metricChangePoint: Change Point Detection in Metric Spaces

In speegled/cpbaywave: Bayesian Wavelet based Change Point Detection

Change Point Detection in Metric Spaces

Description

Determines the most likely locations of a change point in distribution of elements in a metric space. Forms a new time series which is the pairwise distance matrix; the diagonal is replaced by average value of nearby elements. Then, dimension reduction change point detection is called. If useBootstrap is TRUE, then a bootstrapped confidence interval for possible change points is computed.

Usage

metricChangePoint(
  multiSeries,
  distance,
  useGaussian = TRUE,
  useBFIC = TRUE,
  useMetricProject = FALSE,
  setdetail,
  reducedDim = 10,
  useBootstrap = FALSE
)

Arguments

multiSeries	The high dimensional time series. Should be a matrix, with each row being an observation of the time series.
distance	A function that returns the distance between two rows of multiSeries.
useGaussian	Set to TRUE if you want to use a random Gaussian projection. Set to FALSE for random Bernoulli matrix.
useBFIC	Optional argument to use BFIC to decide change point location.
setdetail	Optional argument to set the detail level you wish to use. Default is all details except one.
reducedDim	The dimension you want to project onto. Should be less than the dimension of the time series. Default is 10.
useBootstrap	Set to true if you wish to bootstrap a cutoff for significance of indices.
...	Additional parameters passed to the distance function

Examples

dim=10
sig1=diag(dim)
mu1=rep(0,dim)
mu2=rep(1.5,dim)
n=128
tau=30
series1=createTimeSeries(mu1, mu2, sig1, n, tau)
sn=sin(2.5*pi*1:n/n)
series2=sn+series1
myDistance <- function(x,y) {
  sum(abs(x - y))
}
metricChangePoint(series2, myDistance) #True change point is at t = 30

speegled/cpbaywave documentation built on Feb. 19, 2024, 11:13 a.m.