XDZhelheim / HimNet

[KDD'24] Official code for our paper "Heterogeneity-Informed Meta-Parameter Learning for Spatiotemporal Time Series Forecasting".

[KDD'24] Heterogeneity-Informed Meta-Parameter Learning for Spatiotemporal Time Series Forecasting

Our work is already accepted by KDD2024 research track. The citation information will be updated when the official proceeding is online.

Citation

@article{dong2024heterogeneity,
  title={Heterogeneity-Informed Meta-Parameter Learning for Spatiotemporal Time Series Forecasting},
  author={Dong, Zheng and Jiang, Renhe and Gao, Haotian and Liu, Hangchen and Deng, Jinliang and Wen, Qingsong and Song, Xuan},
  journal={arXiv preprint arXiv:2405.10800},
  year={2024}
}

Performance on Spatiotemporal Forecasting Benchmarks

Required Packages

pytorch>=1.12
numpy
pandas
matplotlib
pyyaml
pickle
torchinfo

Training Commands

cd scripts/
python train.py -d <dataset> -g <gpu_id>

<dataset>:

• METRLA
• PEMSBAY
• PEMS04
• PEMS07
• PEMS08