改进的时间序列关键点表示算法

Improved Representation Alogrithm of Key Points in Times Series

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。

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.