trajectories rich knowledge, which may enable many smart location-based services, such as generic travel recommenders and personalized friend & location recommendations.