Bernstein Polynomial Processes for Continuous Time Change Detection
Dan Cunha, Mark Friedl, Luis Carvalho · 2025
Read Bernstein Polynomial Processes for Continuous Time Change Detection by Dan Cunha, Mark Friedl, Luis Carvalho (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-2504-17876.txt.
Description
Dan Cunha, Mark Friedl, and Luis Carvalho build change detectors for data that arrive at uneven times, in a 2025 paper. Most methods assume tidy, evenly spaced observations; in continuous time, two close readings should rarely straddle a change, while distant ones often do. Their model is a Markov chain whose jump odds stretch with the gap, specified without informative priors.
Forward-backward algorithms then locate the changes exactly, no slow MCMC sampling required. A new loss function and its Bayes estimator suit the setting. On synthetic data and on satellite time series, with a weather-aware mean curve to suppress false alarms, the method performed well across three land-change tasks.
Catalog record
| CREATOR | Dan Cunha, Mark Friedl, Luis Carvalho |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 113 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2504-17876