Early Classification of Time Series in Non-Stationary Cost Regimes
Aurélien Renault, Antoine Cornuéjols, Alexis Bondu, Vincent Lemaire · 2026
Read Early Classification of Time Series in Non-Stationary Cost Regimes by Aurélien Renault, Antoine Cornuéjols, Alexis Bondu, Vincent Lemaire (2026) 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-2602-00918.txt.
Description
Aurélien Renault and colleagues adapt early classification to online learning, updating only the trigger model while keeping the classifier fixed. They study early classification of time series under cost non-stationarity. Most methods assume decision costs are known, fixed, and right, but training and deployment objectives then mismatch. They cover two cost non-stationarity forms: drift in the error-versus-delay balance, and random cost cases deviating from the nominal model. They propose online updates and baselines, with bandit-based and RL-based variants.
Catalog record
| CREATOR | Aurélien Renault, Antoine Cornuéjols, Alexis Bondu, Vincent Lemaire |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 40822 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602.00918