Skip to content

[Neural Networks] RGDAN: A random graph diffusion attention network for traffic prediction

Notifications You must be signed in to change notification settings

wengwenchao123/RGDAN

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

10 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

[Neural Networks] RGDAN: A random graph diffusion attention network for traffic prediction

This is a PyTorch implementation of RGDAN: A random graph diffusion attention network for traffic prediction, as described in our paper: Jin Fan, Weng, Wenchao, Hao Tian, Huifeng Wu , Fu Zhu, Jia Wu RGDAN: A random graph diffusion attention network for traffic prediction,Neural Networks 2024.

PWC PWC PWC

Note

The original code for this paper was lost due to server damage a year ago, and there was a lack of awareness to save relevant data at that time. The current code has been reconstructed based on memory to provide a version for research reference. While it achieves good results, it may not match the performance reported in the paper due to unknown reasons. We appreciate your understanding.

Data Preparation

The relevant datasets have been placed in the "data" folder. To run the program, simply unzip the "PeMS.zip" and "METR.zip" files.

Requirements

Python 3.6.5, Pytorch 1.9.0, Numpy 1.16.3, argparse and configparser

Model Training

# METR
python train.py --dataset METR --adjdata data/adj_mx.pkl

# PeMS
python train.py --dataset PeMS --adjdata data/adj_mx_bay.pkl

#BJ
python train_BJ.py 

Cite

If you find the paper useful, please cite as following:

@article{fan2024rgdan,
  title={RGDAN: A random graph diffusion attention network for traffic prediction},
  author={Fan, Jin and Weng, Wenchao and Tian, Hao and Wu, Huifeng and Zhu, Fu and Wu, Jia},
  journal={Neural networks},
  pages={106093},
  year={2024},
  publisher={Elsevier}
}

More Related Works

About

[Neural Networks] RGDAN: A random graph diffusion attention network for traffic prediction

Topics

Resources

Stars

Watchers

Forks

Packages

No packages published

Languages