The code for paper "Traffic Prediction with Transfer Learning: A Mutual Information-based Approach"
- This version is a reconstruct code for TrafficTL, so it omits some complicated operations in papers like replace nodes in source city with nodes in target city.
include but not limited
torch
tqdm
dtaidistance
logging
pyyaml
numpy
sklearn
pathlibFor data privacy, please apply follow line to generate sample data for going through whole process.
python data_generate.pyif you want to go through the pipeline, please use next command.
python main.py --src_city 'src' --trg_city 'trg' --device 0if you want to use it on your own data, please place your data on the file 'data' and specify data path in 'config.yaml'.
python main.py --src_city 'xxx' --trg_city 'xxx' --device 0please use city name replace 'xxx'.
Thanks for repository codes IIC and DCRNN
@ARTICLE{10105852,
author={Huang, Yunjie and Song, Xiaozhuang and Zhu, Yuanshao and Zhang, Shiyao and Yu, James J. Q.},
journal={IEEE Transactions on Intelligent Transportation Systems},
title={Traffic Prediction With Transfer Learning: A Mutual Information-Based Approach},
year={2023},
volume={24},
number={8},
pages={8236-8252},
doi={10.1109/TITS.2023.3266398}}
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent