基于形态相似距离的时间序列相似性度量

Morphology similarity distance based time series similarity measurement

时间序列的相似性度量是时间序列分析的基础工作之一,是进行相似匹配的关键。针对欧几里德距离描述分段趋势的不足和各种模式距离对应分段之间距离值的离散化问题,提出一种基于形态相似距离的时间序列相似性度量方法,标准数据集上完成的识别和聚类实验表明了该方法的可行性和有效性。

Time series similarity measurement is one of the fundamental tasks in time series data analyzing, and the key to similarity matching. In view of shortcomings of Euclidean distance can not compare segment trend similarity and pattern distance measure or its transformations existing discretization problem, the morphology similarity distance based time series similarity measurement is presented in this paper. Experimental results of reorganization and clustering on standard data sets show that the proposed method is feasible and effective.