基于改进MOEN算法的时序数据主旨模式挖掘

Find Time Series Motifs Based on Improved MOEN Algorithm

主旨模式挖掘常用于发现时间序列中具有代表性的相似子序列,其中MOEN算法(Efficient Enumeration of Motifs,MOEN)基于枚举的思想来发现指定长度范围内的主旨模式(motifs),采用候选相似子序列的方法降低了计算所需资源。本研究对距离矩阵的生成策略加以改进,进一步降低计算成本,并通过实验验证其有效性。

Motifs mining is often used to find representative similar subsequences in time series.MOEN algorithm(efficiency enumeration of motifs,Moen)is based on the idea of enumeration to find the motifs within the specified length range.The method of candidate similar subsequences reduces the computing resources.In this study,the generation strategy of distance matrix is improved to further reduce the calculation cost,and its effectiveness is verified by experiments.