【摘要】 提出了一种新的基于重要点的分段方法,将时间序列数据转换为趋势序列。在进行相似性比较时先进行趋势相似的比较,然后对结果进行欧氏距离的比较。实验结果表明该算法是有效的。更多还原

【Abstract】 The paper proposes a segmentation method based on important points, by which time series data can betransformed into trend sequence.When we perform time series similarity searching, trend similarity may be calculatedfirst,then the results are compared based on Euclidean distance.Experiments show that this hierarchical approach is ef-fective.更多还原