jasaSIMULATION: Simulation studies for JASA submission

In hychen-uic/TEV: Total Explained Variation

Simulation studies for JASA submission

Description

Simulation studies for JASA submission

Usage

jasaSimulation(
  cindep = "T",
  sparse = "F",
  n = 400,
  N = 200,
  p = 200,
  p1 = 100,
  cs = 1,
  powx = 1,
  powy = 1,
  xsig = 1,
  errsig = 1,
  method = "our",
  rho = 0.9,
  nrep = 1000,
  rept = 1000
)

Arguments

cindep	determine if covariates are independent: cindep="T" if independent, cindep="F" if dependent.
sparse	determine whether effects are sparse: spare="T" if effects are sparse, sparse="F" if effects are dense
n	sample size of main data, e.g., n=400.
N	sample size for supplementary data, e.g., N=200, it is set to N=p in the simulation.
p	dimension of the covariates, for example, p=200,
p1	subdimension of covariates having nonzero effects p1<=p, e.g. p1=100 when p=200. in the simulation of sparse effects, p1=4.
cs	the constant for determining regression coefficients. it needs to be adjusted to achieve R2 wanted.
powx	powertansformation of normal covariates, powx=1 means normal, powx=2 means chi-square
powy	powertansformation of normal random error, powx=1 means normal, powx=2 means chi-square
xsig	standard deviation of covariates
errsig	standard deviation of random error.
method	method for dependence generate: method="our" means our approach, otherwise Cai and Guo's approach.
rho	correlation coefficient in the AR(1) covariate dependence model. This has effects only when method!="our".
nrep	number of replicates in simulation sample for variance estimation nrep=1000 by default.
rept	number of replicates in simulation rept=1000 by default.

Value

Estimate of the proportion of the explained variation and confidence intervals for the proportion.

Examples

## Not run: jasaSIMULATION(cindep="T",sparse="F", p=200, p1=100, n=400,N=200,
                                  powx=1,powy=1,xsig=1,errsig=1,rho=0.9,rept=100,nrep=1000)
## End(Not run)

hychen-uic/TEV documentation built on Jan. 24, 2025, 9:14 p.m.