Integrating data decomposition and machine learning methods: An empirical proposition and analysis for renewable energy generation forecasting

作者：

Highlights：

• An ensemble data-driven STL-LSTM forecasting framework is proposed.

• This proposed model is used for renewable energy generation forecasting.

• The new model can grasp the nonlinear, trend, and periodic patterns in datasets.

• The proposed method strikingly outperforms a range of prevalent benchmarks.

摘要

•An ensemble data-driven STL-LSTM forecasting framework is proposed.•This proposed model is used for renewable energy generation forecasting.•The new model can grasp the nonlinear, trend, and periodic patterns in datasets.•The proposed method strikingly outperforms a range of prevalent benchmarks.

论文关键词：Long-short term memory,Seasonal-trend decomposition procedure based on loess,Ensemble framework,Renewable energy generation,Grid flexibility enhancement

论文评审过程：Received 18 June 2021, Revised 9 May 2022, Accepted 19 May 2022, Available online 24 May 2022, Version of Record 28 May 2022.

论文官网地址：https://doi.org/10.1016/j.eswa.2022.117635