Research map / Time series / Causal and generative time series
Locally Linear Continual Learning for Time Series Based on VC-Theoretical Generalization Bounds
Yan V. G. Ferreira, Igor Barbosa Lima, Pedro H. G. Mapa S., Felipe V. Campos, A. P. Braga
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 model aggregation (exponential-distance weighted interpolation) to SyMPLER (ablation) | SyMPLER · GEFCom2014 electric load | prediction error (RMSE): 21.06 -> 21.01 | Improved |
| add model aggregation to SyMPLER (ablation) | SyMPLER · GEFCom2014 electric load | forgetting ratio: 0.11 -> 0.14 | Hurt |
| add model aggregation to SyMPLER (ablation) | SyMPLER · GEFCom2014 electric load | fitting error (RMSE): 21.91 -> 21.09 | Improved |
| use SyMPLER's performance-based model addition instead of eTS clustering | eTS · GEFCom2014 electric load | prediction error (RMSE): 28.87 -> 21.06 | Improved |
| use SyMPLER instead of ICN | ICN · GEFCom2014 electric load | prediction error (RMSE): 39.07 -> 21.06 | Improved |
| use SyMPLER instead of EWC-SWFT | EWC-SWFT · GEFCom2014 electric load | prediction error (RMSE): 55.64 -> 21.06 | Improved |
| use SyMPLER instead of offline XGBoost (forgetting) | XGBoost · GEFCom2014 electric load | forgetting ratio: 0.66 -> 0.11 | Improved |
| use offline XGBoost instead of SyMPLER (accuracy) | SyMPLER · GEFCom2014 electric load | prediction error (RMSE): 21.06 -> 18.56 | Improved |
| switch model selection to recent-prediction-error criterion under concept drift | SyMPLER · Pendulum with rod length change | prediction error | Improved |
| increase spurious input variables from 0 to 100 (ablation) | SyMPLER · Pendulum with spurious variables | test MSE: 0 to 100 added variables | Hurt |
| increase ridge regularization under high output noise (ablation) | SyMPLER · Pendulum noisy outputs (std 0 to 2) | coefficient distance to Taylor | Improved |
| increase ridge regularization at low noise (ablation) | SyMPLER · Pendulum noisy outputs (std 0 to 2) | coefficient distance to Taylor | Hurt |
Closest papers by meaning
- An Analysis of Linear Time Series Forecasting Models (ICML 2024 · similarity 0.86)
- PAMod: Modeling Cyclical Shifts via Phase-Amplitude Modulation for Non-stationary Time Series Forecasting (NeurIPS 2026 · similarity 0.86)
- TimePerceiver: An Encoder-Decoder Framework for Generalized Time-Series Forecasting (NeurIPS 2025 · similarity 0.86)
- Improving Time Series Forecasting via Instance-aware Post-hoc Revision (NeurIPS 2025 · similarity 0.86)
- Online Time Series Forecasting with Theoretical Guarantees (NeurIPS 2025 · similarity 0.85)
- Towards a General Time Series Forecasting Model with Unified Representation and Adaptive Transfer (ICML 2025 · 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.
