resamp
. For each parameter, a histogram of replicates
with a smooth density estimate is plotted. A solid line indicates the observed
parameter value for the original data, and (when appropriate)
a dotted line indicates the mean of the replicates.
For the bootstrap, the difference between these two lines is the estimated bias.
plot.resamp(x, nrow = NULL, grid.layout = T, rugplot = F, nclass.func = nclass.fd, bandwidth.func = bandwidth.nrd, subset.statistic = 1:p, key = T, corner=c(1,1), ..., inside=F, nclass, main)
resamp
, e.g. produced by
or
.
nrow
is NULL and
x
has a
dim.obs
component, it is used to determine
nrow
. Otherwise, two rows are
used unless there is only a single parameter.
nclass.fd
is used. This function uses the
Freedman and Diaconis rule as described in Venables and Ripley.
density
when
constructing the density estimate. By default the function
bandwidth.nrd
,
is used. This function uses the
normal reference density bandwidth estimate as described by Venables and
Ripley.
TRUE
(the default), then a legend is added,
describing the solid and dotted lines indicating the observed
value and mean of the replicates, and plotting characters are
added at the bottom of the lines to help distinguish them.
c(0,1)
indicates the top left.
hist
and
barplot
.
TRUE
then lines are drawn between histogram bars;
default
FALSE
.
nclass.func
.
main=""
to avoid a title.
Venables, W.N. and Ripley, B.D. (1997), Modern Applied Statistics with S-PLUS, Second Edition, Springer-Verlag.
temp <- bootstrap(stack.loss, var) plot(temp)