A feature based method for trajectory dataset segmentation and profiling

摘要: The pervasiveness of location-acquisition and mobile computing techniques has generated massive spatial trajectory data, which brought great challenges to the management analysis such a big data. In this paper, we focus on sub-trajectory dataset profiling problem, aim extract representative sub-trajectories from raw as subset, called profile, can best describe whole dataset. This problem is very challenging subject finding most set by trading off size quality profile. To tackle model features aspects density, speed direction flow. Meanwhile present our two-step method select trajectories based feature model. First, novel segmentation algorithm applied identify segments concerning their representativeness automatically estimate number segment borders. Then, performed yield in dataset, local heuristic evolution strategy. We evaluate extensive experiments using two real-world datasets over 12,000 taxicabs Beijing Shanghai. results demonstrate efficiency effectiveness methods different applications.