Draws non-parametric bootstrap samples
Usage
# S3 method for class 'lvm'
bootstrap(x,R=100,data,fun=NULL,control=list(),
p, parametric=FALSE, bollenstine=FALSE,
constraints=TRUE,sd=FALSE, mc.cores,
future.args=list(future.seed=TRUE),
...)
# S3 method for class 'lvmfit'
bootstrap(x,R=100,data=model.frame(x),
control=list(start=coef(x)),
p=coef(x), parametric=FALSE, bollenstine=FALSE,
estimator=x$estimator,weights=Weights(x),...)Arguments
- x
lvm-object.- R
Number of bootstrap samples
- data
The data to resample from
- fun
Optional function of the (bootstrapped) model-fit defining the statistic of interest
- control
Options to the optimization routine
- p
Parameter vector of the null model for the parametric bootstrap
- parametric
If TRUE a parametric bootstrap is calculated. If FALSE a non-parametric (row-sampling) bootstrap is computed.
- bollenstine
Bollen-Stine transformation (non-parametric bootstrap) for bootstrap hypothesis testing.
- constraints
Logical indicating whether non-linear parameter constraints should be included in the bootstrap procedure
- sd
Logical indicating whether standard error estimates should be included in the bootstrap procedure
- mc.cores
Optional number of cores for parallel computing. If omitted future.apply will be used (see future::plan)
- future.args
arguments to future.apply::future_lapply
- ...
Additional arguments, e.g. choice of estimator.
- estimator
String definining estimator, e.g. 'gaussian' (see
estimator)- weights
Optional weights matrix used by
estimator
Examples
m <- lvm(y~x)
d <- sim(m,100)
e <- estimate(lvm(y~x), data=d)
## Reduce Ex.Timings
B <- bootstrap(e,R=50,mc.cores=1)
B
#> Non-parametric bootstrap statistics (R=50):
#>
#> Estimate Bias Std.Err 2.5 % 97.5 %
#> y -0.053821882 0.003045923 0.101716706 -0.197865173 0.162723203
#> y~x 0.900266219 0.023337709 0.085382077 0.750455514 1.086770784
#> y~~y 1.054489157 -0.026246424 0.126152114 0.807799324 1.290297648
#>