节点文献
一种时间序列连续分段多项式模式表示方法
Continuous piecewise polynomial model representation method of time series
【摘要】 在时间序列分段模式表示中,序列的分段多项式模式表示(PPR)是一种比分段线性化表示(PLR)误差更小的时间序列模式表示方法,但PPR在相邻分段点处存在间断非连续问题,给进一步的序列相似性搜索和子序列聚类分析,带来在分段点处数据处理的困难。针对此问题,分析产生这种非连续的原因,提出一种连续分段多项式模式表示(CPPR)方法。最后通过水质时间序列CPPR实验,证明了CPPR能够较好地解决PPR间断非连续问题,且成功应用于水质时间序列数据的预处理。
【Abstract】 In piecewise polynomial model representation of time series,the piecewise polynomial representation( PPR) of time series is a time series representation method with the error smaller than that of piecewise linear representation( PLR). However,the PPR has discontinuous problem at adjacent subsection points,which brings difficulties in the data processing at the subsection points for further time series comparability search and subsequent sub-series clustering analysis. Aiming at this problem,the reason causing the discontinuous problem is analyzed,and a continuous polynomial piecewise representation( CPPR) method is proposed. Finally,the experiment on water quality time series was conducted,experiment results prove that the CPPR method can solve the discontinuous problems of PPR method and has been applied in water quality time series data preprocessing successfully.
【Key words】 time series; piecewise model representation; piecewise polynomial; time series sub-section point;
- 【文献出处】 仪器仪表学报 ,Chinese Journal of Scientific Instrument , 编辑部邮箱 ,2014年05期
- 【分类号】O211.61
- 【被引频次】14
- 【下载频次】383