基于时空标签轨迹的k近邻模式匹配查询knearestneighborpattern.PDF

第39 卷第4 期 通 信 学 报 Vol.39 No.4 2018 年4 月 Journal on Communications April 2018 基于时空标签轨迹的k 近邻模式匹配查询 许建秋，梁珺秀，秦小麟 （南京航空航天大学计算机科学与技术学院，江苏南京 211106 ） 摘 要：时空标签轨迹在传统的时空轨迹数据基础之上融入了具有语义含义的标签信息，丰富了移动对象数据。 针对该数据提出k 近邻模式匹配查询，即在给定时间区间内匹配相应的模式且距离查询轨迹最近的k 条轨迹。设 计并实现标签R 树（LR-Tree ），即增加标签表并在R 树每项中添加标签位图，及基于LR-Tree 的k 近邻模式匹配 查询算法。通过真实数据和合成数据将LR-Tree 与3DR-Tree、SETI 及TB-Tree 进行对比，实验表明LR-Tree 具有 更好的剪枝能力，从而验证了所提算法及索引的有效性。 关键词：时空标签轨迹；k 近邻算法；模式匹配；索引 中图分类号：TP311 文献标识码：A doi: 10.11959/j.issn.1000-436x.2018063 k nearest neighbor pattern match queries over spatio-temporal label trajectories XU Jianqiu, LIANG Junxiu, QIN Xiaolin College of Computer Science and Technology, Nanjing University of Aeronautics and Astronautics, Nanjing 211106, China Abstract: Spatio-temporal label trajectories extended traditional spatio-temporal trajectories with semantic labels. k nearest neighbor pattern match was proposed to return the k nearest trajectories that fulfilled the temporal pattern condi- tion. The Label R-Tree (LR-Tree for short) was proposed, which appending a label table and adding label bitmap in each entry, and k nearest neighbor pattern match query algorithm based on LR-Tree was designed. Using both real and syn- thetic datasets, the LR-Tree was extensively evaluated in comparison with 3DR-Tree, SETI and TB-Tree. The experimental re- sults demonstrate that LR-Tree showing better pruning ability, and verify the effectiveness of proposed algorithm and index. Key words: spatio-temporal label trajectories, k nearest neighbor algorithm, pattern match, index