[R] Jackknife estimates of predict.lda success rate

[Prof Brian Ripley]

Reposting the same message is not making it any easier for your audience 
to understand.

You need to tell us what you mean by your subject line and 'jackknife 
reclassification'.  Given these are far from standard terms, either you 
mean something else (and 'jackknife' is very frequently misused for 'leave 
one out cross-validation) or it is likely that you will need to code the 
method yourself.

On Wed, 28 Mar 2007, Mat.Vanderklift at csiro.au wrote:

> Dear all

> I have used the lda and predict functions to classify a set of objects 
> of unknown origin. I would like to use a jackknife reclassification to 
> assess the degree to which the outcomes deviate from that expected by 
> chance. However, I can't find any function that allows me to do this. 
> Any suggestions of how to generate the jackknife reclassification to 
> assess classification accuracy? (BTW, commands I used to generate 
> predictions appended below.)

> Many thanks
> Mat Vanderklift
>> library(MASS)
>> Sep04 <- read.table("Sep04 dataframe2.txt", header=TRUE)
>> trainSep04 <- Sep04[1:125,]
>> Sep04.lda <- lda(Reef ~ ., trainSep04)
>> Sep04.lda
>> predict.Sep04 <- predict(Sep04.lda, Sep04) $class
>  
>
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>

-- 
Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595