Deep Learning Network-Temporal Models For Traffic Prediction

arXiv cs.LG / 3/13/2026

📰 NewsIdeas & Deep AnalysisModels & Research

共有:

Key Points

The paper introduces two deep learning models for multivariate time series in network traffic: a customized network-temporal graph attention network (GAT) and a fine-tuned multi-modal large language model (LLM) with clustering, aiming to learn both temporal patterns and network topology.
Both models are evaluated against an LSTM baseline on a real-world network dataset, with the LLM-based model achieving superior overall prediction and generalization, while the GAT reduces prediction variance across time series and horizons.
The study reveals correlation variability and prediction distribution discrepancies over time series and across different prediction horizons.
The results suggest potential practical benefits for network intelligent control and management by enabling more accurate and robust traffic prediction.

Abstract

Time series analysis is critical for emerging net- work intelligent control and management functions. However, existing statistical-based and shallow machine learning models have shown limited prediction capabilities on multivariate time series. The intricate topological interdependency and complex temporal patterns in network data demand new model approaches. In this paper, based on a systematic multivariate time series model study, we present two deep learning models aiming for learning both temporal patterns and network topological correlations at the same time: a customized network-temporal graph attention network (GAT) model and a fine-tuned multi-modal large language model (LLM) with a clustering overture. Both models are studied against an LSTM model that already outperforms the statistical methods. Through extensive training and performance studies on a real-world network dataset, the LLM-based model demonstrates superior overall prediction and generalization performance, while the GAT model shows its strength in reducing prediction variance across the time series and horizons. More detailed analysis also reveals important insights into correlation variability and prediction distribution discrepancies over time series and different prediction horizons.