时间序列的多尺度小波变换特征提取

Withdraw characteristic of time series with multiple scales wavele transform

本文基于小波变换,提出了一种提取时间序列特征的方法,用小波变换的低频逼近信号表示原时间序列。与傅里叶变换降维比较,该方法更能体现原时间序列的细部特征。

This paper proposed a method to get characteristic of time series which based on the wavelet transform.,It approaches the original times series with lower_frequency signal of wavelet transform.Compared with the dimensionality reduction of fourier transform,this method is better than fourier transform in express minute part of time series.