A STAM-LSTM model for wind power prediction with feature selection

Wangbin Cao, Guangxing Wang, Xiaolin Liang and Zhengwei Hu

Energy, 2024, vol. 296, issue C

Abstract: In an effort to enhance the precision of wind power prediction, this study proposes a wind power prediction model with a secondary-weighted attention mechanism, which is based on feature selection. During the pre-processing stage, wavelet denoising is employed on the original dataset to eliminate noise and enhance the convergence rate of the model. As for model improvement, a secondary-weighted time attention mechanism-LSTM (STAM-LSTM) model is proposed. Additionally, the random forest algorithm is employed to analyse the feature correlation, leading to the best feature combination for the construction of the final input vector. In the comparative experiments, the STAM-LSTM model shows good performance and stability compared to the other nine models and three reference methods. In addition, to validate the effectiveness of feature selection, different combinations of features are entered for prediction. The results show that the model metrics reach a better level after feature selection. Finally, the effects of the STAM mechanism and the Random Forest feature screening assistance strategy, on the model are analysed through ablation experiments.

Keywords: Wavelet noise reduction; Hybrid model; Secondary-weighted attention mechanism; Feature selection; Wind power prediction (search for similar items in EconPapers)
Date: 2024
References: View references in EconPapers View complete reference list from CitEc
Citations:

Downloads: (external link)
http://www.sciencedirect.com/science/article/pii/S0360544224008028
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:296:y:2024:i:c:s0360544224008028

DOI: 10.1016/j.energy.2024.131030

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 ().