FreDF: Learning to Forecast in Frequency Domain

Hao Wang, Licheng Pan, Zhichao Chen, Degui Yang, Sen Zhang, Yifei Yang, Xinggao Liu, Haoxuan Li, Dacheng Tao

Abstract

Time series modeling is uniquely challenged by the presence of autocorrelation in both historical and label sequences. Current research predominantly focuses on handling autocorrelation within the historical sequence but often neglects its presence in the label sequence. Specifically, emerging forecast models mainly conform to the direct forecast (DF) paradigm, generating multi-step forecasts under the assumption of conditional independence within the label sequence. This assumption disregards the inherent autocorrelation in the label sequence, thereby limiting the performance of DF-based models. In response to this gap, we introduce the Frequency-enhanced Direct Forecast (FreDF), which bypasses the complexity of label autocorrelation by learning to forecast in the frequency domain. Our experiments demonstrate that FreDF substantially outperforms existing state-of-the-art methods including iTransformer and is compatible with a variety of forecast models.
BibTeX
@inproceedings{wang2025fredf,
  author = {Hao Wang and Licheng Pan and Zhichao Chen and Degui Yang and Sen Zhang and Yifei Yang and Xinggao Liu and Haoxuan Li and Dacheng Tao},
  title = {{FreDF: Learning to Forecast in Frequency Domain}},
  booktitle = {International Conference on Learning Representations},
  year = {2025},
  url = {https://openreview.net/forum?id=4A9IdSa1ul},
}
Figure