/ODKL

This resp presents a probabilistic and online forecasting model. In detail, a deep kernel is proposed by integrating the deep soft Spiking Neural Networks into the Gaussian kernel, which is then applied to perform sparse Gaussian Process regression.

Primary LanguagePythonGNU General Public License v3.0GPL-3.0

ODKL: An Online-Offline Deep Kernel Learning Method

Official Implementation of the paper published in IEEE Transactions on Power Systems Volume: 39, Issue: 2, March 2024

image

📖 Implementation:

  • Install the packages in requirement.txt
  • The demo.py shows how to apply the proposed model for load forecasting, to run this demo:
    • download file '2012-2013 Solar home electricity data v2.csv' from ausgrid resident dataset
    • run demo.py

🌟 Dataset:

3 publicly available residential load datasets are applied:

  • Ausgrid Resident: Loads of 300 households in the Australian distribution network are released for public utilization.
  • UMass Smart: Multiple smart meters’ readings of 7 homes are collected by the UMass Smart Home project in America from 2014 to 2016.
  • SGSC Customer Trial: This dataset stems from the Smart Grid Smart City (SGSC) project in Australia since 2010.

🤗 Citation

If you use ODKL in your research, please consider citing us.

@article{li2023residential,
  title={Residential load forecasting: An online-offline deep kernel learning method},
  author={Li, Yuanzheng and Zhang, Fushen and Liu, Yun and Liao, Huilian and Zhang, Hai-Tao and Chung, Chiyung},
  journal={IEEE Transactions on Power Systems},
  year={2023},
  publisher={IEEE}
}