/STGAGRTN

This is a PyTorch implementation of STGAGRTN in the Spatial-Temporal Graph Attention Gated Recurrent Transformer Network for Traffic Flow Forecasting.

Primary LanguagePython

STGAGRTN

This is a PyTorch implementation of STGAGRTN in the Spatial-Temporal Graph Attention Gated Recurrent Transformer Network for Traffic Flow Forecasting. Spatial-Temporal Graph Attention Gated Recurrent Transformer Network for Traffic Flow Forecasting has been published in the IEEE Internet of Things Journal.

figure1

核心文件介绍

  • prepareData.py 负责处理原始数据，并将其划分为训练集、验证集与测试集。运行完成后得到npz文件，data文件夹下包含处理得到的npz文件，无需重复运行。
  • train.py 负责训练模型，并生成params参数文件。04.params 与 08.params 分别是模型在PEMS04与PEMS08两个数据集下训练得到的参数文件。
  • predict.py 负责得到预测结果。
  • model.py 与 GAT.py 两个文件是模型主体部分。

Requirements

  • Python 3.8
  • numpy == 1.21.0
  • pandas == 1.2.4
  • torch == 1.10.2+cu113
  • torch-geometric == 2.0.4

Citation

If you find this repository useful in your research, please consider citing the following paper:

@ARTICLE{10347394,
  author={Wu, Di and Peng, Kai and Wang, Shangguang and Leung, Victor C. M.},
  journal={IEEE Internet of Things Journal}, 
  title={Spatial-Temporal Graph Attention Gated Recurrent Transformer Network for Traffic Flow Forecasting}, 
  year={2023},
  volume={},
  number={},
  pages={1-1},
  doi={10.1109/JIOT.2023.3340182}}