3 packages on CRAN
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.
Optimal partitioning algorithm for change-in-slope problem with continuity constraint and a finite number of states. Some constraints can be enforced in the inference: isotonic, unimodal or smoothing. With the function slopeSN() (segment neighborhood) the number of segments to infer is fixed by the user and does not depend on a penalty value.
Detect abrupt changes in time series with local fluctuations as a random walk process and autocorrelated noise as an AR(1) process. See Romano, G., Rigaill, G., Runge, V., Fearnhead, P. (2020) <arXiv:2005.01379>.