Online Transfer Learning with MLP-Assisted Graph Convolution Network for Traffic Flow Forecasting: A Solution for Edge Intelligent Devices

Jingru Sun, Chendingying Lu, Yichuang Sun, Hongbo Jiang, Zhu Xiao

Research output: Contribution to journalArticlepeer-review

30 Downloads (Pure)

Abstract

Traffic flow prediction is crucial for intelligent transportation, aiding route planning and navigation. However, existing studies often focus on prediction accuracy improvement, while neglect external influences, and practical issues like resource constraints and data sparsity on edge devices. We propose an Online Transfer Learning with Multi-layer Perceptron (MLP) assisted Graph Convolution Network (GCN) framework (OTL-GM), consisting of two parts: transferring source domain features to edge devices and using online learning to bridge domain gaps. Experiments on 4 datasets demonstrate OTL’s effectiveness, and compared with the model without OTL, convergence time of the OTL-model increased from 24.77% to 95.32%.
Original languageEnglish
JournalFrontiers of Information Technology & Electronic Engineering
Publication statusAccepted/In press - 3 Mar 2025

Fingerprint

Dive into the research topics of 'Online Transfer Learning with MLP-Assisted Graph Convolution Network for Traffic Flow Forecasting: A Solution for Edge Intelligent Devices'. Together they form a unique fingerprint.

Cite this