Abstract:
Recent advances in wireless and positioning technologies and the prevalence of portable devices have fostered many novel tracking and monitoring applications. It can be anticipated that large amounts of spatio-temporal trajectory data will be quickly generated and accumulated. Much valuable knowledge may hide in these data, which is worth a deep exploring to analyze object’s movement behavior and the discovered knowledge can be further utilized to create more novel applications.
In this talk, I would like to share our recent study on group movement pattern (GMP) mining and semantic data compression. The GMP mining problem is to discover the group relationships of moving objects based on the regularity hidden in the trajectory data while semantic data compression is to utilize the application semantics, i.e., the group relationships and the movement regularity, to further compress the data to be transmitted. To solve the GMP mining and semantic data compression problems, we propose the DGMPMine algorithm and 2D2P algorithm respectively. The experimental results show that the DGMPMine algorithm can discover the group movement patterns effectively and efficiently and the 2D2P compression algorithm can utilize the mining results to achieve better compressibility.
