SLIP.mcp.generator: Generate n-by-p matrix with at least one changes in each...

In MengtaoWen/SLIP: Change Testing for Large-Scale Datastreams with FDR Control

Generate n-by-p matrix with at least one changes in each abnormal datastream

Description

Generate n-by-p matrix with at least one changes in each abnormal datastream

Usage

SLIP.mcp.generator(
  n,
  p,
  cov = "Identity",
  rho = NULL,
  dist = "Gaussian",
  ratio = 0.15,
  delta = 1,
  varrho = 0.05,
  cpn = 4,
  Sigma = NULL,
  param = NULL
)

Arguments

n	the number of observations
p	the number of datastreams
cov	covariance type c("CS", "Factor", "AR1", "Block", "Sparse", "Identity", "Given")
rho	optional when using cov from c("CS", "AR1", "Block"), rho in [0, 1]
dist	noise distribution c("Gaussian", "t", "exp", "chisq")
ratio	abnormal streams/total streams = ratio, ratio in [0, 1]
delta	the magnitudes of changes (delta >= 0) lie in [delta-0.1, delta+0.1] with equally probable sign from +, -
varrho	the parameter avoiding the boundary problem, [0, 0.5)
cpn	the parameter of Poisson distribution related to the number of change-points at each of abnormal streams
Sigma	the covariance matrix, optional only when cov=('Given')
param	the parameter when using dist from c("t", "chisq", "exp")

Value

data with changes in some datastreasm

dat	n-by-p data matrix
index	those datastreams containing changes
cpnum	a vector of the number of change-points
mu	a list of values of changes
loc	a list of locations of changes

MengtaoWen/SLIP documentation built on May 3, 2022, 6:45 a.m.