/DMVST-Net

DMVST-Net for AAAI 2018

Primary LanguagePythonMIT LicenseMIT

Deep Multi-View Spatial-Temporal Network for Taxi Demand Prediction

Source code of the paper Deep Multi-View Spatial-Temporal Network for Taxi Demand Prediction

If you find the code is userful, please cite our paper

@inproceedings{yao2018deep,
  title={Deep Multi-View Spatial-Temporal Network for Taxi Demand Prediction},
  author={Yao, Huaxiu and Wu, Fei and Ke, Jintao and Tang, Xianfeng and Jia, Yitian and Lu, Siyu and Gong, Pinghua and Ye, Jieping and Li Zhenhui},
  booktitle={The Thirty-Second AAAI Conference on Artificial Intelligence},
  year={2018}
}

Current Version

I only put the model part on it. Other parts will be put soon.

The model part is just for reference

Requirements

  • keras>=2.0.8
  • tensorflow>=1.3.0
  • python 2.7 (Recommend Anaconda)

Data

This work is an intern work in Didi Chuxing. Due to the privacy policy, we do not provide the raw data. Maybe you can apply a similar type of data in https://outreach.didichuxing.com/appEn-vue/Personal.