FPOPDp: FPOPDp

In lpishchagina/FPOPapprox2D: FPOPapprox2D

Description

Detection changepoints using the Functional Pruning Optimal Partitioning method (FPOP) in p-variate time series in a p-variable time series of length n.

Usage

FPOPDp(data, penalty, type)

Arguments

data	is a matrix of data(p-rows x n-columns).
penalty	is a value of penalty (a non-negative real number).
type	is a value defining the type of geometry for FPOP-pruning: type=1: ("intersection" of sets), approximation - rectangle; type=2:("intersection" of sets)"minus"("union" of sets), approximation - rectangle; type=3: (last disk)"minus"("union" of sets), approximation - disk.

Value

a list of elements = (changepoints, means, globalCost).

chpts

is the changepoint vector that gives the last index of each segment for the p-variate time series.

means

is the list of successive means for the p-variate time series.

globalCost

is a number equal to the global cost.

Examples

FPOPDp(data = data_genDp(p = 3, n = 100, chpts = 100, means = matrix(c (0, 0, 0), nrow = 3), noise = 1), penalty = 2*log(100), type = 2) 

lpishchagina/FPOPapprox2D documentation built on Dec. 21, 2021, 11:49 a.m.