Research map / Time series / Long-term forecasting
ProCAST: A Projection Framework for Coupled Aggregation Constrained Multivariate Time Series Forecasting
Jiaqi Xue, Hongji Dong, Yucen Gao, Xiaofeng Gao, Guihai Chen
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 oblique projection | PatchTST · E-Commerce (14 top-level series) | RMSE: 1924 -> 1881 | Improved |
| add orthogonal projection | PatchTST · E-Commerce (14 top-level series) | RMSE: 1924 -> 1924 | No clear effect |
| add oblique projection | PatchTST · RH (7 top-level series) | RMSE: 864.3 -> 817.9 | Improved |
| add oblique projection | Informer · E-Commerce (14 top-level series) | RMSE: 3702 -> 2347 | Improved |
| add oblique projection | DLinear · RH (7 top-level series) | RMSE: 9510 -> 7539 | Improved |
| add oblique projection | PatchTST · E-Commerce (14 top-level series) | MAE: 1041 -> 1057 | Hurt |
| add orthogonal projection | PatchTST · RH (7 top-level series) | MAE: 589.7 -> 589.6 | No clear effect |
| add average ProCAST projection over base models | all base models · E-Commerce (14 top-level series) | relative RMSE reduction: 5.15% | Improved |
| add average ProCAST projection over base models | all base models · RH (7 top-level series) | relative RMSE reduction: 7.10% | Improved |
| add Bottom-Up reconciliation | PatchTST · E-Commerce (14 top-level series) | RMSE: 1924 -> 2030 | Hurt |
| add MinT reconciliation | PatchTST · E-Commerce (14 top-level series) | RMSE: 1924 -> 1945 | Hurt |
| increase history window h from 40 to 80 (ablation) | oblique projection · E-Commerce (14 top-level series) | RMSE | Improved |
Closest papers by meaning
- Multivariate Time Series Forecasting needs Cross Variable Loss (NeurIPS 2026 · similarity 0.86)
- Structured Matrix Basis for Multivariate Time Series Forecasting with Interpretable Dynamics (NeurIPS 2024 · similarity 0.85)
- Learning Optimal Projection for Forecast Reconciliation of Hierarchical Time Series (ICML 2024 · similarity 0.85)
- TimeBase: The Power of Minimalism in Efficient Long-term Time Series Forecasting (ICML 2025 · similarity 0.85)
- TiRex-2: Generalizing TiRex to Multivariate Data and Streaming (NeurIPS 2026 · similarity 0.85)
- Revitalizing Multivariate Time Series Forecasting: Learnable Decomposition with Inter-Series Dependencies and Intra-Series Variations Modeling (ICML 2024 · similarity 0.84)
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.
