List of functionals (summaries of a bootstrap distribution)

List of functionals (summaries of a bootstrap distribution). These accept a matrix of bootstrap replicates as input, and possibly the observed value and weights, and calculate a summary such as quantiles, centered quantiles, standard error, mean or bias, for each column.

USAGE:

resampFunctionalList[["mean"]](x,              weights, ...) 
resampFunctionalList[["bias"]](x,    observed, weights, ...) 
resampFunctionalList[["se"]](x,                weights, ...) 
resampFunctionalList[["bias&se"]](x, observed, weights, ...) 
resampFunctionalList[["mean&se"]](x,           weights, ...) 
resampFunctionalList[["quantiles"]](x,                        weights,  
                              probs = c(0.025, 0.16, 0.5, 0.84, 0.975), ...) 
resampFunctionalList[["centered quantiles"]](x,     observed, weights, 
                              probs = c(0.025, 0.16, 0.5, 0.84, 0.975), ...) 
resampFunctionalList[["standardized quantiles"]](x, observed, weights, 
                              probs = c(0.025, 0.16, 0.5, 0.84, 0.975), ...) 

REQUIRED ARGUMENTS:

x
matrix with B rows (number of bootstrap samples) and p columns (length of the observed statistic), containing the bootstrap sample.
observed
vector of length p of observed statistics. This argument is required for functionals that use it.

OPTIONAL ARGUMENTS:

weights
vector of weights of length B, or NULL signifying equal weights. Calculations are done for the weighted distributions.
...
optional arguments, passed to the corresponding function, e.g. colMeans for "mean".

VALUE:

A vector or matrix of summary statistics, for example quantiles of each column. The "mean", "bias", and "se" examples return vectors of length p, the others return matrices with p columns and two or five (by default) columns.

DETAILS:

These are used by and .

You may define your own functional to pass to those functions; it should have the same initial arguments. The output need not have p columns or length p; however the plotting routines will work best if it does. If the output has names or dimnames they will be used in printing and plotting.

SEE ALSO:

These functionals are called by and

EXAMPLES:

x <- qexp(ppoints(19)) 
boot <- bootstrap(x, mean) 
plot(boot) # slightly skewed 
jab <- jackknifeAfterBootstrap(boot) 
plot(jab) 
plot(jab, xaxis = "L") 
plot(jab, xaxis = "data") 
 
jab <- jackknifeAfterBootstrap(boot, functional = "Centered Quantiles") 
plot(jab) 
 
# See also extensive examples in the help files for jackknifeAfterBootstrap 
# and tiltAfterBootstrap