s-news
[Top] [All Lists]

Problems with bootstrapping arguments

To: <s-news@lists.biostat.wustl.edu>
Subject: Problems with bootstrapping arguments
From: René Holst <dsl8955@vip.cybercity.dk>
Date: Wed, 20 Feb 2002 04:10:20 +0100
Importance: Normal
Reply-to: <rene@constat.dk>
Hi'

S+ 4.5 R.2, Win2K

I have a number, say n, of r*j matrices, each representing a sampling unit.
I want to bootstrap them, i.e apply a function to the sum of the matrices.
The matrices have been "cbinded" into a big r*(n*j) matrix dat, which is
used as the argument for the "bootstrap func.

boot.fun<-function(bdat){
....
#First pick out which matrices to use for the bootstrap sample by

boot.idx_sample(n,n,T)

#next pick out the matrices and pool them by

boot.dat<-matrix(apply(sapply(j*boot.idx-(j-1),function(i,locdat){locdat[,i:
(i+4)]},bdat),1,sum),ncol=j)
...
}

The code works fine from the commandline, but when called within the
boot.fun from by

bootstrap(dat,boot.fun)

I get the message : Error in [: Need 2 subscripts for array: .Dim=c(36, ..)

(r=36)

Hope someone can explain to me what goes wrong.

Another question: The manual and helpfile say that the data argument to the
bootstrap function must be a vector, a matrix or a dataframe. Does bootstrap
not accept arrays (3 dimensional) ?

Thanks in advance

Rene



*****************************************
René Holst
ConStat
Groenspættevej 10
DK-9800 Hjoerring
Denmark
Tel: +45 98 92 19 79
Mob: +45 24 40 79 41
Fax: +45 98 92 19 56
Web: http://www.constat.dk
SMS: 24407941@sms.tdm.dk (max 160 characters)
*****************************************



<Prev in Thread] Current Thread [Next in Thread>
  • Problems with bootstrapping arguments, René Holst <=