Probabilistic Reconciliation of Hierarchical Forecast via Bayes' Rule

Giorgio Corani, Dario Azzimonti, João P. S. C. Augusto, Marco Zaffalon

ECML-PKDD 2020 pp. 211-226

doi:10.1007/978-3-030-67664-3_13 /ecmlpkdd/2020/corani2020ecmlpkdd-probabilistic/

Abstract

We present a novel approach for reconciling hierarchical forecasts, based on Bayes rule. We define a prior distribution for the bottom time series of the hierarchy, based on the bottom base forecasts. Then we update their distribution via Bayes rule, based on the base forecasts for the upper time series. Under the Gaussian assumption, we derive the updating in closed-form. We derive two algorithms, which differ as for the assumed independencies. We discuss their relation with the MinT reconciliation algorithm and with the Kalman filter, and we compare them experimentally.

PDF ECML-PKDD Semantic Scholar

Cite

Text

Corani et al. "Probabilistic Reconciliation of Hierarchical Forecast via Bayes' Rule." European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, 2020. doi:10.1007/978-3-030-67664-3_13

Markdown

[Corani et al. "Probabilistic Reconciliation of Hierarchical Forecast via Bayes' Rule." European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, 2020.](https://mlanthology.org/ecmlpkdd/2020/corani2020ecmlpkdd-probabilistic/) doi:10.1007/978-3-030-67664-3_13

BibTeX

@inproceedings{corani2020ecmlpkdd-probabilistic,
  title     = {{Probabilistic Reconciliation of Hierarchical Forecast via Bayes' Rule}},
  author    = {Corani, Giorgio and Azzimonti, Dario and Augusto, João P. S. C. and Zaffalon, Marco},
  booktitle = {European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases},
  year      = {2020},
  pages     = {211-226},
  doi       = {10.1007/978-3-030-67664-3_13},
  url       = {https://mlanthology.org/ecmlpkdd/2020/corani2020ecmlpkdd-probabilistic/}
}