Rethinking Fourier Transform from A Basis Functions Perspective for Long-term Time Series Forecasting. (NeurIPS 2024)

This is the offical implementation of FBM-L, FBM-NL and FBM-NP model.

Implement the project

1. Install requirements. pip install -r requirements.txt

2. Download data. You can download the ETTh1, ETTh2, ETTm1, ETTm2, Electricity and Traffic data from Autoformer and WTH data from Google Drive Create a seperate folder ./dataset and put all the csv files in the directory.

3. Training. All the scripts are in the directory ./scripts/FBM/file_to_implement.sh

sh ./scripts/FBM/ETTh1.sh

You can adjust the hyperparameters based on your needs.

Fourier Basis Mapping

Main Results

Acknowledgement

We appreciate the following github repo very much for the valuable code base and datasets:

https://github.com/cure-lab/LTSF-Linear

https://github.com/zhouhaoyi/Informer2020

https://github.com/thuml/Autoformer

https://github.com/MAZiqing/FEDformer

https://github.com/alipay/Pyraformer

https://github.com/yuqinie98/PatchTST

https://github.com/ServiceNow/N-BEATS

https://github.com/aikunyi/FreTS

https://github.com/hqh0728/CrossGNN

https://github.com/thuml/iTransformer

https://github.com/kwuking/TimeMixer

https://github.com/VEWOXIC/FITS

Citation

If you find this repository useful, please consider citing our paper.

If you have any questions, feel free to contact: runze.y@sjtu.edu.cn