README.md

A Unifying Framework of Attention-based Neural Load Forecasting

Contact me with [email protected].

Introduction

The official code for paper [A Unifying Framework of Attention-based Neural Load Forecasting]

We propose a unifying deep learning framework for short-term load forecasting. Our novel approach includes time-varying feature weighting, hierarchical temporal attention, and feature-reinforced error correction. The proposed modular design features good generalization capability, which achieves superior performance over the SOTA.

Attention-based load forecasting Framework with LSTM implementation (PM-LSTM)

Run load forecasting model training, update model training and test result for ISO-NE dataset:

python ANLFF.py --data ISONE --final_run --updateCkpName checkpoint_update_ISO --logname ANLFF_ISO --subName ANLFF_ISO

Run load forecasting model training, update model training and test result for NAU dataset:

python ANLFF.py --data Utility --final_run --updateCkpName checkpoint_update_NAU --logname ANLFF_NAU --subName ANLFF_NAU

License

The project is only free for academic research purposes, but needs authorization for commerce. For commerce permission, please contact [email protected].

Citation

If you use our code/model, please cite our [paper].

@ARTICLE{10122506,
  author={Xiong, Jing and Zhang, Yu},
  journal={IEEE Access}, 
  title={A Unifying Framework of Attention-Based Neural Load Forecasting}, 
  year={2023},
  volume={11},
  number={},
  pages={51606-51616},
  doi={10.1109/ACCESS.2023.3275095}}
}