EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

A Bayesian-Optimized Mixture of Experts Framework for Short-Term Traffic Flow Prediction

Jianqing Wu, Jiaao Ren, Hui Wang, Fei Xie, Shaohan Chen, Mengjie Jiang

Modelling · 16.03.2026 · 7 (2) · 55

DOI: 10.3390/modelling7020055 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Modelling, online erschienen am 16.03.2026. Verfasst von Jianqing Wu und 5 weiteren Autor:innen. Die Arbeit nennt 27 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: National Natural Science Foundation of China, Doctoral Scientific Research Foundation of Jiangxi University of Science and Technology, Jiangxi Provincial Key Laboratory of Multidimensional Intelligent Perception and Control.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Accurate and reliable short-term traffic flow prediction is crucial for managing urban congestion but is challenged by the complex spatio-temporal dependencies inherent in traffic systems. Conventional single models, such as Long Short-Term Memory (LSTM) and Temporal Convolutional Network (TCN), often fail to capture these nonlinear dynamics. To address this, we propose a novel Bayesian-Optimized Mixture of Experts (BO-MoE) framework. This hybrid architecture utilizes a Mixture of Experts (MoE) to dynamically integrate multiple specialized deep learning models, allowing it to adapt to diverse and complex traffic patterns. Bayesian Optimization (BO) is further integrated to automate hyperparameter tuning, significantly enhancing predictive accuracy and model efficiency. We evaluated BO-MoE on three real-world traffic datasets. Empirical results demonstrate that our model consistently outperforms strong baselines, including TCN. Specifically, on PEMS04, it reduces MAE, RMSE, and MAPE by 1.97%, 1.19%, and 3.23%, respectively, while on PEMS08, the corresponding reductions reach 3.83%, 1.26%, and 5.49%. On the NZ dataset, BO-MoE also achieves superior performance, with improvements comparable to those on PEMS benchmarks.

Förderung

  • National Natural Science Foundation of China
  • Doctoral Scientific Research Foundation of Jiangxi University of Science and Technology
  • Jiangxi Provincial Key Laboratory of Multidimensional Intelligent Perception and Control

Zitieren

APA

Wu, J., Ren, J., Wang, H., Xie, F., Chen, S., & Jiang, M. (2026). A Bayesian-Optimized Mixture of Experts Framework for Short-Term Traffic Flow Prediction. Modelling, 7(2), 55. https://doi.org/10.3390/modelling7020055

Vancouver

Wu J, Ren J, Wang H, Xie F, Chen S, Jiang M. A Bayesian-Optimized Mixture of Experts Framework for Short-Term Traffic Flow Prediction. Modelling. 2026;7(2):55. doi:10.3390/modelling7020055

BibTeX
@article{Wu2026,
  author = {Wu, Jianqing and Ren, Jiaao and Wang, Hui and Xie, Fei and Chen, Shaohan and Jiang, Mengjie},
  title = {A Bayesian-Optimized Mixture of Experts Framework for Short-Term Traffic Flow Prediction},
  journal = {Modelling},
  year = {2026},
  volume = {7},
  number = {2},
  pages = {55},
  doi = {10.3390/modelling7020055}
}
Kostenlos starten →