基于时间约束的Hausdorff距离的时空轨迹相似度量

摘要

从海量的轨迹数据中发现用户的相似轨迹是实现基于位置的个性化服务推荐的关键技术之一.位置和时间是轨迹数据的两个重要属性,而已有的基于Hausdorff距离的轨迹相似度量只着重关注于位置属性.针对在计算轨迹间相似性时忽略了轨迹的时间属性这一问题,提出一种基于时间约束的Hausdorff距离的时空轨迹相似度量,利用滑动窗口挖掘两条较长轨迹中所有相似的子轨迹,进而判断较长轨迹间的相似性.实验结果表明,该时空轨迹相似度量具有更好的效果.%Finding the similar trajectories in trajectory data was one of the key technologies to realize the personalized service recommendation.Position and time were two important attributes of trajectory data,and the existing Hausdorff distance,which were used to compare the similar trajectories,only focused on the location.Aiming to the problem which was ignoring time of trajectories,this paper proposed a new Hausdorff distance about spatio-temporal trajectory similarity based on time restriction.To compare the similarity of two long trajectories,all similar sub-trajectories were mined by using the sliding window in these trajectories.The experimental results show that the spatio-temporal trajectory similarity measure has better effect.