Spatial uncertainty trajectory dataset mining based on two-stages dynamic division

Due to the measurement precision, transmission delay and so on, we could only obtain uncertainty position information of moving objects. Spatial uncertainty trajectory data is uncertain in location of mobile objects. It is leading to the challenge of modeling uncertainty trajectory data and mining usable knowledge about movement pattern. In this paper, we propose a two-stages dynamic division method to dealing with the spatial uncertainty trajectory. The approach presents the notions of adjacent boundary cells and shared cells, and merges these cells into basic cells through distance and density membership degree. A comprehensive performance study on synthetic datasets shows that the proposed method in both effectiveness and scalability.