Research map / Time series / Long-term forecasting
NHITS: Neural Hierarchical Interpolation for Time 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 |
|---|---|---|---|
| add hierarchical interpolation and multi-rate sampling (ablation) | N-HiTS · Five datasets, H=720 | MSE: 0.548 -> 0.484 | Helped |
| add multi-rate sampling (ablation) | N-HiTS · Five datasets, H=720 | MSE: 0.518 -> 0.484 | Helped |
| add hierarchical interpolation (ablation) | N-HiTS · Five datasets, H=720 | MSE: 0.498 -> 0.484 | Helped |
| replace max pooling with average pooling (ablation) | N-HiTS · Long-horizon forecasting | forecast accuracy | Hurt |
| replace linear interpolation with nearest-neighbor interpolation (ablation) | N-HiTS · Long-horizon forecasting | forecast accuracy | Hurt |
| replace linear interpolation with cubic interpolation (ablation) | N-HiTS · Long-horizon forecasting | forecast accuracy | Hurt |
| synthesize and remove low-frequency components before modeling fine-grained components (ablation) | N-HiTS hierarchical interpolation · Long-horizon forecasting | forecast accuracy | Helped |
| use N-HiTS instead of the strongest baseline | Long-horizon forecasting · ECL, H=720 | MSE: 0.231 -> 0.243 | Hurt |
| use N-HiTS instead of the strongest baseline | Long-horizon forecasting · ETTm2, H=720 | MSE: 0.411 -> 0.401 | Helped |
Closest papers by meaning
- HDT: Hierarchical Discrete Transformer for Multivariate Time Series Forecasting (AAAI 2025 · similarity 0.88)
- FilterNet: Harnessing Frequency Filters for Time Series Forecasting (NeurIPS 2024 · similarity 0.87)
- Hierarchical Classification Auxiliary Network for Time Series Forecasting (AAAI 2025 · similarity 0.86)
- Learning Optimal Projection for Forecast Reconciliation of Hierarchical Time Series (ICML 2024 · similarity 0.86)
- Graph-based Time Series Clustering for End-to-End Hierarchical Forecasting (ICML 2024 · similarity 0.86)
- Towards a General Time Series Forecasting Model with Unified Representation and Adaptive Transfer (ICML 2025 · similarity 0.86)
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.