Bootstrap Dialog Functions

DESCRIPTION:

These functions are used by the Bootstrap dialog. menuBootstrap calls tabSummary.bootstrap , tabPlot.bootstrap and tabJab.bootstrap if summary, plotting and jackknife-after-bootstrap results are requested.

USAGE:

menuBootstrap(data, statistic, B, group, seed, block.size, trace, 
     assign.frame1, save.indices, probs=c(25, 50, 950, 975)/1000, 
     print.short.p=T, print.emp.p=F, print.bca.p=F, print.cor.p=F, 
     plot.p=T, plotQQ.p=F, functional.jab=mean, print.jab.p=F, 
     plot.jab.p=F, save.name.jab=NULL, ...) 
tabSummary.bootstrap(boot.obj, probs=c(25, 50, 950, 975)/1000, 
     print.short.p=T, print.emp.p=F, print.bca.p=F, print.cor.p=F) 
tabPlot.bootstrap(boot.obj, plot.p, plotQQ.p) 
tabJab.bootstrap(boot.obj, functional.jab=mean, print.jab.p=F, 
     plot.jab.p=F, save.name.jab=NULL, call.new.graphsheet=T) 

REQUIRED ARGUMENTS:

data
data to be bootstrapped. May be a vector, matrix, or data frame.
statistic
statistic to be bootstrapped. It may be a function which accepts data as the first argument and returns a vector or matrix, or a call referring to the data which evaluates to a vector or matrix. If it is an expression and the data argument is the name of an object, then the object should also be referred to by name in the expression. If the data are constructed within the call to bootstrap, then data should be referred to as data in the expression. (See examples for bootstrap.)
boot.obj
an object of class bootstrap.

OPTIONAL ARGUMENTS:

B
number of bootstrap resamples to be drawn. We recommend at least 250 to estimate standard errors and 1000 to estimate percentiles.
group
allows stratified sampling and bootstrapping multi-sample problems. The unique values of this vector determine groups. For each resample, a bootstrap sample is drawn separately for each group, and the observations are combined to give the full resample. The statistic is calculated for the resample as a whole.
seed
seed for generating resampling indices. May be a legal random number seed or an integer between 0 and 1000 which will be passed to set.seed.
block.size
control variable specifying the number of resamples to calculate at once. bootstrap uses an lapply() within a for() loop. For small sample sizes, a single lapply() is reasonable, while for large sample sizes, a series of separate lapply()s is more efficient. Using sampler=samp.boot.bal gives balancing done separately within each group of resamples.
trace
logical flag indicating whether the algorithm should print a message indicating which set of replicates is currently being drawn.
assign.frame1
logical flag indicating whether the resampled data should be assigned to frame 1 before evaluating the statistic. This may be necessary if the statistic is reevaluating the call of a model object. If all bootstrap estimates are identical, try setting assign.frame1=T. Note that this will slow down the algorithm.
save.indices
logical flag indicating whether to save the matrix of resampling indices.
probs
probability levels at which to calculate percentiles of confidence intervals.
print.short.p
logical flag indicating whether to print basic summary statistics.
print.emp.p
logical flag indicating whether to print empirical percentiles.
print.bca.p
logical flag indicating whether to print BCa percentiles.
print.cor.p
logical flag indicating whether to print correlation matrix of statistic if statistic is a vector.
plot.p
logical flag indicating whether to plot the replicates with plot.resamp.
plotQQ.p
logical flag indicating whether to plot a QQ-plot of the replicates with qqnorm.resamp.
functional.jab
functional of the bootstrap distribution for which standard error is to be estimated when using jackknife-after-bootstrap. May be a character string specifying a column of boot.obj$stats (i.e. "Bias", "Mean", or "SE") or a function expecting a vector as its first argument and returning a scalar, e.g. max.
print.jab.p
logical flag indicating whether to print jackknife-after-bootstrap summaries.
plot.jab.p
logical flag indicating whether to plot a jackknife-after-bootstrap influence plot.
save.name.jab
name under which to save the jack.after.bootstrap object produced by the jackknife-after-bootstrap analysis.
...
other arguments to pass to the bootstrap function.

VALUE:

menuBootstrap returns an object of class bootstrap. See the bootstrap help files for details. tabSummary.bootstrap and tabPlot.bootstrap return the boot.obj. tabJab.bootstrap returns the jack.after.bootstrap object.

SIDE EFFECTS:

If save.name.jab is specified, the jack-after-bootstrap is assigned to the working database under the specified name.

SEE ALSO:

, , , qqnorm.resamp', , .

EXAMPLES:

menuBootstrap(fuel.frame, var(Weight))