Research map / Time series / Causal and generative time series
Characteristic Root Analysis and Regularization for Linear Time Series Forecasting
Zheng Wang, Kaixuan Zhang, Wanfang Chen, Xiaonan Lu, Longyuan Li, Tobias Schlagenhauf
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 Root Purge regularizer to linear model (INC) (ablation) | INC linear model · ETTh1 H=96 | MSE: 0.397 -> 0.357 | Improved |
| add Root Purge regularizer to linear model (CI) (ablation) | CI linear model · ETTh1 H=96 | MSE: 0.375 -> 0.359 | Improved |
| add Root Purge regularizer to linear model (INC) (ablation) | INC linear model · ETTm1 H=96 | MSE: 0.297 -> 0.292 | Improved |
| add Root Purge to CI linear model (ablation) | CI linear model · ETTm2 H=96 | MSE: 0.162 -> 0.162 | No clear effect |
| individual channel modeling instead of channel independent (ablation) | linear model · ETTh1 H=96 | MSE: 0.375 -> 0.397 | Hurt |
| individual channel modeling instead of channel independent (ablation) | linear model · ETTm1 H=96 | MSE: 0.306 -> 0.297 | Improved |
| add rank reduction (RRR) vs OLS | linear model · synthetic roots | root distance: 0.064 -> 0.036 | Improved |
| add Root Purge vs OLS | linear model · synthetic roots | root distance: 0.064 -> 0.045 | Improved |
| Root Purge vs RRR in frequency domain | linear model · Weather H=96 | MSE: 0.140 -> 0.142 | Hurt |
| Root Purge vs FITS | FITS · Exchange H=720 | MSE: 0.936 -> 0.941 | Hurt |
| increase lambda from small to moderate then large (ablation) | Root Purge · ETTh1 and ETTm1 | MSE | Mixed |
| increase lambda (ablation) | Root Purge · singular values | singular value spectrum | 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.
- SparseTSF — SparseTSF: Modeling Long-term Time Series Forecasting with *1k* Parameters (ICML 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
- An Analysis of Linear Time Series Forecasting Models (ICML 2024 · similarity 0.88)
- Parsimony or Capability? Decomposition Delivers Both in Long-term Time Series Forecasting (NeurIPS 2024 · similarity 0.87)
- TimeBase: The Power of Minimalism in Efficient Long-term Time Series Forecasting (ICML 2025 · similarity 0.86)
- SIN: Selective and Interpretable Normalization for Long-Term Time Series Forecasting (ICML 2024 · similarity 0.86)
- Enhancing Foundation Models for Time Series Forecasting via Wavelet-based Tokenization (ICML 2025 · similarity 0.86)
- Rethinking Channel Dependence for Multivariate Time Series Forecasting: Learning from Leading Indicators (ICLR 2024 · 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.
