TimePred: efficient and interpretable offline change point detection for high volume data - with application to industrial process monitoring
Simon Leszek Technische Universität Berlin - Machine Learning Group Straße des 17. Juni 135, 10623 Berlin - Germany · 2025
Read TimePred: efficient and interpretable offline change point detection for high volume data - with application to industrial process monitoring by Simon Leszek Technische Universität Berlin - Machine Learning Group Straße des 17. Juni 135, 10623 Berlin - Germany (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-2512-01562.txt.
Description
Simon Leszek presents TimePred, a fast way to find shift points in big data streams. Streams with many channels are hard to scan. Tests must be sound, fast, and clear to read. TimePred learns to guess each sample's time index in normalized form.
That trick turns a multi-channel hunt into a one-channel mean-shift hunt. Old proven tools can then finish the job offline. The method scales to high-volume data. It serves in fields like factory watch and finance, where a shift in regime must be caught and explained.
Catalog record
| CREATOR | Simon Leszek Technische Universität Berlin - Machine Learning Group Straße des 17. Juni 135, 10623 Berlin - Germany |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 11 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2512-01562