This repository contains the data and code required to reproduce the results in ICLR2025 paper "Error-quantified Conformal Inference for Time Series", where we borrow or extend some code from PID.

Please clone this repo and run following command locally for install the environment:

conda create --name eci
pip install -r requirements.txt

Please run following command locally for local test:

cd tests
python base_test.py configs/AMZN_test.yaml
python base_plots.py results/AMZN_test.pkl

The plot results will be saved in test/plots folder. More commands can be seen in tests/expbook.ipynb. Users can modify the YAML files in the configs folder to configure the experimental settings.

If you find this work useful, you can cite it with the following BibTex entry:

@inproceedings{wu2025error,
  title={Error-quantified Conformal Inference for Time Series},
  author={Wu, Junxi and Hu, Dongjian and Bao, Yajie and Xia, Shu-tao and Zou, Changliang},
  booktitle={The Thirteenth International Conference on Learning Representations},
  year={2025}
}