Research map / Time series / Long-term forecasting
FreqCycle: A Multi-Scale Time-Frequency Analysis Method for Time Series Forecasting
Boya Zhang, Shuaijie Yin, Huiwen Zhu, Xing He
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| add FECF module (ablation) | FreqCycle · Traffic (L=96, H=96) | MSE: 0.501 -> 0.438 | Improved |
| add FECF module (ablation) | FreqCycle · Weather (L=96, H=96) | MSE: 0.179 -> 0.159 | Improved |
| add FECF module (ablation) | FreqCycle · ETTh1 (L=96, H=96) | MSE: 0.375 -> 0.369 | Improved |
| use SFPL instead of MLP (ablation) | FreqCycle · ETTh2 (L=96, H=96) | MSE: 0.302 -> 0.282 | Improved |
| use SFPL instead of MLP (ablation) | FreqCycle · Traffic (L=96, H=96) | MSE: MAE 0.262 -> 0.262 | No clear effect |
| use SFPL instead of FITS low-pass filter (ablation) | FreqCycle · ECL (L=96, H=96) | MSE: 0.143 -> 0.139 | Improved |
| use SFPL instead of PaiFilter (ablation) | FreqCycle · Traffic (L=96, H=96) | MSE: 0.445 -> 0.438 | Improved |
| use FECF instead of moving-average decomposition (ablation) | FreqCycle · ECL (L=96, avg over H) | MSE: 0.204 -> 0.168 | Improved |
| use FECF instead of local decomposition (ablation) | FreqCycle · Traffic (L=96, avg over H) | MSE: 0.527 -> 0.448 | Improved |
| increase lookback window from 96 to 168 (ablation) | CycleNet · ETTh2 | MSE: 0.383 -> 0.388 | Hurt |
| increase lookback window from 96 to 672 (ablation) | CycleNet · ETTm2 | MSE: 0.267 -> 0.262 | Improved |
| use MFreqCycle with longer lookback 168 (ablation) | FreqCycle · ETTh2 | MSE: 0.371 -> 0.241 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- TimeMixer — TimeMixer: Decomposable Multiscale Mixing for Time Series Forecasting (ICLR 2024) · better on 5 of 7 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- FilterNet: Harnessing Frequency Filters for Time Series Forecasting (NeurIPS 2024 · similarity 0.89)
- MSGNet: Learning Multi-Scale Inter-series Correlations for Multivariate Time Series Forecasting (AAAI 2024 · similarity 0.87)
- TimeKAN: KAN-based Frequency Decomposition Learning Architecture for Long-term Time Series Forecasting (ICLR 2025 · similarity 0.87)
- Rethinking Fourier Transform from A Basis Functions Perspective for Long-term Time Series Forecasting (NeurIPS 2024 · similarity 0.87)
- CFPT: Empowering Time Series Forecasting through Cross-Frequency Interaction and Periodic-Aware Timestamp Modeling (ICML 2025 · similarity 0.87)
- CycleNet: Enhancing Time Series Forecasting through Modeling Periodic Patterns (NeurIPS 2024 · similarity 0.87)
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.
