Collective periodic pattern discovery for understanding human mobility

Periodic behaviors are essential to understanding objects’ movements. In real world situations, the collective movement of moving objects hides useful periodic patterns that people are more interested in. Discovering such periodic patterns is helpful in exploring human mobility, which can benefit many applications, such as urban planning, traffic management and public security. However, the previous works mainly focused on detecting individual periodic behaviors, and rarely studied collective periodicities. This paper proposes a novel algorithm, called CPMine, which adopts filter-refine paradigm to mining collective periodic patterns. In the filter phase, CPMine filters the initial candidates generated by sub-patterns, and refines them to determine final results in the refinement phase. In order to improve the performance of pattern growth, this paper further proposes GMine_S algorithm that develops a pruning algorithm based on spatial proximity to rapidly filter enormous invalid candidates. To greatly reduce search space, CPMine_I algorithm is proposed to support more efficient trajectory queries by a specialized index structure and its update algorithm. Moreover, this paper employs spatial indexing techniques to speed up clustering process. Finally, experiments on three real trajectory datasets have verified the effectiveness and efficiency of our proposed algorithms respectively. Experiment results show that the improved algorithm CPMine-IS using pruning and index outperforms the other three algorithms significantly.