Research map / Time series / Long-term forecasting
FEDformer: Frequency Enhanced Decomposed Transformer for Long-term Series Forecasting
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use FEDformer instead of Autoformer | Multivariate forecasting · Six benchmark datasets, all reported horizons | MSE: 14.8% | Helped |
| Use FEDformer instead of Autoformer | Univariate forecasting · Six benchmark datasets, all reported horizons | MSE: 22.6% | Helped |
| Use FEDformer-f for self-attention only, retaining AutoCorr cross-attention (ablation) | FEDformer · ETTm1 and ETTm2 multivariate, 16 cases | MSE: 10/16 cases | Mixed |
| Use FEDformer-f for cross-attention only, retaining AutoCorr self-attention (ablation) | FEDformer · ETTm1 and ETTm2 multivariate, 16 cases | MSE: 12/16 cases | Mixed |
| Use FEA-f for both self-attention and cross-attention (ablation) | FEDformer · ETTm1 and ETTm2 multivariate, 16 cases | MSE: 7/16 cases | Mixed |
| Use both FEB and FEA blocks instead of the Autoformer attention setup (ablation) | FEDformer-f · ETTm1 and ETTm2 multivariate, 16 cases | MSE: 16/16 cases | Helped |
| Use mixture-of-experts decomposition instead of a single decomposition scheme (ablation) | FEDformer · ETT and Weather datasets | improvement: 2.96% | Helped |
| Select Fourier modes randomly rather than keeping only the lowest-frequency modes (ablation) | FEDformer-f · ETT full benchmark | MSE | Helped |
| Use an appropriate random subset of modes rather than all modes (ablation) | FEDformer-f · ETT full benchmark | MSE | Helped |
| Use Fourier rather than wavelet basis (ablation) | FEDformer · Different datasets and horizons | MSE | Mixed |
Closest papers by meaning
- SAMformer: Unlocking the Potential of Transformers in Time Series Forecasting with Sharpness-Aware Minimization and Channel-Wise Attention (ICML 2024 · similarity 0.87)
- FilterNet: Harnessing Frequency Filters for Time Series Forecasting (NeurIPS 2024 · similarity 0.86)
- DeformableTST: Transformer for Time Series Forecasting without Over-reliance on Patching (NeurIPS 2024 · similarity 0.86)
- Pathformer: Multi-scale Transformers with Adaptive Pathways for Time Series Forecasting (ICLR 2024 · similarity 0.86)
- Rethinking Fourier Transform from A Basis Functions Perspective for Long-term Time Series Forecasting (NeurIPS 2024 · similarity 0.86)
- Parsimony or Capability? Decomposition Delivers Both in Long-term Time Series Forecasting (NeurIPS 2024 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.