文章摘要
改进的时间序列关键点表示算法
Improved Representation Alogrithm of Key Points in Times Series
  
DOI:10.16018/j.cnki.cn32-1650/n.202104008
中文关键词: 时间序列  关键点  转折点  
英文关键词: time series  key points  turning points
基金项目:
作者单位
刘永志 福州职业技术学院阿里巴巴大数据学院 南京航空航天大学计算机科学与技术学院 
摘要点击次数: 5644
全文下载次数: 4158
中文摘要:
      针对时间序列数据表示存在压缩率不高的问题,首先给出了时间序列定义和时间序列的9种基本形态;然后通过对极值点的优化处理,提出了关键极值点、剔除了轻微变化的极值点,并给出了结合转折点算法的IRAKPTS算法;最后通过实验验证,通过IRAKPTS算法的时间序列数据,较好地保留了时间序列的外形轮廓,并提高了压缩效率。
英文摘要:
      Aiming at the problem of low compression rate in time series data representation,firstly,the definition of time series and nine basic forms of time series are given. Then,through the optimization of extreme points,the key extreme points and the extreme points with slight changes are proposed,and the IRAKPTS algorithm combined with the turning point algorithm is given. Finally,it is verified by experiments that the time series data of the IRAPKTS algorithm retains the contours of the time series well and improves the compression efficiency.
查看全文   查看/发表评论  下载PDF阅读器
关闭