Photonics Global Conference 2025Spatio-temporal forecasting Zhengzhou Beijing University of Posts and

An Edge-Aware Graph Attention and Transformer Network for LEO Satellite Optical Networks Traffic Prediction

ST-EGAT predicts LEO satellite optical network traffic via edge-aware graph attention and transformers, improving MAE, RMSE, and MAPE.

Zihao Qin, Shaobo Qin, Pengfei Lv, Yajuan Qin, Yanyan Xie, Xiaoke Sun, Haicheng Li, Ruijie Zhu

Published Dec 5, 2025Paper ↗

65%

OverallHighly rated

OverallHighly ratedVote to see the score

Readers

–

Only vote on papers you've read. Sign in to vote.

AI panel2/20reviewers recommend it

lenient 1/5

medium 0/10

strict 1/5

AI panel?Vote to see what the 20 AI reviewers said

Abstract

We propose a Spatio-Temporal Edge-aware Graph Attention and Transformer Network (ST-EGAT) for satellite network traffic prediction. Experimental results demonstrate that our model improves MAE, RMSE and MAPE metrics.