Research map / Time series / Long-term forecasting
PESD-TSF: A Period-Aware and Explicit Structured Decomposition Framework for Long-Term Time Series Forecasting
Hua Wang, Xianhao Jiao, Fan Zhang
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 period-aware gating (ablation) | PESD-TSF · ETTh1 (L=720) | MSE: 0.430 -> 0.394 | Improved |
| add period-aware gating (ablation) | PESD-TSF · ETTh2 (L=720) | MSE: 0.423 -> 0.339 | Improved |
| add CSCA (ablation) | PESD-TSF · ETTh1 (L=720) | MSE: 0.415 -> 0.394 | Improved |
| add RLC regularization (ablation) | PESD-TSF · ETTh1 (L=720) | MSE: 0.405 -> 0.394 | Improved |
| add hierarchical multi-scale sampling (ablation) | PESD-TSF · ETTm2 (L=720) | MSE: 0.284 -> 0.244 | Improved |
| add RLC regularization (ablation) | PESD-TSF · ETTm1 (L=720) | MAE: MSE 0.401 -> 0.343; MAE 0.380 vs 0.370 | Mixed |
| keep stage order S1->S2->S3 instead of reordering (ablation) | PESD-TSF · ETTh1 (L=720, O=720) | MSE/MAE: 0.448/0.471 -> 0.437/0.463 | Improved |
| include Stage 2 PatchSampling (ablation) | PESD-TSF · Traffic (L=720, O=720) | MSE/MAE: 0.394/0.269 -> 0.376/0.257 | Improved |
| include CSCA in stage order test (ablation) | PESD-TSF · Traffic (L=720, O=720) | MSE/MAE: 0.388/0.265 -> 0.376/0.257 | Improved |
| RLC weight lambda too large (>1e-2) rather than 1e-3 (ablation) | PESD-TSF · ETTh1/ETTh2/Traffic/Electricity | MSE | Hurt |
| periodic coefficient gamma near 0.5 vs extremes (ablation) | PESD-TSF · Traffic, ETTh1 | MSE | Improved |
| replace channel-independent attention with CSCA cross-variable modeling vs best Transformer | PESD-TSF · Traffic (L=720) | MSE/MAE: 0.380/0.267 -> 0.354/0.242 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 4 of 4 on most shared results.
- TimeBase — TimeBase: The Power of Minimalism in Efficient Long-term Time Series Forecasting (ICML 2025) · better on 6 of 6 shared results
- TimeBridge — TimeBridge: Non-Stationarity Matters for Long-term Time Series Forecasting (ICML 2025) · better on 3 of 4 shared results
- TimeMixer — TimeMixer: Decomposable Multiscale Mixing for Time Series Forecasting (ICLR 2024) · better on 3 of 4 shared results
- iTransformer — iTransformer: Inverted Transformers Are Effective for Time Series Forecasting (ICLR 2024) · better on 1 of 1 shared result
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
- TStruct: Learning Shared Temporal Structures for Long-Term Time-Series Forecasting (NeurIPS 2026 · similarity 0.89)
- TimePerceiver: An Encoder-Decoder Framework for Generalized Time-Series Forecasting (NeurIPS 2025 · similarity 0.89)
- TimeCMA: Towards LLM-Empowered Multivariate Time Series Forecasting via Cross-Modality Alignment (AAAI 2025 · similarity 0.87)
- Enhancing Foundation Models for Time Series Forecasting via Wavelet-based Tokenization (ICML 2025 · similarity 0.87)
- Time-MoE: Billion-Scale Time Series Foundation Models with Mixture of Experts (ICLR 2025 · similarity 0.87)
- TimeBase: The Power of Minimalism in Efficient Long-term Time Series Forecasting (ICML 2025 · 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.
