This is a testing PyTorch version implementation of Graph Multi-Attention Network in the following paper: Chuanpan Zheng, Xiaoliang Fan, Cheng Wang, and Jianzhong Qi.
Refer to the implementation of VincLee8188
- Python
- PyTorch
- Pandas
- Matplotlib
- Numpy
- gensim
- networkx
在Perms-bay这个数据集上效果不错,测试结果有MAE:1.93(预测60分钟),但在METR-LA数据集上效果不是很好,不清楚是我的代码实现有问题,还是其他原因,如果有大佬能指出原因,不胜感激
This version of implementation is only for learning purpose. For research, please refer to and cite from the following paper:
@inproceedings{ GMAN-AAAI2020,
author = "Chuanpan Zheng and Xiaoliang Fan and Cheng Wang and Jianzhong Qi"
title = "GMAN: A Graph Multi-Attention Network for Traffic Prediction",
booktitle = "AAAI",
pages = "1234--1241",
year = "2020"
}