Research map / Time series / Long-term forecasting
XLinear: A Lightweight and Accurate MLP-Based Model for Long-Term Time Series Forecasting with Exogenous Inputs
Xinyang Chen, Huidong Jin, Yu Huang, Zaiwen Feng
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 global tokens carrying exogenous information to the endogenous sequence (ablation) | XLinear · Weather multivariate h96 | MSE: 0.175 -> 0.149 | Improved |
| add global tokens carrying exogenous information (ablation) | XLinear · ETTh1 multivariate h96 | MSE: 0.370 -> 0.369 | No clear effect |
| add endogenous temporal sequence to global tokens (ablation) | XLinear · Weather multivariate h96 | MSE: 0.153 -> 0.149 | Improved |
| add endogenous temporal sequence to global tokens (ablation) | XLinear · GTDS h96 | MSE: 0.012 -> 0.012 | No clear effect |
| use global tokens alone instead of endogenous sequence alone (ablation) | XLinear · ETTm1 / Weather / GTDS | MSE: ETTm1 0.328 -> 0.326; Weather 0.175 -> 0.153 | Improved |
| use XLinear instead of TimeXer | XLinear · Electricity multivariate h96 | MSE: 0.140 -> 0.138 | Improved |
| use XLinear instead of iTransformer on high-dimensional Traffic | XLinear · Traffic multivariate h96 | MSE: 0.395 -> 0.432 | Hurt |
| use XLinear instead of patch-based TimeXer and PatchTST on volatile PEMS data | XLinear · PEMS03 h12 | MSE: 0.033 -> 0.030 | Improved |
| use XLinear on PEMS03 at longest horizon | XLinear · PEMS03 h96 | MSE: 0.129 -> 0.133 | Hurt |
| use lookback 512 with XLinear instead of PatchTST at horizon 720 | XLinear · ETTh1 lookback 512 h720 | MSE: 0.447 -> 0.451 | No clear effect |
| use XLinear instead of TimeXer for training speed | XLinear · Electricity univariate | training time ms/iter: 37 -> 17 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 3 on most shared results.
- TimeXer — TimeXer: Empowering Transformers for Time Series Forecasting with Exogenous Variables (NeurIPS 2024) · better on 8 of 10 shared results
- iTransformer — iTransformer: Inverted Transformers Are Effective for Time Series Forecasting (ICLR 2024) · better on 3 of 5 shared results
- TimeMixer — TimeMixer: Decomposable Multiscale Mixing for Time Series Forecasting (ICLR 2024) · better on 4 of 4 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
- OLinear: A Linear Model for Time Series Forecasting in Orthogonally Transformed Domain (NeurIPS 2025 · similarity 0.87)
- TimeXer: Empowering Transformers for Time Series Forecasting with Exogenous Variables (NeurIPS 2024 · similarity 0.87)
- Unlocking the Power of Patch: Patch-Based MLP for Long-Term Time Series Forecasting (AAAI 2025 · similarity 0.86)
- ARM: Refining Multivariate Forecasting with Adaptive Temporal-Contextual Learning (ICLR 2024 · similarity 0.86)
- xLSTM-Mixer: Multivariate Time Series Forecasting by Mixing via Scalar Memories (NeurIPS 2025 · similarity 0.86)
- TimeBase: The Power of Minimalism in Efficient Long-term Time Series Forecasting (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.
