sbs.thr: Bootstrapping for threshold selection in SBS algorithm

In hdbinseg: Change-Point Analysis of High-Dimensional Time Series via Binary Segmentation

Bootstrapping for threshold selection in SBS algorithm

Description

Generate thresholds for SBS algorithm via bootstrapping

Usage

sbs.thr(
  z,
  interval = c(1, dim(z)[2]),
  cp.type = 1,
  do.clean.cp = TRUE,
  scales = NULL,
  diag = FALSE,
  sgn = NULL,
  B = 1000,
  q = 0.01,
  do.parallel = 4
)

Arguments

z	input data matrix, with each row representing the component time series
interval	a vector of two containing the start and the end points of the interval from which the bootstrap test statistics are to be calculated
cp.type, scales, diag, B, q, do.parallel	see sbs.alg
do.clean.cp	if do.clean.cp = TRUE pre-change-point cleaning is performed
sgn	if diag = FALSE, wavelet transformations of the cross-covariances are computed with the matching signs

Value

if cp.type = 1, a vector of length nrow(z), each containing the threshold applied to the CUSUM statistics from the corresponding coordinate of z if cp.type = 2, a vector of length length(scales)*nrow(z) (when diag = TRUE) or length(scales)*nrow(z)*(nrow(z)+1)/2 (when diag = FALSE), each containing the threshold applied to the CUSUM statistics of the corresponding wavelet transformation of z

hdbinseg documentation built on Aug. 17, 2023, 5:06 p.m.