统计与决策2025，Vol.41Issue(8)：59-63,5.
基于EEMD与WT组合的变点检测研究
Research on Change-point Detection Based on Combination of EEMD and WT
摘要
Abstract
Accurately detecting change point in data is of great value for statistical analysis.This paper constructs an adap-tive change-point detection method that does not require the observation data itself.By applying ensemble empirical mode decom-position to the original sequence and analyzing the variance contribution ratio,the main influencing factors of system evolution are determined.The excellent localization characteristics of Wavelet Transform(WT)are utilized to determine the location of the main influencing factors'change point.According to the temporal dependency relationship between the system change time and that of main influencing factors of the system,the change point of the original sequence is located.The results show that compared with classical regression analysis and direct use of WT to detect change point,the combined change-point detection method pro-posed in the paper can significantly improve the detection rate of change point.关键词
变点/聚合经验模态分解/主要影响因子/小波变换Key words
change point/ensemble empirical mode decomposition/main influencing factors/Wavelet transform分类
社会科学引用本文复制引用
郭建平,赵立龙..基于EEMD与WT组合的变点检测研究[J].统计与决策,2025,41(8):59-63,5.基金项目
国家社会科学基金一般项目(22BTJ061) （22BTJ061）
