ATFNet

Official implement of ATFNet: Adaptive Time-Frequency Ensembled Network for Long-term Time Series Forecasting

Model Architecture

Main Results

• Multi-variate long-term time series forecasting results on 8 datasets. The best results are in bold and the second best results are underlined.

• Forecasting Results Visualization

(Visualization of the input-192-predict-192 forecasting results on ETTm2 dataset.)

Acknowledgement

We acknowledge with great appreciation the significant contributions of the following GitHub repositories, which have provided invaluable code bases and datasets for our research.

• https://github.com/thuml/Time-Series-Library
• https://github.com/cure-lab/SCINet.
• https://github.com/aikunyi/FreTS.
• https://github.com/VEWOXIC/FITS
• https://github.com/microsoft/StemGNN/
• https://github.com/thuml/Autoformer.
• https://github.com/MAZiqing/FEDformer