AGC-Net: Adaptive Graph Convolution Networks for Traffic Flow Forecasting

AGC-Net (Adaptive Graph Convolution Networks) is an advanced model designed to predict traffic flow. The paper is available here.

Citation

If you find this work useful for your research, please cite our paper:

@article{li2023adaptive,
      title={Adaptive Graph Convolution Networks for Traffic Flow Forecasting}, 
      author={Zhengdao Li and Wei Li and Kai Hwang},
      year={2023},
      eprint={2307.05517},
      archivePrefix={arXiv},
      primaryClass={cs.LG}
}

Installation

Before proceeding with the model training, ensure all necessary packages are installed. To install the requirements, run the following command:

pip install -r requirements -i http://mirrors.aliyun.com/pypi/simple/ --trusted-host mirrors.aliyun.com

Data Preparation

To prepare the data, please make sure you have the METR-LA dataset placed inside the ./data/ directory in a sub-directory named METR-LA-12. If you don't have the dataset, you can download it from the DCRNN (Note: replace with the appropriate link). The feature_len parameter denotes the feature length of the dataset. Here, we use a feature_len of 3.

Training

Once the data is prepared, you can train the AGC-Net model. The following is a sample command to initiate the training:

python main.py --predict_len=12 --cuda --att --data_path=./data/METR-LA-12 --feature_len=3 --wavelets_num=20 --transpose --epochs=1 --best_model_save_path=best_model_12_30w

Here is a brief explanation of the command-line arguments:

• --predict_len: The number of future time steps to be predicted by the model (12 in this case).

• --cuda: If present, use GPU for training.

• --att: If present, use the attention mechanism in the model.

• --data_path: The path to the directory where the dataset is stored.

• --feature_len: The feature length of the dataset.

• --wavelets_num: The number of wavelet functions to be used (20 in this case).

• --transpose: If present, transpose the input data.

• --epochs: The number of epochs to train the model.

• --best_model_save_path: The path where the model with the best validation performance should be saved.

The best_model_12_30w will be saved in the provided path upon successful training of the model.

Feel free to explore and adapt the model to suit your own requirements. We look forward to your contribution and feedback.

License

MIT