bootLongSubsampling: bootLongSubsampling

In PratheepaJ/bootLong: The Block Bootstrap Method for Longitudinal Count Data

Description

Finding optimal block size.

Usage

bootLongSubsampling(ps, main_factor, time_var, subjectID_var, sampleID_var,
  lI, R, RR, omega = 0.6, lC1 = 1, lC2 = NULL, ncores,
  psi.hat.lI = FALSE, psi.hat.lI.val = NULL,
  compStatParallel = FALSE)

Arguments

ps	phyloseq object.
main_factor	Character string. The name of the covariate variable.
time_var	Character string. The name of the time variable.
subjectID_var	Character string. The name of the subject ID variable.
sampleID_var	Character string. The name of the sample ID variable.
lI	A numeric. The initial block size.
R	A numeric. The number of block bootstrap realizations.
RR	A numeric. The number of double block bootstrap realizations.
omega	A numeric. The proportion between 0 and 1 of repeated observations for subsampling.
lC1	A numeric. (optional) any block size to start with. Defualt is 1.
lC2	A numeric. (optional) any block size to end with. defualt is (lI-1).
ncores	A numeric. The number of cores to use in mclapply.
psi.hat.lI	A logical. psi is estiamted already with the initial block size.
psi.hat.lI.val	A list. If psi is calculated already, give the list of values.
compStatParallel	A logical. True is used to use parallel in computeStat and disable parallel in bootLongPsi or bootLongMethod.

Value

A list of “MSE“ in calculating K with block sizes lC=1:(lI-1), “Khat“ values with block sizes lC=1:(lI-1) and “Khat.obs“ with lI

PratheepaJ/bootLong documentation built on April 5, 2020, 2:19 p.m.