EnsembleCI

Installation

To install the necessary dependencies, please follow the instructions at AutoGluon Installation Guide.

Usage

Running carbonForecast.py

To run the carbon forecasting script, use the following command in src:

python3 carbonForecast.py <region> <d/l> <model> <l/t>

Example: python3 carbonForecast.py CISO d GBM t

<region>:

US:

• CISO: California ISO
• PJM: Pennsylvania-Jersey-Maryland Interconnection
• ERCO: Electric Reliability Council of Texas
• ISNE: ISO New England
• EPE: El Paso Electric
• MISO: Midcontinent Independent System Operator

Europe:

• DE: Germany
• SE: Sweden
• PL: Poland
• NL: Netherlands
• ES: Spain

<d/l>:

• d: direct_emission
• l: lifecycle_emission

<model>:

• GBM: LightGBM
• FASTAI: neural network with FastAI backend （https://auto.gluon.ai/stable/_modules/autogluon/tabular/models/fastainn/tabular_nn_fastai.html#NNFastAiTabularModel）
• CAT: CatBoost
• XGB: XGBoost
• AUTO: Autogluon weightedEnsamble

<l/t> (optional)

• l: load existing model
• t: train new model

Citing EnsembleCI

If you use EnsembleCI, please consider citing our paper. The BibTex format is as follows:

@inproceedings{yan2025ensembleCI,
  title = {EnsembleCI: Ensemble Learning for Carbon Intensity Forecasting},
  author = {Yan, Leyi and Wang, Linda and Liu, Sihang and Ding, Yi},
  booktitle = {Proceedings of the 15th ACM International Conference on Future and Sustainable Energy Systems (e-Energy)},
  year = {2025},
  doi = {10.1145/3679240.3734630},
}`