基於小波之水文時間序列多尺度熵複雜度分析

Wavelet-based Multi-scale Entropy Complexity Analysis of Hydrological Time Series

周建明(Chien-Ming Chou) ； 鄭士仁(Shin-Jen Cheng)

《明道學術論壇》 6卷1期 (2010/03) Pp. 43-57

https://doi.org/10.6953/MJ.201003.0042

摘要

水文時間序列之分析及模擬已有多種新穎方法與模式，然而大多數方法與模式均由水文時間序列自身進行分析處理，因此實有必要對水文時間序列進行多尺度研究。小波轉換與多尺度熵兩者都是常見之多尺度時間序列分析方法，本文應用上述兩種方法於水文時間序列，分析在多尺度下不同雨量站水文時間序列之差異。另外，藉由多尺度熵亦可判斷各測站最合適之小波分解層數，能解決分解層數如何確定之問題。研究結果說明小波轉換能擷取水文時間序列中之細微訊息，並闡釋多尺度熵在偵測水文時間序列隱藏的特徵之能力，臻以提供水資源規劃應用之參考。

關鍵字

小波轉換 ； 多尺度熵 ； 水文時間序列

並列摘要

Most of the methods and models applied hydrological time series themselves to carry out research. It is necessary to implement the multi-scale analysis to the modeling of hydrological time series. Wavelet transform and multi-scale entropy can be applied to the analysis of multiscale time series. The above methods are applied to the hydrological time series in this paper. Comparison among hydrological time series complexity analysis of different stations at multiscale is obtained. In addition, the resolution number of wavelet decomposition can be decided using multi-scale entropy. The results illustrate that the benefit of a wavelet transform of a hydrological time series lies in its capacity to highlight details of the signal with time-frequency resolution at different scales. Furthermore, it is worthwhile to note that analyzing hydrological time series using multi-scale entropy analysis can help us detect some hidden characteristics of the original time series. The above results can provide the reference for water resource planning and application.

並列關鍵字

Wavelet transform ； Multi-scale entropy ； Hydrological time series