EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Extreme-Aware Time-Series Forecasting via Weak-Label-Guided Mixture of Experts

Jialou Wang iD, Jacob Sanderson iD, Wai Lok Woo iD

Sensors · 02.03.2026 · 26 (5) · 1571

DOI: 10.3390/s26051571 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Sensors, online erschienen am 02.03.2026. Verfasst von Jialou Wang und 2 weiteren Autor:innen. Beteiligte Einrichtung: School of Computer Science, Northumbria University, Newcastle upon Tyne NE1 8ST, UK. Die Arbeit wurde bisher 5-mal zitiert (Crossref) und nennt 43 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Department for Environment Food and Rural Affairs, UK Department for Environment, Food and Rural Affairs.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Deep time-series forecasting models can achieve strong average accuracy under normal conditions, yet they often struggle with rare, high-impact extremes, where severe class imbalance biases learning toward majority dynamics. Although infrequent, these extremes frequently correspond to critical events such as natural disasters or power outages. We address this challenge with a weak-label-guided mixture of experts (WL-MoE) that routes each input window to lightweight specialists designed to capture distinct temporal regimes. To prevent routing collapse during early optimisation, WL-MoE follows a two-stage training curriculum. In Stage I, cluster-derived weak labels encourage diverse expert utilisation and promote specialisation under imbalance. In Stage II, guidance is removed and training proceeds solely with the forecasting objective, ensuring that inferences remain fully data-driven. The expert-based structure also supports interpretable routing via expert-usage profiling, enabling regime-level auditing of model behaviour in high-stakes settings. Across seven benchmark datasets, WL-MoE reduces the average MSE by approximately 7.9% and the extreme-case MSE by approximately 23.58% relative to the best baseline. In a UK flood forecasting study, it reduces the all-water MSE by 31.6% and the high-water MSE by approximately 35.0%. These results indicate that weak-label guidance can stabilise specialisation and improve reliability under rare extremes while keeping model behaviour auditable for real-world deployment.

Schlagwörter

Stichwörter der Autor:innen

extreme-aware time-series forecastingheuristic unsupervised routing clusteringweak-label-guided mixture of experts

Autor:innen und Institutionen

  • School of Computer Science, Northumbria University, Newcastle upon Tyne NE1 8ST, UK.

ORCID-Profile: Jialou Wang, Jacob Sanderson, Wai Lok Woo

Förderung

  • Department for Environment Food and Rural Affairs
  • UK Department for Environment, Food and Rural Affairs

Zitieren

APA

Wang, J., Sanderson, J., & Woo, W. L. (2026). Extreme-Aware Time-Series Forecasting via Weak-Label-Guided Mixture of Experts. Sensors, 26(5), 1571. https://doi.org/10.3390/s26051571

Vancouver

Wang J, Sanderson J, Woo WL. Extreme-Aware Time-Series Forecasting via Weak-Label-Guided Mixture of Experts. Sensors. 2026;26(5):1571. doi:10.3390/s26051571

BibTeX
@article{Wang2026,
  author = {Wang, Jialou and Sanderson, Jacob and Woo, Wai Lok},
  title = {Extreme-Aware Time-Series Forecasting via Weak-Label-Guided Mixture of Experts},
  journal = {Sensors},
  year = {2026},
  volume = {26},
  number = {5},
  pages = {1571},
  doi = {10.3390/s26051571}
}
Kostenlos starten →