Sarath pointed out (privately) that the standard error is not actually
stored as part of the bootstrap object.
Instead, it is calculated on the fly when you print the bootstrap object;
this is done by
boot:::print.boot
Similarly for bias.
(The 'boot:::' is necessary because
"'print.boot' is not
Do
names(bootObj)
to find out what the components are, and use $ or [[ to extract
components.
Do
help(boot)
for a description of components of the object (look in the Value section).
That is general advice in R, applying to all kinds of objects -
boot, and many other functions such as lm(), re
Sarath,
Maybe someone can show a direct access to the statistic. One way to
get around is to access 't' component of the boot object which
contains individual estimates. so you can extract standard error by
sqrt(var(bootObj$t))
Weidong Gu
On Fri, Jul 22, 2011 at 11:36 AM, Sarath Gamini Bannehek
3 matches
Mail list logo
