PELT.online.update: PELT.online.update (Pruned Exact Linear Time) - Only intended...

In rkillick/changepoint.online: Online detection of changepoints

Description

Implements the PELT method for identifying changepoints in a given set of summary statistics for a specified cost function and penalty. This function is used when we have existing data and want to update it by adding new data.

This function is called by ocpt.mean.update, ocpt.var.update and ocpt.var.update. This is not intended for use by regular users of the package. It is exported for developers to call directly for speed increases or to fit alternative cost functions.

WARNING: No checks on arguments are performed!

Usage

PELT.online.update(previousanswer,newdata)

Arguments

previousanswer	This is a summary of data produced by ocpt.mean.initialise or similar.
newdata	A new set of data in the form of a vector or a column/row of a matrix.

Details

This function is used as a wrapper function to implement the PELT algorithm in C. It simply creates the necessary worker vectors, ensures all inputs are the correct type, and passes everything to the C function.

This function is exported for developer use only. It does not perform any checks on inputs (other than type coersion) and is simply a wrapper function for the C code.

Value

A list is returned with elements:

cost_func	The cost function that has been used.
sumstat	The summary statistics for the time series.
ndone	Length of the time series when analysis begins.
nupdate	Length of the time series to be analysed in this update.
penalty	Penalty used to decide if a changepoint is significant.
cptsout	Vector of locations of the identified changepoints up to (ndone+nupdate).
error	0 by default, nonzero indicates error in code.
shape	Only used when cost_func is the gamma likelihood. Otherwise 1.
minseglen	Minimum segment length.
lastchangelike	Vector of lenght n containing the likelihood of the optimal segmentation up to each timepoint.
lastchangecpts	Vector of length n containing the last changepoint prior to each timepoint.
checklist	Vector of locations of the potential last changepoint for next iteration (to be updated), max length=(ndone+nupdate).
nchecklist	Number in the checklist currently (to be updated).
numchangecpts	Stores the current number of changepoints detected.

Author(s)

Andrew Connell Rebecca Killick

References

PELT Algorithm: Killick R, Fearnhead P, Eckley IA (2012) Optimal detection of changepoints with a linear computational cost, JASA 107(500), 1590–1598

CROPS: Haynes K, Eckley IA, Fearnhead P (2014) Efficient penalty search for multiple changepoint problems (in submission), arXiv:1412.3617

See Also

PELT.online,PELT.online.initialise,ocpt

Examples

#This function should only be used by developers, 
#see its use in ocpt.mean.update, ocpt.var.update 
#and ocpt.meanvar.update.

rkillick/changepoint.online documentation built on Sept. 26, 2020, 11:01 p.m.