摘要
目的设计基于DTW的高效过滤算法,提高时间序列数据流的过滤查询的效率.方法提出基于DTW的双边界的概念,并在此基础上定义新的更紧密的基于DTW的下界距离.结果实验证明基于DTW双边界的过滤算法在不发生错误丢失的情况下改进了算法的效率.对于模式间差异较大的情况,算法性能更好.结论基于DTW的双边界算法可以有效地过滤时间序列数据流.
In order to improve the efficiency of filtering algorithms for time series data stream, this paper proposes a new more efficient streaming time series query filtering algorithm for DTW. In the algorithm, Double Wedge for DTW is defined and a new more tight lower bounding distance based on DTW is introduced. Extensive experiments demonstrate that the algorithm has achieved tremendous improvements in the streaming time series query filtering with guaranteed no false dismissal. Especially, the larger differences predefined patterns have, the more efficient algorithm is. The result of the paper shows that based on double wedges for DTW it can effectively filter the time series data stream.
出处
《沈阳建筑大学学报(自然科学版)》
CAS
北大核心
2009年第6期1188-1192,共5页
Journal of Shenyang Jianzhu University:Natural Science
基金
国家十一五科技支撑计划项目(2008BAJ08B08-04
2006BAJ11B07-01
2006BAJ06B08-03)
辽宁自然科学基金项目(20071004)
辽宁省教育厅攻关计划(2008596
2008600)
关键词
数据流
过滤查询
DTW
下界距离
streams
query filtering
DTW
lower bounding distance