TPGCN

The official implementation of Time-aware Personalized Graph Convolutional Network for Multivariate Time Series Forecasting.

TPGCN- Official PyTorch Implementation

The official implementation of Time-aware Personalized Graph Convolutional Network for Multivariate Time Series Forecasting. (paper).

Requirements

• python 3
• see requirements.txt

Data Preparation

Raw Dataset

Traffic data https://github.com/LeiBAI/AGCRN

Time series data https://github.com/laiguokun/multivariate-time-series-data

Train Commands

For traffic datasets (PeMSD3, PeMSD4, PeMSD7, PeMSD8):

python Pems4/train_pems.py --gcn_bool --dataset

For time series datasets:

python Time_series/train_series.py --gcn_bool --dataset

Experimental result

I uploaded the model's training log on the PeMSD8 dataset

Citing

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

@article{li2024time,
  title={Time-aware personalized graph convolutional network for multivariate time series forecasting},
  author={Li, ZhuoLin and Gao, ZiHeng and Zhang, XiaoLin and Zhang, GaoWei and Xu, LingYu},
  journal={Expert Systems with Applications},
  volume={240},
  pages={122471},
  year={2024},
  url={https://doi.org/10.1016/j.eswa.2023.122471},
  doi={10.1016/j.eswa.2023.122471},
  timestamp={Mon, 15 April 2024},
  publisher={Elsevier}
}