Details
| Original language | English |
|---|---|
| Title of host publication | Proceedings of the 18TH Learning and Intelligent Optimization Conference (LION) |
| Editors | Paola Festa, Daniele Ferone, Tommaso Pastore, Ornella Pisacane |
| Pages | 456–471 |
| Number of pages | 16 |
| ISBN (electronic) | 978-3-031-75623-8 |
| Publication status | Published - 3 Jan 2025 |
Publication series
| Name | Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) |
|---|---|
| Volume | 14990 LNCS |
| ISSN (Print) | 0302-9743 |
| ISSN (electronic) | 1611-3349 |
Abstract
In today’s data-driven landscape, time series forecasting is pivotal in decision-making across various sectors. Yet, the proliferation of more diverse time series data, coupled with the expanding landscape of available forecasting methods, poses significant challenges for forecasters. To meet the growing demand for efficient forecasting, we introduce auto-sktime, a novel framework for automated time series forecasting. The proposed framework uses the power of automated machine learning (AutoML) techniques to automate the creation of the entire forecasting pipeline. The framework employs Bayesian optimization to automatically construct pipelines from statistical, machine learning (ML) and deep neural network (DNN) models. Furthermore, we propose three essential improvements to adapt AutoML to time series data. First, pipeline templates to account for the different supported forecasting models. Second, a novel warm-starting technique to start the optimization from prior optimization runs. Third, we adapt multi-fidelity optimizations to make them applicable to a search space containing statistical, ML and DNN models. Experimental results on 64 diverse real-world time series datasets demonstrate the effectiveness and efficiency of the framework, outperforming traditional methods while requiring minimal human involvement.
Keywords
- Automated Machine Learning, Forecasting, Time Series
ASJC Scopus subject areas
- Mathematics(all)
- Theoretical Computer Science
- Computer Science(all)
- General Computer Science
Cite this
- Standard
- Harvard
- Apa
- Vancouver
- BibTeX
- RIS
Proceedings of the 18TH Learning and Intelligent Optimization Conference (LION). ed. / Paola Festa; Daniele Ferone; Tommaso Pastore; Ornella Pisacane. 2025. p. 456–471 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 14990 LNCS).
Research output: Chapter in book/report/conference proceeding › Conference contribution › Research › peer review
}
TY - GEN
T1 - auto-sktime: Automated Time Series Forecasting
AU - Zöller, Marc
AU - Lindauer, Marius
AU - Huber, Marco
N1 - Marc Zöller was funded by the Federal Ministry for Economic Affairs and Climate Action in the project AutoQML, Marius Lindauer by the Federal Ministry for Economic Affairs and Energy of Germany in the project CoyPu (Grant 01MK21007L), and Marco Huber by the Baden-Wuerttemberg Ministry for Economic Affairs, Labour and Tourism in the project KI-Fortschrittszentrum “Lernende Systeme und Kognitive Robotik” (Grant 036-140100).
PY - 2025/1/3
Y1 - 2025/1/3
N2 - In today’s data-driven landscape, time series forecasting is pivotal in decision-making across various sectors. Yet, the proliferation of more diverse time series data, coupled with the expanding landscape of available forecasting methods, poses significant challenges for forecasters. To meet the growing demand for efficient forecasting, we introduce auto-sktime, a novel framework for automated time series forecasting. The proposed framework uses the power of automated machine learning (AutoML) techniques to automate the creation of the entire forecasting pipeline. The framework employs Bayesian optimization to automatically construct pipelines from statistical, machine learning (ML) and deep neural network (DNN) models. Furthermore, we propose three essential improvements to adapt AutoML to time series data. First, pipeline templates to account for the different supported forecasting models. Second, a novel warm-starting technique to start the optimization from prior optimization runs. Third, we adapt multi-fidelity optimizations to make them applicable to a search space containing statistical, ML and DNN models. Experimental results on 64 diverse real-world time series datasets demonstrate the effectiveness and efficiency of the framework, outperforming traditional methods while requiring minimal human involvement.
AB - In today’s data-driven landscape, time series forecasting is pivotal in decision-making across various sectors. Yet, the proliferation of more diverse time series data, coupled with the expanding landscape of available forecasting methods, poses significant challenges for forecasters. To meet the growing demand for efficient forecasting, we introduce auto-sktime, a novel framework for automated time series forecasting. The proposed framework uses the power of automated machine learning (AutoML) techniques to automate the creation of the entire forecasting pipeline. The framework employs Bayesian optimization to automatically construct pipelines from statistical, machine learning (ML) and deep neural network (DNN) models. Furthermore, we propose three essential improvements to adapt AutoML to time series data. First, pipeline templates to account for the different supported forecasting models. Second, a novel warm-starting technique to start the optimization from prior optimization runs. Third, we adapt multi-fidelity optimizations to make them applicable to a search space containing statistical, ML and DNN models. Experimental results on 64 diverse real-world time series datasets demonstrate the effectiveness and efficiency of the framework, outperforming traditional methods while requiring minimal human involvement.
KW - Automated Machine Learning
KW - Forecasting
KW - Time Series
UR - http://www.scopus.com/inward/record.url?scp=85216093690&partnerID=8YFLogxK
U2 - 10.1007/978-3-031-75623-8_35
DO - 10.1007/978-3-031-75623-8_35
M3 - Conference contribution
SN - 978-3-031-75622-1
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 456
EP - 471
BT - Proceedings of the 18TH Learning and Intelligent Optimization Conference (LION)
A2 - Festa, Paola
A2 - Ferone, Daniele
A2 - Pastore, Tommaso
A2 - Pisacane, Ornella
ER -