关键词:  , 时间序列, 向量, 夹角余弦, 相似性度量, 离均差

Abstract:  The similarity measure of time series is a hot spot in the field of data mining. High dimensional multivariate time series data generally contain a large amount of noise which is not conducive to the comparison of similarity. In view of the problems in the existing time series measure method, the paper presents a new angle cosine algorithm based on the piecewise linear representation of time series shortly called ACMS. ACMS algorithm will be equivalent to a multi-dimensional vector of time series, the direction and size characteristics of the two vectors are fully considered, which enhances the robustness of time and amplitude variation, reduces human disturbance, and is helpful to the clustering and prediction in data mining.

Key words: time series, vector quantity, cosine algorithm, similarity measure, mean deviation