EasyDGL

The official implementation for "EasyDGL: Encode, Train and Interpret for Continuous-time Dynamic Graph Learning".

What's news

[2023.07.28] We release the pre-version of Pytorch-DGL for traffic forecasting.

[2023.07.10] We release the pre-version of Pytorch-DGL for link prediction.

Results for Link Prediction

Results

Below we report the HR@50, NDCG@50 and NDCG@100 results on the above provided dataset.

Model	HR@50	NDCG@50	NDCG@100
EasyDGL (Tensorflow)	0.48320	0.23104	0.24476
EasyDGL (Torch)	0.48252	0.23055	0.24378

Results for Traffic Forecasting

Results

Below we summarize the performance with regards to different metrics and horizons on the METR-LA dataset (see DCRNN).

		H=3	H=6	H=12
DCRNN	MAE	2.77	3.15	3.60
	RMSE	5.38	6.45	7.60
	MAPE(%)	7.30	8.80	10.50
AGCRN	MAE	2.87	3.23	3.62
	RMSE	5.58	6.58	7.51
	MAPE(%)	7.70	9.00	10.38
EasyDGL	MAE	2.74	3.10	3.55
	RMSE	5.21	6.17	7.17
	MAPE(%)	6.96	8.19	9.81

Folder Specification

• conf/: configurations for logging
• data/: dataset for model training and evaluation
• runme.sh: train or evaluate EasyDGL and baseline models
• src/: codes for model definition

Run the Code

Here are commands for training the model for link prediction and traffic forecasting respectively.

python src/demo_recsys.py --config conf/model/Netflix/EasyDGL.yaml

python src/demo_traffic.py --config conf/model/METR-LA/EasyDGL.yaml

Citation

If you find our codes useful, please consider citing our work

@inproceedings{chen2021learning,
  title={Learning Self-Modulating Attention in Continuous Time Space with Applications to Sequential Recommendation},
  author={Chen, Chao and Geng, Haoyu and Yang, Nianzu and Yan, Junchi and Xue, Daiyue and Yu, Jianping and Yang, Xiaokang},
  booktitle={Proceedings of the International Conference on Machine Learning (ICML '21)},
  pages={1606--1616},
  year={2021},
  organization={PMLR}
}

@article{chen2023easydgl,
  title={EasyDGL: Encode, Train and Interpret for Continuous-time Dynamic Graph Learning},
  author={Chen, Chao and Geng, Haoyu and Yang, Nianzu and Yang, Xiaokang and Yan, Junchi},
  journal={arXiv preprint arXiv:2303.12341},
  year={2023}
}