gfpop: Graph-Constrained Functional Pruning Optimal Partitioning

Penalized parametric change-point detection by functional pruning dynamic programming algorithm. The successive means are constrained using a graph structure with edges of types null, up, down, std or abs. To each edge we can associate some additional properties: a minimal gap size, a penalty, some robust parameters (K,a). The user can also constrain the inferred means to lie between some minimal and maximal values. Data is modeled by a quadratic cost with possible use of a robust loss, biweight and Huber (see edge parameters K and a). Other losses are also available with log-linear representation or a log-log representation.

Version: 1.0.2
Depends: R (≥ 3.5.0)
Imports: Rcpp (≥ 1.0.0)
LinkingTo: Rcpp
Published: 2020-12-01
Author: Vincent Runge [aut, cre], Toby Hocking [aut], Guillem Rigaill [aut], Gaetano Romano [aut], Fatemeh Afghah [aut], Paul Fearnhead [aut], Michel Koskas [ctb], Arnaud Liehrmann [ctb]
Maintainer: Vincent Runge <vincent.runge at>
License: MIT + file LICENSE
NeedsCompilation: yes
SystemRequirements: C++11
CRAN checks: gfpop results


Reference manual: gfpop.pdf
Package source: gfpop_1.0.2.tar.gz
Windows binaries: r-devel:, r-release:, r-oldrel:
macOS binaries: r-release: gfpop_1.0.2.tgz, r-oldrel: gfpop_1.0.2.tgz
Old sources: gfpop archive


Please use the canonical form to link to this page.