Comparison study of sub-trajectory clustering in data mining

3rd International Conference on Advances in Energy, Environment and Chemical Engineering

能源学;生态环境科学;化学工业

Yang, Guodong^1 ; Huang, Zhitao^1 ; Wang, Xiang^1

School of Electronic Science and Engineering, National University of Defense Technology, Changsha

410073, China^1

关键词: Asynchronous sensor;    Comparative experiments;    Multi-sensor information fusion;    Positioning accuracy;    Similarity measure;    Trajectory analysis;    Trajectory clustering;    Trajectory data minings;   

Others  :  https://iopscience.iop.org/article/10.1088/1755-1315/69/1/012143/pdf
DOI  :  10.1088/1755-1315/69/1/012143

学科分类：环境科学（综合）

来源: IOP

【 摘 要 】

Trajectory clustering is an important method to achieve moving object data mining, multi-sensor information fusion and trajectory knowledge discovery. Sub-trajectory clustering is an important method to extract useful information from a large number of trajectory data in trajectory analysis. In this paper, comparative experiments are made on the time consumption, similarity measure and clustering performance based on the existing sub-trajectory clustering methods. Based on the comparisons, the advantages and disadvantages of different methods are presented and an improved method is proposed for dealing with trajectories with low positioning accuracy and correlating tracklets from asynchronous sensors. Besides, a general framework of trajectory data mining is discussed.