mlp model with RSNNS package masks confusionMatrix, train

[tobigithub]

HI,
downloading the test or example data from > model <- "mlp" the following issue occurs:

> library(RSNNS)
Loading required package: Rcpp

Attaching package: ‘RSNNS’

The following objects are masked from ‘package:caret’:

    confusionMatrix, train

this will lead from the correct

> apropos("confusion")
[1] "confusionMatrix"       "confusionMatrix.train"

to the incorrect:

> apropos("confusion")
[1] "confusionMatrix"       "confusionMatrix"       "confusionMatrix.train"

This will further down the road lead to the following error for all packages that are used after the "mlp" training model such as

Error in confusionMatrix :  $ operator is invalid for atomic vectors
or
Error in toNumericClassLabels(targets) :    (list) object cannot be coerced to type 'double'

Still need to verify that but it is under RSNNS_0.4-7, Rcpp_0.12.0 , caret_6.0-52
under PRO 3.2.2 under Windows. Basically, code that was running OK, will after the install of the mlp training model, generate only errors such as confusionMatrix $ operator is invalid for atomic vectors or others. One solution would be to use:

caret::confusionMatrix(...)

But I have not tested that for all scenarios.
Cheers
Tobias

[topepo]

I checked in a change. Please give it a try to see if it fixes the issue for you.

Thanks,

Max

[tobigithub]

Cool. Thank you
Tobias

[topepo]

Looking at my changes, there really isn't anything else that I can do. The function in RSNNS isn't object oriented so that does not solve it. You will have to call using caret::confusionMatrix to get the right version when you have them both loaded.