Mining frequent trajectory patterns in road network based on similar trajectory

Abstract

Mining Trajectory Patterns plays an important role in moving objects. In many practical applications, the movement of objectives is always constrained by spatial space (e.g. road network). Therefore, it has more realistic significance to work on frequent trajectory patterns. This paper proposes a frequent trajectory patterns mining algorithm based on similar trajectory (named SimTraj-PrefixSpan). Since the trajectory data with the same frequent trajectory pattern may be not exactly the same, a trajectory similarity is utilized to measure whether considered trajectories have the same pattern. Computational results on simulated data and real data verify that the proposed algorithm can mine more complete and continuous frequent trajectory patterns, and shows the superiority of the proposed algorithm over traditional trajectory patterns mining algorithms in terms of mining efficiency.