MsPELT

MsPELT is an exact segmentation algorithm that extends inequality pruning idea from PELT to a least squares criterion with a multiscale penalty promoting well-spread changepoints. Such a penalty has been proposed in Verzelen et al. (2020), and it achieves optimal rates for changepoint detection and changepoint localization (arXiv:2010.11470).

Usage

MsPELT(y, beta, alpha, wt)

Arguments

y

A vector of observations.

beta

A constant used in the penalty.

alpha

A constant used in the penalty. (recommended: gamma+beta*log(length(y))).

wt

A vector of weights associated to the observations.

Value

A list that includes the estimated position of the changepoints.