HetGAN

Implementation of Heterogeneous Graph Attention Network

This is the code repository of paper “Prediction of Metro Ridership During the COVID-19 Pandemic: A Deep Learning Approach Based on Heterogeneous Graph” which is submitted to Transportation Rsearch Part C: Emgerging Technologies.

Datasets

  • New York City COVID case number
  • New York City Metro Ridership Number

Model Framework

How to Run?

  • Single run:

cd Experiments/
python main.py --epoch 200 --seed 1230 --log_file='./logs/running_log.log'
  • Grid Search:

    cd Experiments/
    python grid_search.py --epoch 200 --seed 1230 --log_file='./logs/grid_search.log'

GitHub

View Github