This is the pytorch implementation of STIDGCN. I hope these codes are helpful to you!
STIDGCN is accepted by TITS (IEEE Transactions on Intelligent Transportation Systems).
The code is built based on Python 3.9.12, PyTorch 1.11.0, and NumPy 1.21.2.
We provide preprocessed datasets that you can access here. If you need the original datasets, please refer to STSGCN (including PEMS03, PEMS04, PEMS07, and PEMS08) and ESG (including NYCBike and NYCTaxi).
It's easy to run! Here are some examples, and you can customize the model settings in train.py.
nohup python -u train.py --data PEMS08 > PEMS08.log &
nohup python -u train.py --data bike_drop > bike_drop.log &
Our model is built based on model of Graph WaveNet and SCINet.