计算机工程与应用 ›› 2009, Vol. 45 ›› Issue (22): 4-6.DOI: 10.3778/j.issn.1002-8331.2009.22.002

• 博士论坛 • 上一篇    下一篇

时间序列度量的斜率偏离距离方法研究

梁建海1,张建业2,杨 峰,潘 泉1   

  1. 1.西北工业大学 自动化学院,西安 710072
    2.空军工程大学,西安 710051
  • 收稿日期:2009-04-08 修回日期:2009-05-08 出版日期:2009-08-01 发布日期:2009-08-01

Study of distance-measuring method in time series based on slope departure

LIANG Jian-hai1,ZHANG Jian-ye2,YANG Feng1,PAN Quan1
  

  1. 1.School of Automation,Northwestern Polytechnical University,Xi’an 710072,China
    2.Air Force Engineering University,Xi’an 710051,China
  • Received:2009-04-08 Revised:2009-05-08 Online:2009-08-01 Published:2009-08-01

摘要: 针对数据挖掘领域中时间序列的相似性度量问题,提出一种斜率复合偏离距离方法。以大量噪声的高维多元时间序列数据为目标,提出了一种基于斜率偏离度的时间序列相似性度量方法。该方法主要是在分段线性的基础上,基于角度和斜率进行偏离度计算,解决普通斜率距离度量的局限性,物理意义更为明确,实际度量更为准确。证明了斜率复合偏离的完备性和连续性,最后用仿真算例对算法的有效性进行了验证。

Abstract: Aiming at the measuring method of similarity search in time series,a similarity measuring method based on slope departure is proposed in this paper.To dispose the time series data with high dimension,multi-property,noise and default,a similarity measuring method is used to caculate the slope departure based on PLR solving the localization of slope distance measurement.In fact,the physical concept of this method proves more clearly and nicely.The maturity and the continuity of slope departure are proved.The example shows its validity.