InspectChangepoint (version 1.2)

High-Dimensional Changepoint Estimation via Sparse Projection

Description

Provides a data-driven projection-based method for estimating changepoints in high-dimensional time series. Multiple changepoints are estimated using a (wild) binary segmentation scheme.

Copy Link

Version

Install

install.packages('InspectChangepoint')

Monthly Downloads

261

Version

1.2

License

GPL-3

Maintainer

Tengyao Wang

Last Published

May 3rd, 2022

Functions in InspectChangepoint (1.2)

random.UnitVector

Generate a random unit vectors in R^n

plot.inspect

Plot function for 'inspect' class objects

vector.norm

Norm of a vector

plot.hdchangeseq

Plot function for 'hdchangeseq' class

vector.clip

Clipping a vector from above and below

cusum.univariate.missing

MissCUSUM transformation of a single vector with missing entries

printPercentage

Print percentage

print.inspect

Print function for 'inspect' class objects

rescale.variance

Noise standardisation for multivariate time series.

summary.inspect

Summary function for 'inspect' class objects

sparse.svd.missing

Computing the sparse leading left singular vector of a matrix with missing entries

vector.normalise

Normalise a vector

single.change

Generating high-dimensional time series with exactly one change in the mean structure

sparse.svd

Computing the sparse leading left singular vector of a matrix

vector.soft.thresh

Soft thresholding a vector

locate.change.missing

Single changepoint estimation with missing data

cusum.transform.missing

MissCUSUM transformation of a matrix with missing entries

PiS

Matrix projection onto the nuclear norm unit sphere

cusum.transform

CUSUM transformation

compute.threshold

Computing threshold used in inspect

inspect

Informative sparse projection for estimation of changepoints (inspect)

multi.change

Generating a high-dimensional time series with multiple changepoints

PiW

Projection onto the standard simplex

locate.change

Single changepoint estimation