Large Language Models for Wireless Cellular Traffic Prediction: A Multi-timespan Approach
Bibliographic record
Abstract
Wireless cellular traffic prediction is essential for efficient network management and monitoring, yet it is a challenging task due to the spatial and temporal characteristics of traffic. Recently, machine learning based traffic prediction algorithms have been proposed in the literature. However, these algorithms lack good generalization ability as they cannot adapt to frequent changes in traffic distribution typically encountered in wireless networks. In this paper, we propose a traffic prediction algorithm using large language models (LLMs). We first analyze the temporal characteristics of traffic and identify those timespans in the historical traffic information which are important for traffic prediction. We use a clustering algorithm to identify cells with similar traffic patterns. To predict the traffic in a cell, we incorporate the multi-timespan historical traffic information of the cell as well as those cells with similar traffic patterns into natural language sentences and provide them as input to the LLM. Using our proposed framework, we fine-tune three popular LLMs (BART, BigBird, and PEGASUS) on the traffic prediction task. Experimental results show that our proposed LLM framework outperforms a state-of-the-art graph neural network (GNN) baseline and achieves up to 12.32% improvement in terms of the mean absolute error (MAE). Moreover, the proposed LLM framework has excellent generalization ability under the zero-shot setting, reducing the MAE by up to 46.84% compared to the baseline. The ablation studies reveal that providing information from multiple timespans to the model reduces the MAE by up to 15.05% compared to only providing information from the most recent timespan.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".