limits.bca(boot.obj, probs=c(0.025, 0.05, 0.95, 0.975),
details=F, z0=NULL, acceleration=NULL,
group.size=NULL, frame.eval.jack=sys.parent(1))
bootstrap.
z0,
acceleration,
group.size, and the empirical probability levels for each BCa confidence
limit
along with the BCa limits. If
detail=F, only the BCa confidence limits are
returned.
limits.bca.
bca.percent.
Note that jackknifing is performed to calculate this value,
so a statistic must allow samples of reduced size.
floor(n/20), which yields roughly 20
groups of points, and hence 20 jackknife replicates. For the full jackknife,
use
group.size=1.
jackknife. By default, the frame from which
limits.bca was called. This is where S-PLUS will look for the data used
to construct the bootstrap object.
detail=F, returns a matrix of BCa confidence limits with one row per
parameter of interest. Otherwise, returns a list with these limits as
the
limits.bca component, plus the empirical probability levels,
z0
,
acceleration, and
group.size.
Efron, B. and Tibshirani, R. J. (1993).
An Introduction to the Bootstrap.
San Francisco: Chapman & Hall.
Shao, J. and Tu, D. (1995).
The Jackknife and Bootstrap.
New York: Springer-Verlag.
temp <- bootstrap(stack.loss, var) limits.bca(temp) limits.bca(temp, detail=T)