MHDS-Part: An Algorithm for Trajectory Partitioning

摘要：

The data mining of moving object trajectory is the basis for LBS but the part of local similarity will be ignored inevitably if the data of trajectory is mined as a whole. However, reasonable trajectory partitioning is an important way to find local similarity. According to the existing problems in related algorithms of present trajectory partitioning, the paper presents an algorithm of trajectory partitioning, Modified Hausdorff distance for Spatial trajectory matching (MHDS). If the distance between the sub-trajectory, between the actual position of the moving object and previous characteristic point, and the polyine, between the actual position of the moving object and previous characteristic point, exceeds the preset threshold, the previous location updating point of the actual point is set as a characteristic point. Finally, the experiment shows that the algorithm has better preciseness and conciseness.

关键词： Moving Object Sub-Trajectory partition characteristic point

作者: Fei SHAO Junzhong GU

作者单位: Institute of Computer Applications East China Normal University (ICA-ECNU) 200241, Shanghai, China

会议类型: 国际会议

会议名称: 2010 Second Asia-Pacific Conference on Information Processing(2010年第二届亚太地区信息处理国际会议 APCIP 2010)

会议地点: 南昌

会议语种:英文

页码: 119-125

在线出版日期: 2010-09-17（万方平台首次上网日期，不代表论文的发表时间）