R devel - Nov 2001 - confusion() -- generic function in package mda -- generally useful?

For a private package ("soon" to be on CRAN),
I started looking at and thinking about the function confusion() from the
mda package and wondered if it wouldn't make sense to provide the generic
function in base R and also think about making the default for the `true'
argument equal to  fitted(object).

One could write methods for confusion() for almost all classifiers 
which would be a nice universal approach.
Also, the value (a confusion matrix contingency table with some
"useful"
attributes) could be made a bit more explicit.

Good or not so good idea?  alternatives?

Martin Maechler <maechler@stat.math.ethz.ch>
http://stat.ethz.ch/~maechler/
Seminar fuer Statistik, ETH-Zentrum  LEO D10	Leonhardstr. 27
ETH (Federal Inst. Technology)	8092 Zurich	SWITZERLAND
phone: x-41-1-632-3408		fax: ...-1228			<><
>>   Confusion Matrices
>> 
>>   Description:
>> 
>>        Compute the confusion matrix between two factors, or for an fda
or
>>        mda object.
>> 
>>   Usage:
>> 
>>        confusion(object, true, ...)
>>        confusion(object, data, ...)
>> 
>>   Arguments:
>> 
>>     object: the predicted factor, or an fda or mda model object.
>> 
>>       true: the true factor.
>> 
>>       data: a data frame (list) containing the test data.
>> 
>>        ...: further arguments to be passed to or from methods.
>> 
>>   Details:
>> 
>>        This is a generic function.
>> 
>>   Value:
>> 
>>        For the default method essentially `table(object, true)',
but with
>>        some useful attribute(s).
>> 
>>   See Also:
>> 
>>        `fda', `predict.fda'
>> 
>>   Examples:
>> 
>>        data(iris)
>>        irisfit <- fda(Species ~ ., data = iris)
>>        confusion(predict(irisfit, iris), iris$Species)
>>        ##            Setosa Versicolor Virginica 
>>        ##     Setosa     50          0         0
>>        ## Versicolor      0         48         1
>>        ##  Virginica      0          2        49
>>        ## attr(, "error"):
>>        ## [1] 0.02
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-devel mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To:
r-devel-request@stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._

>>>>> Martin Maechler writes:
> For a private package ("soon" to be on CRAN),
> I started looking at and thinking about the function confusion() from
> the mda package and wondered if it wouldn't make sense to provide the
> generic function in base R and also think about making the default for
> the `true' argument equal to fitted(object).
> One could write methods for confusion() for almost all classifiers
> which would be a nice universal approach.  Also, the value (a
> confusion matrix contingency table with some "useful" attributes)
> could be made a bit more explicit.
> Good or not so good idea?  alternatives?
I think it would be nice to unify some of the existing code for
classification, e.g. the types in the predict methods.

Re confusion(), what would this do exactly?  The stuff in mda does too
much as exemplified by

\method{confusion}{default}(object, true, \dots)
\method{confusion}{fda}(object, data, \dots)

so the default and fda methods have surprising arguments ...?  (I think
you only want the versions for fitted model objects.)

-k
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-devel mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To:
r-devel-request@stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._