how adjust parameters in method of train function (caret package)

Question

I'm trying to do a cross validation resampling using the caret package but the function I want to train need some parameters. I want to use the function nmf from the package NMF:

res <- nmf(dfrm, 6, nrun=20)

and the code

inTraining <- createDataPartition(dfrm, p = .8, list = FALSE, times = 200)
training <- dfrm[ inTraining,]
testing  <- dfrm[-inTraining,]

fitControl <- trainControl(## 5-fold cv
                       method = "cv",
                       number = 5
                       )

cv_result <- train(dfrm.T, res, 
             data = training,
             method = "nmf",
             trControl = fitControl)

but if I just put nmf in the methods it's not going to use the 6 and nrun=20. How can I adjust it? thanks

Answer 1

nmf is not an available method in caret.

Using additional function parameters in train is possible. For example the ntree argument with randomForest. Tuning is only on mtry, but you can add additional function parameters.

train(Species ~ ., data = iris, method = "rf", ntree = 10)