A decomposition-ensemble model with regrouping method and attention-based gated recurrent unit network for energy price prediction

Hongli Niu, Kunliang Xu and Cheng Liu

Energy, 2021, vol. 231, issue C

Abstract: Accurately forecasting the energy price has increasingly attracted attention of researchers. A novel hybrid forecasting model, termed as ICEEMDAN-R-AttGRU, is for the first time proposed to enhance the forecasting accuracy of energy price series (Brent oil, DCE coke and NYMEX natural gas). In the proposed model, the improved CEEMDAN is taken to decompose the raw prices into multiple subcomponents. Then, a novel regrouping method based on frequency is put forward to reconstruct the subcomponents to reduce the forecasting workload and diminish the chance of errors. The attention-based GRU network is adopted to perform the forecasting task for each component, in which the attention mechanism is applied to allocate and optimize weights to the input elements in GRU. The empirical results measured by various performance metrics verify that the predictive accuracy is evidently improved by ICEEMDAN-R-AttGRU model in most cases compared with single models, ICEEMDAN-based and ICEEMDAN-R-based hybrid models. Besides, a new multi-scale composite complexity synchronization (MCCS) statistic is introduced into model measurement, which further confirms the competitive prediction ability of ICEEMDAN-R-AttGRU in different exponent and time scales.

Keywords: Hybrid models; Attention mechanism; Regrouping; Gated recurrent units; Energy price forecasting (search for similar items in EconPapers)
Date: 2021
References: View references in EconPapers View complete reference list from CitEc
Citations: View citations in EconPapers (10)

Downloads: (external link)
http://www.sciencedirect.com/science/article/pii/S0360544221011890
Full text for ScienceDirect subscribers only

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:eee:energy:v:231:y:2021:i:c:s0360544221011890

DOI: 10.1016/j.energy.2021.120941

Access Statistics for this article

Energy is currently edited by Henrik Lund and Mark J. Kaiser

More articles in Energy from Elsevier
Bibliographic data for series maintained by Catherine Liu ().