sth4k / graph_convolutional_lstm

High-Order Graph Convolutional Recurrent Neural Network

High-Order Graph Convolutional Recurrent Neural Network

A Deep Learning Framework for Network-Scale Traffic Learning and Forecasting

Environment

• Python 3.6.1
• PyTorch 0.3.0

Model Structure

Results

The results can be found in the WiKi

Data

To run the code, you need to download the loop detector data from my GitHub link: https://github.com/zhiyongc/Seattle-Loop-Data and put them in the "Data" folder. For confidentiality, the INRIX data can not be shared.

Cite

Please cite our paper if you use this code or data in your own work: High-Order Graph Convolutional Recurrent Neural Network: A Deep Learning Framework for Network-Scale Traffic Learning and Forecasting

@misc{1802.07007,
  Author = {Zhiyong Cui and Kristian Henrickson and Ruimin Ke and Yinhai Wang},
  Title = {High-Order Graph Convolutional Recurrent Neural Network: A Deep Learning Framework for Network-Scale Traffic Learning and Forecasting},
  Year = {2018},
  Eprint = {arXiv:1802.07007},
}