# [R] using jackknife in linear models

Simon Knapp sleepingwell at gmail.com
Fri Dec 19 05:56:47 CET 2008

The definition of jackknife implies that the result of theta must be a
numeric scalar. So, presuming you wanted to jackknife a coefficient,
theta would need to be something like:

theta <- function(x, xdata, coefficient) coef(lm(model.lm,
data=xdata[x,]))[coefficient]

and would be called something like:
results <- jackknife(1:n,theta,xdata=DF,coefficient="(Intercept)")

To do all coefficients, you could write a function like like:
jackknife.apply <- function(x, xdata, coefs) sapply(coefs,
function(coefficient) jackknife(x, theta, xdata=xdata,
coefficient=coefficient), simplify=F)

and call it like:
results <- jackknife.apply(1:n, DF, c("(Intercept)", "V1", "V2", "V3"))

which would give you a list with the jackknife output for each coefficient.

note that you don't need to attach DF for your code to work :-)

Regards,
Simon Knapp.

On Fri, Dec 19, 2008 at 7:02 AM, Mark Heckmann <mark.heckmann at gmx.de> wrote:
> Hi R-experts,
>
> I want to use the jackknife function from the bootstrap package onto a
> linear model.
> I can't figure out how to do that. The manual says the following:

...

> So I tried:
>
>  DF <- as.data.frame(matrix(rnorm(250), ncol=5))
>  attach(DF)
>  model.lm <- formula(V1 ~ V2 + V3 + V4)
>  n <- 50
>  theta <- function(x, xdata){ lm(model.lm, data=xdata) }
>  results <- jackknife(1:n,theta,xdata=DF)