Exact and Robust Conformal Inference Methods for Predictive Machine Learning with Dependent Data

Abstract

We extend conformal inference to general settings that allow for time series data. Our proposal is developed as a randomization method and accounts for potential serial dependence by including block structures in the permutation scheme. As a result, the proposed method retains the exact, model-free validity when the data are i.i.d. or more generally exchangeable, similar to usual conformal inference methods. When exchangeability fails, as is the case for common time series data, the proposed approach is approximately valid under weak assumptions on the conformity score.

Cite

Text

Chernozhukov et al. "Exact and Robust Conformal Inference Methods for Predictive Machine Learning with Dependent Data." Annual Conference on Computational Learning Theory, 2018. doi:10.1920/WP.CEM.2018.1618

Markdown

[Chernozhukov et al. "Exact and Robust Conformal Inference Methods for Predictive Machine Learning with Dependent Data." Annual Conference on Computational Learning Theory, 2018.](https://mlanthology.org/colt/2018/chernozhukov2018colt-exact/) doi:10.1920/WP.CEM.2018.1618

BibTeX

@inproceedings{chernozhukov2018colt-exact,
  title     = {{Exact and Robust Conformal Inference Methods for Predictive Machine Learning with Dependent Data}},
  author    = {Chernozhukov, Victor and Wüthrich, Kaspar and Zhu, Yinchu},
  booktitle = {Annual Conference on Computational Learning Theory},
  year      = {2018},
  pages     = {732-749},
  doi       = {10.1920/WP.CEM.2018.1618},
  url       = {https://mlanthology.org/colt/2018/chernozhukov2018colt-exact/}
}