小波分析在时间序列中的分析应用

Application of Wavelet Analysis in Time-series Analysis

为了研究时间序列真实的周期性和突变性,建立小波分解重构的消噪方法。本文以紫坪埔1951-2010年径流系列数据为例,先对样本时间序列进行小波分解,再逐层除去高频噪声部分,最后进行重构,获得全新的时间序列。在此基础上,对新时间序列进行连续小波变换,探讨小波变换系数随时间的变化过程,据此分析时间序列的奇异性(突变、跳跃等)变化。计算分析结果表明,提出的方法是可行的。

In order to study the real cyclical and mutability of time series, de-noising method of wavelet decomposition reconstruction is established in this paper. The runoff series data from 1951 to 2010 year is taken as an example. Firstly, the wavelet method is used to decompose the time series, then the high frequency noise section is removed step by step, finally, the sequence is reconstruc- ted to obtain a new time series. On this basis, continuous wavelet transform is conducted with the new time series and the change process of wavelet transform coefficient with the change of time is discussed to study the singularity, including mutations, jumping, etc. , of the time series changes. The calculation and analysis result shows that the proposed method is feasible.