Frequent Semantic Trajectory Sequence Pattern Mining in Location-Based Social Networks.

At present, more and more researchers have focused on the study of the frequent trajectory sequence pattern mining in location-based social network (LBSN), in which the trajectories of contributing frequent patterns in users’ trajectory database must have same or similar the location coordinates and conform to the semantics and time constraints. In this paper, we focus on the study of users’ daily frequent mobile pattern. Excessive limitations on location information may limit the results of mining users’ frequent mobile pattern. Therefore, based on the frequent trajectory sequence pattern mining in LBSNs, we first define a new frequent semantic trajectory sequence pattern mining (FSTS-PM) problem that focuses on the study of mining users’ frequent mobile pattern. FSTS-PM problem does not consider the location coordinates of the trajectory points, but uses the distance and time constraints among the trajectory points in a trajectory sequence to optimize the user’s frequent mobile pattern mining results. Then, we propose the modified PrefixSpan (MP) algorithm which integrates the distance and time filtering mechanism based on the original PrefixSpan to find frequent semantic trajectory sequence pattern. Finally, the extensive experiments verify the performance of MP algorithm.