Wireless Traffic Prediction with Large Language Model

TL;DR

This paper tackles city-scale wireless traffic forecasting, highlighting the limitations of purely temporal models to capture spatial dependencies. It introduces TIDES, a two-phase framework that clusters regions by traffic similarity and then applies an LLM backbone (DeepSeek) with prompt-based representations and a cross-domain, spatial-attention mechanism to jointly model spatial-temporal dynamics. The approach combines region-aware clustering, RevIN-based prompts, graph-Laplacian filtering, and cross-modal alignment to yield accurate, robust predictions with efficient fine-tuning. Experiments on real-world city data show that TIDES significantly outperforms state-of-the-art baselines across multiple metrics and scenarios, signaling strong potential for autonomous, scalable network management in 6G environments.

Abstract

The growing demand for intelligent, adaptive resource management in next-generation wireless networks has underscored the importance of accurate and scalable wireless traffic prediction. While recent advancements in deep learning and foundation models such as large language models (LLMs) have demonstrated promising forecasting capabilities, they largely overlook the spatial dependencies inherent in city-scale traffic dynamics. In this paper, we propose TIDES (Traffic Intelligence with DeepSeek-Enhanced Spatial-temporal prediction), a novel LLM-based framework that captures spatial-temporal correlations for urban wireless traffic prediction. TIDES first identifies heterogeneous traffic patterns across regions through a clustering mechanism and trains personalized models for each region to balance generalization and specialization. To bridge the domain gap between numerical traffic data and language-based models, we introduce a prompt engineering scheme that embeds statistical traffic features as structured inputs. Furthermore, we design a DeepSeek module that enables spatial alignment via cross-domain attention, allowing the LLM to leverage information from spatially related regions. By fine-tuning only lightweight components while freezing core LLM layers, TIDES achieves efficient adaptation to domain-specific patterns without incurring excessive training overhead. Extensive experiments on real-world cellular traffic datasets demonstrate that TIDES significantly outperforms state-of-the-art baselines in both prediction accuracy and robustness. Our results indicate that integrating spatial awareness into LLM-based predictors is the key to unlocking scalable and intelligent network management in future 6G systems.

Figures (7)

• Figure 1: Traffic patterns vary greatly among different places. Left: City boundary and the three selected areas, i.e., WuJiaBu, PuJi, and ShuangQuan. Middle: The temporal traffic dynamics of WuJiaBu, PuJi, and ShuangQuan. Right: Pearson correlation coefficient of forty randomly selected areas from the city.
• Figure 2: Wireless traffic prediction with a two-phase process. The first phase (left) involves a spatial-aware clustering process, and the second phase (right) utilizes the TIDES framework, which performs LLM-driven spatial-temporal learning.
• Figure 3: Correlation coefficient between predictions and ground truth values. The results from left to right are obtained by Time-LLM, DLinear, and our proposed TIDES, respectively.
• Figure 4: Normalized prediction error distribution of different algorithms.
• Figure 5: City-level prediction performance comparison from the spatial view (upper side) and temporal view (lower side).