ORIGINAL RESEARCH article
Front. Commun. Netw.
Sec. Wireless Communications
Volume 6 - 2025 | doi: 10.3389/frcmn.2025.1658461
This article is part of the Research TopicMachine Learning-Based Spectrum Occupancy Prediction and Resource Allocation/Management for Wireless Communication SystemsView all 4 articles
Spatio-Temporal Beam-Level Traffic Forecasting in 5G Wireless Systems Using Multi-Task Learning
Provisionally accepted
- Prairie View A&M University, Prairie View, United States
Select one of your emails
You have multiple emails registered with Frontiers:
Notify me on publication
Please enter your email address:
If you already have an account, please login
You don't have a Frontiers account ? You can register here
Accurate beam-level traffic forecasting is critical for 5G network optimization, yet challenged by inherent data sparsity and multi-scale temporal patterns. This paper presents a Gated Recurrent Unit (GRU)-based Multi-task Learning (MTL) framework with ensemble enhancement for spatiotemporal traffic prediction. Through systematic evaluation of six models (Linear Regression, DLinear, XGBoost, Echo State Network (ESN), Long Short-Term Memory (LSTM), and GRU-MTL) across three sequence lengths (168/24/8-hour), we demonstrate that: (1) MTL outperforms conventional approaches (LSTM MAE=0.3223 vs. MTL MAE=0.2136 for 168-hour sequences), leveraging shared representations to address intermittent beam activity; (2) Longer sequences (168-hour) yield 56% lower MAE than 8-hour windows by mitigating sparsity effects; and (3) A weighted ensemble of top-performing models (MTL, XGBoost, Linear Regression) achieves further 1.45% improvement (MAE=0.2105). Using real-world beam-level data from ITU's AI for Good initiative, we establish practical guidelines: weekly (168-hour) horizons optimize accuracy for capacity planning, while our ensemble maintains robustness across temporal scales. These results provide a framework for sparse spatio-temporal forecasting in 5G networks, balancing accuracy with operational feasibility for resource allocation and congestion control.
Keywords: 5G, Traffic forecasting, time series prediction, GRU, Multi-task learning, LSTM, ESN, DLinear
Received: 02 Jul 2025; Accepted: 25 Aug 2025.
Copyright: © 2025 Tommy, Akinola, Li and Qian. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
* Correspondence:
Israel Tommy, Prairie View A&M University, Prairie View, United States
Lijun Qian, Prairie View A&M University, Prairie View, United States
Disclaimer: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.