What is Ms.Fpop ?
The MsFPOP R package implements an exact segmentation algorithm that extends functionnal pruning ideas of Rigaill (2015) and Maidstone et al. (2017) to a least squares criterion with a multiscale penalty described in (Verzelen et al. 2020).
The motivation for this method as well as technical details and simulations are described in Liehrmann and Rigaill (2023). Briefly, this penalty promotes well spread changepoints. We propose an efficient C++ implementation interfaced with R of Ms.Fpop allowing to segment a profile of up to 106 datapoints in a matter of seconds.
How can I get Ms.FPOP ?
Make sure that remotes is installed by running install.packages("remotes"), then type
remotes::install_github("aLiehrmann/MsFPOP")Where can I learn more?
See the Quick Start Guide for a minimal exampe of how to use the main function. Details regarding the multiscale penalty can be found in Verzelen et al. (2020). Details regarding functional pruning ideas can be found in Rigaill (2015) and Maidstone et al. (2017).
References
Verzelen, N., Fromont, M., Lerasle, M., Reynaud-Bouret, P. Optimal Change-Point Detection and Localization (2020) https://arxiv.org/abs/2010.11470.
Maidstone R, Hocking T, Rigaill G, Fearnhead P. On optimal multiple changepoint algorithms for large data. Stat Comput 27(2), 519-533 (2017) doi:10.1007/s11222-016-9636-3.
Rigaill G. A pruned dynamic programming algorithm to recover the best segmentations with 1 to Kmax change-points. Journal de la Société Française de Statistique 156(4), (2015) http://www.numdam.org/item/JSFS_2015__156_4_180_0/.
Liehrmann, A. and Rigaill, G. Ms.Fpop: An Exact and Fast Segmentation Algorithm With a Multiscale Penalty (2023) doi:10.48550/arXiv.2303.08723.