Skip to content

Latest commit

 

History

History
112 lines (74 loc) · 5.6 KB

README.md

File metadata and controls

112 lines (74 loc) · 5.6 KB
Raw

Decoupled Dynamic Spatial-Temporal Graph Neural Network for Traffic Forecasting

Code for our VLDB'22 paper: "Decoupled Dynamic Spatial-Temporal Graph Neural Network for Traffic Forecasting".

D2STGNN

We all depend on mobility, and vehicular transportation affects the daily lives of most of us. Thus, the ability to forecast the state of traffic in a road network is an important functionality and a challenging task. Traffic data is often obtained from sensors deployed in a road network. Recent proposals on spatial-temporal graph neural networks have achieved great progress at modeling complex spatial-temporal correlations in traffic data, by modeling traffic data as a diffusion process. However, intuitively, traffic data encompasses two different kinds of hidden time series signals, namely the diffusion signals and inherent signals. Unfortunately, nearly all previous works coarsely consider traffic signals entirely as the outcome of the diffusion, while neglecting the inherent signals, which impacts model performance negatively. To improve modeling performance, we propose a novel Decoupled Spatial-Temporal Framework (DSTF) that separates the diffusion and inherent traffic information in a data-driven manner, which encompasses a unique estimation gate and a residual decomposition mechanism. The separated signals can be handled subsequently by the diffusion and inherent modules separately. Further, we propose an instantiation of DSTF, Decoupled Dynamic Spatial-Temporal Graph Neural Network (D2STGNN), that captures spatial-temporal correlations and also features a dynamic graph learning module that targets the learning of the dynamic characteristics of traffic networks. Extensive experiments with four real-world traffic datasets demonstrate that the framework is capable of advancing the state-of-the-art.

1. Table of Contents

configs         ->  training Configs and model configs for each dataset
dataloader      ->  pytorch dataloader
datasets        ->  raw data and processed data
model           ->  model implementation and training pipeline
output          ->  model checkpoint

2. Requirements

pip install -r requirements.txt

3. Data Preparation

3.1 Download Data

For convenience, we package these datasets used in our model in Google Drive or BaiduYun.

They should be downloaded to the code root dir and replace the raw_data and sensor_graph folder in the datasets folder by:

cd /path/to/project
unzip raw_data.zip -d ./datasets/
unzip sensor_graph.zip -d ./datasets/
rm {sensor_graph.zip,raw_data.zip}
mkdir log output

Alterbatively, the datasets can be found as follows:

  • METR-LA and PEMS-BAY: These datasets were released by DCRNN[1]. Data can be found in its GitHub repository, where the sensor graphs are also provided.

  • PEMS04 and PEMS08: These datasets were released by ASTGCN[2] and ASTGNN[3]. Data can also be found in its GitHub repository.

3.2 Data Process

python datasets/raw_data/$DATASET_NAME/generate_training_data.py

Replace $DATASET_NAME with one of METR-LA, PEMS-BAY, PEMS04, PEMS08.

The processed data is placed in datasets/$DATASET_NAME.

4. Training the D2STGNN Model

python main.py --dataset=$DATASET_NAME

E.g., python main.py --dataset=METR-LA.

5 Loading a Pretrained D2STGNN Model

Check the config files of the dataset in configs/$DATASET_NAME, and set the startup args to test mode.

Download the pre-trained model files in Google Drive or BaiduYun into the output folder and run the command line in 4.

6 Results and Visualization

TheTable

Visualization

7. More Related Works

8. Citing

If you find this repository useful for your work, please consider citing it as follows:

@article{DBLP:journals/pvldb/ShaoZWWXCJ22,
  author    = {Zezhi Shao and
               Zhao Zhang and
               Wei Wei and
               Fei Wang and
               Yongjun Xu and
               Xin Cao and
               Christian S. Jensen},
  title     = {Decoupled Dynamic Spatial-Temporal Graph Neural Network for Traffic
               Forecasting},
  journal   = {Proc. {VLDB} Endow.},
  volume    = {15},
  number    = {11},
  pages     = {2733--2746},
  year      = {2022}
}

References

[1] Atwood J, Towsley D. Diffusion-convolutional neural networks[J]. Advances in neural information processing systems, 2016, 29: 1993-2001.

[2] Guo S, Lin Y, Feng N, et al. Attention based spatial-temporal graph convolutional networks for traffic flow forecasting[C]//Proceedings of the AAAI Conference on Artificial Intelligence. 2019, 33(01): 922-929.

[3] Guo S, Lin Y, Wan H, et al. Learning dynamics and heterogeneity of spatial-temporal graph data for traffic forecasting[J]. IEEE Transactions on Knowledge and Data Engineering, 2021.