摘要
在对时间序列经过符号化处理获得两类字符串集合的基础上,提出了一种非同步时间序列特征模式提取算法,该算法的核心是利用约简方法快速找到一个能充分区分两类字符串集合的模式,从而可以发现所有包含此模式的时间序列。实验结果表明,该算法在解决诸如从大量的飞行数据中搜索设备故障模式或飞行状态模式等问题时具有良好的应用前景。
On the basis of two sets of strings ,which are produced by symbol representation of time series, an algorithm for finding pattern of non synchronous time series is put forward. The main idea of the algorithm is to obtain fleetly a pattern which can distinguish two sets of strings completely by the use of pruning method. So every time series including this pattern can be find quickly. The experiment result demonstrates that the algorithm has a good application foreground in the area of search device fault pattern and flight state pattern of aircraft from large numbers of flight data.
出处
《计算机应用研究》
CSCD
北大核心
2005年第5期87-89,共3页
Application Research of Computers
关键词
时间序列
符号化
部分模式匹配
特征提取
Time Series
Symbol Representation
Episode Pattern Matching
Characteristic Drawing
