A Hybrid Framework Combining Autoregression and Common Factors for Matrix Time Series
Zhiyun Fan †, Xiaoyu Zhang ‡, Di Wang † · 2025
Read A Hybrid Framework Combining Autoregression and Common Factors for Matrix Time Series by Zhiyun Fan †, Xiaoyu Zhang ‡, Di Wang † (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2503-05340.txt.
Introduction
Zhiyun Fan and co-authors merge two modeling styles for data that arrives as a stream of tables, in this 2025 paper. Matrix time series fill economics and finance, where rows and columns both matter. Autoregressive models let input and output spaces differ; factor models force them to match. Their hybrid names the overlap explicitly, splitting each coefficient into shared, input-only, and output-only parts.
That split trims the parameter count while keeping flexibility. A regularized gradient method fits the model at scale, and the theory shows it settles quickly near the answer with sound statistics. Simulations and a global economic forecasting test show the gains. The authors present the framework as a practical bridge between two literatures that rarely meet.
Catalog record
| CREATOR | Zhiyun Fan †, Xiaoyu Zhang ‡, Di Wang † |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 257 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2503-05340