A friend recommendation algorithm based on trajectory mining

In recent years, with the popularity of the running and other sports software, friend recommendation algorithm based on trajectory is gradually becoming a hot research. In this paper, the theta-ADBSCAN algorithm is used to dig the hot trail and the resident points of user's trajectory, then the trajectory segmentation algorithm is described, and the trajectory is replaced by the MTR which is composed of the piecewise points and the hot trail. At last, we further study user behavior similarity based on the trajectory region and the resident point for friend recommendation.