A novel hybrid decomposition-and-ensemble model based on CEEMD and GWO for short-term PM2.5 concentration forecasting

doi:10.1016/j.atmosenv.2016.03.056

	A novel hybrid decomposition-and-ensemble model based on CEEMD and GWO for short-term PM2.5 concentration forecasting
	Niu, Mingfei 1; Wang, Yufang 1; Sun, Shaolong 2; Li, Yongwu 2
刊名	ATMOSPHERIC ENVIRONMENT
	2016-06-01
卷号	134 页码:168-180
关键词	Complementary ensemble empirical mode decomposition Grey wolf optimizer Support vector regression Hybrid decomposition-ensemble model PM2.5 concentration forecasting
ISSN号	1352-2310
DOI	10.1016/j.atmosenv.2016.03.056
英文摘要	To enhance prediction reliability and accuracy, a hybrid model based on the promising principle of "decomposition and ensemble" and a recently proposed meta-heuristic called grey wolf optimizer (GWO) is introduced for daily PM2.5 concentration forecasting. Compared with existing PM2.5 forecasting methods, this proposed model has improved the prediction accuracy and hit rates of directional prediction. The proposed model involves three main steps, i.e., decomposing the original PM2.5 series into several intrinsic mode functions (IMFs) via complementary ensemble empirical mode decomposition (CEEMD) for simplifying the complex data; individually predicting each IMF with support vector regression (SVR) optimized by GWO; integrating all predicted IMFs for the ensemble result as the final prediction by another SVR optimized by GWO. Seven benchmark models, including single artificial intelligence (AI) models, other decomposition-ensemble models with different decomposition methods and models with the same decomposition-ensemble method but optimized by different algorithms, are considered to verify the superiority of the proposed hybrid model. The empirical study indicates that the proposed hybrid decomposition-ensemble model is remarkably superior to all considered benchmark models for its higher prediction accuracy and hit rates of directional prediction. (C) 2016 Elsevier Ltd. All rights reserved.
资助项目	National Natural Science Foundation of China[71501176] ; China Postdoctoral Science Foundation[2015M580141]
WOS研究方向	Environmental Sciences & Ecology ; Meteorology & Atmospheric Sciences
语种	英语
出版者	PERGAMON-ELSEVIER SCIENCE LTD
WOS记录号	WOS:000375504100017
内容类型	期刊论文
源URL	[http://ir.amss.ac.cn/handle/2S8OKBNM/22605]
专题	中国科学院数学与系统科学研究院
通讯作者	Sun, Shaolong
作者单位	1.Lanzhou Univ, Sch Math & Stat, Lanzhou 730000, Peoples R China 2.Chinese Acad Sci, Acad Math & Syst Sci, Beijing 100190, Peoples R China
推荐引用方式 GB/T 7714	Niu, Mingfei,Wang, Yufang,Sun, Shaolong,et al. A novel hybrid decomposition-and-ensemble model based on CEEMD and GWO for short-term PM2.5 concentration forecasting[J]. ATMOSPHERIC ENVIRONMENT,2016,134:168-180.
APA	Niu, Mingfei,Wang, Yufang,Sun, Shaolong,&Li, Yongwu.(2016).A novel hybrid decomposition-and-ensemble model based on CEEMD and GWO for short-term PM2.5 concentration forecasting.ATMOSPHERIC ENVIRONMENT,134,168-180.
MLA	Niu, Mingfei,et al."A novel hybrid decomposition-and-ensemble model based on CEEMD and GWO for short-term PM2.5 concentration forecasting".ATMOSPHERIC ENVIRONMENT 134(2016):168-180.