Multi-layer Stack Ensembles for Time Series Forecasting

Nathanael Bosch, Oleksandr Shchur, Nick Erickson, Michael Bohlke-Schneider, Ali Caner Turkmen
Proceedings of the Fourth International Conference on Automated Machine Learning, PMLR 293:15/1-30, 2025.

Abstract

Ensembling is a powerful technique for improving the accuracy of machine learning models, with methods like stacking achieving strong results in tabular tasks. In time series forecasting, however, ensemble methods remain underutilized, with simple linear combinations still considered state-of-the-art. In this paper, we systematically explore ensembling strategies for time series forecasting. We evaluate 33 ensemble models—both existing and novel—across 50 real-world datasets. Our results show that stacking consistently improves accuracy, though no single stacker performs best across all tasks. To address this, we propose a multi-layer stacking framework for time series forecasting, an approach that combines the strengths of different stacker models. We demonstrate that this method consistently provides superior accuracy across diverse forecasting scenarios. Our findings highlight the potential of stacking-based methods to improve AutoML systems for time series forecasting.

Cite this Paper

BibTeX
@InProceedings{pmlr-v293-bosch25a, title = {Multi-layer Stack Ensembles for Time Series Forecasting}, author = {Bosch, Nathanael and Shchur, Oleksandr and Erickson, Nick and Bohlke-Schneider, Michael and Turkmen, Ali Caner}, booktitle = {Proceedings of the Fourth International Conference on Automated Machine Learning}, pages = {15/1--30}, year = {2025}, editor = {Akoglu, Leman and Doerr, Carola and van Rijn, Jan N. and Garnett, Roman and Gardner, Jacob R.}, volume = {293}, series = {Proceedings of Machine Learning Research}, month = {08--11 Sep}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v293/main/assets/bosch25a/bosch25a.pdf}, url = {https://proceedings.mlr.press/v293/bosch25a.html}, abstract = {Ensembling is a powerful technique for improving the accuracy of machine learning models, with methods like stacking achieving strong results in tabular tasks. In time series forecasting, however, ensemble methods remain underutilized, with simple linear combinations still considered state-of-the-art. In this paper, we systematically explore ensembling strategies for time series forecasting. We evaluate 33 ensemble models—both existing and novel—across 50 real-world datasets. Our results show that stacking consistently improves accuracy, though no single stacker performs best across all tasks. To address this, we propose a multi-layer stacking framework for time series forecasting, an approach that combines the strengths of different stacker models. We demonstrate that this method consistently provides superior accuracy across diverse forecasting scenarios. Our findings highlight the potential of stacking-based methods to improve AutoML systems for time series forecasting.} }
Endnote
%0 Conference Paper %T Multi-layer Stack Ensembles for Time Series Forecasting %A Nathanael Bosch %A Oleksandr Shchur %A Nick Erickson %A Michael Bohlke-Schneider %A Ali Caner Turkmen %B Proceedings of the Fourth International Conference on Automated Machine Learning %C Proceedings of Machine Learning Research %D 2025 %E Leman Akoglu %E Carola Doerr %E Jan N. van Rijn %E Roman Garnett %E Jacob R. Gardner %F pmlr-v293-bosch25a %I PMLR %P 15/1--30 %U https://proceedings.mlr.press/v293/bosch25a.html %V 293 %X Ensembling is a powerful technique for improving the accuracy of machine learning models, with methods like stacking achieving strong results in tabular tasks. In time series forecasting, however, ensemble methods remain underutilized, with simple linear combinations still considered state-of-the-art. In this paper, we systematically explore ensembling strategies for time series forecasting. We evaluate 33 ensemble models—both existing and novel—across 50 real-world datasets. Our results show that stacking consistently improves accuracy, though no single stacker performs best across all tasks. To address this, we propose a multi-layer stacking framework for time series forecasting, an approach that combines the strengths of different stacker models. We demonstrate that this method consistently provides superior accuracy across diverse forecasting scenarios. Our findings highlight the potential of stacking-based methods to improve AutoML systems for time series forecasting.
APA
Bosch, N., Shchur, O., Erickson, N., Bohlke-Schneider, M. & Turkmen, A.C.. (2025). Multi-layer Stack Ensembles for Time Series Forecasting. Proceedings of the Fourth International Conference on Automated Machine Learning, in Proceedings of Machine Learning Research 293:15/1-30 Available from https://proceedings.mlr.press/v293/bosch25a.html.

Related Material