The code repository for SOFTS: Efficient Multivariate Time Series Forecasting with Series-Core Fusion in PyTorch. A scalable pure MLP model that achieves state-of-the-art performance on multivariate time series forecasting benchmarks.
scikit-learn==1.2.2
numpy==1.22.4
pandas==1.2.4
torch==1.10.0+cu111
We refer to this repository for downloading datasets.
To reproduce the main results in Table 2, run the script files under folder scripts/long_term_forecast
.
For example, to reproduce the results of SOFTS on ETTm1 dataset, run the following command:
sh scripts/long_term_forecast/ETT_script/SOFTS_ETTm1.sh
We appreciate the following github repos a lot for their valuable code base or datasets:
https://github.com/zhouhaoyi/Informer2020
https://github.com/thuml/Autoformer
https://github.com/zhouhaoyi/ETDataset
https://github.com/laiguokun/multivariate-time-series-data
https://github.com/thuml/Time-Series-Library
https://github.com/thuml/iTransformer
If you find our work useful in your research, please use the following citation:
@article{han2024softs,
title={SOFTS: Efficient Multivariate Time Series Forecasting with Series-Core Fusion},
author={Han, Lu and Chen, Xu-Yang and Ye, Han-Jia and Zhan, De-Chuan},
journal={arXiv preprint arXiv:2404.14197},
year={2024}
}