文章基本信息

标题：Autocorrelation Function in EOF Analysis
本地全文：下载
作者：Sujata Goswami ; Nico Sneeuw ; Kamal Jain 等
期刊名称：International Journal of Advanced Research in Computer Engineering & Technology (IJARCET)
印刷版ISSN：2278-1323
出版年度：2014
卷号：3
期号：5
页码：1790-1793
出版社：Shri Pannalal Research Institute of Technolgy
摘要：Empirical Orthogonal Function (EOF) analysis has been successfully applied in order to separate the noise from the time-series dataset. In EOF analysis, time-series is decomposed via singular value decomposition. Out of all these decomposed components some of them represent noise and some capture the dominant signals from it. Different rules has been defined to select the modes in order to recover noise free dataset as dominant variance rule, time-series based rules. Dominant variance rule was not able to select all the useful modes whereas time-series based rule sometimes were not able to distinguish clearly between the noise and signal. Then, the need to look into the autocorrelation function arises and finally it has been used as a mode selection tool in EOF analysis. Here, we have explained the use of autocorrelation function in EOF analysis and its advantage over other rules.
关键词：autocorrelation; EOF analysis; ; ; Kolmogorov-Smirnov hypothesis; Bartlett hypothesis; ; Principal Component Analysis (PCA)