Title :
A fast trend extraction for the analysis of temperature data
Author :
Yang Da ; Wang Xiaotong ; Xu Guanlei ; Su Shipeng
Author_Institution :
Navig. Dept., Dalian Navy Acad., Dalian, China
Abstract :
Trend extraction is one of the major contents of time series analysis. This paper employs a novel trend extraction method based on multi-scale extrema of signals to analyze the trend of temperature data. This approach is model-free, adaptive, fast, flexible and free of sifting-process applied in empirical mode decomposition (EMD). The practical temperature data series is analyzed and the changing trend can be extracted as fast as possible. In addition, the comparison with other methods based EMD is also presented to show the advantages of the proposed method in application of trend extraction and analysis for temperature data.
Keywords :
filtering theory; time series; EMD; empirical mode decomposition; multiscale extrema; nonparametric linear filtering approach; novel fast trend extraction method; singular spectrum analysis; temperature data; time series analysis; Binary trees; Data mining; Interpolation; Market research; Temperature distribution; Time series analysis;
Conference_Titel :
Intelligent Control and Information Processing (ICICIP), 2013 Fourth International Conference on
Conference_Location :
Beijing
Print_ISBN :
978-1-4673-6248-1
DOI :
10.1109/ICICIP.2013.6568094
