# Caret and Boruta: exact effect of 'ntree' parameter while calling train() function

I am planning to perform all-relevant feature selection using Boruta R-package. I was also happy to find a nice wrapper for doing this with a simultaneous mtry parameter adjustment implemented in caret. Please, correct me if I'm wrong... But it seems like that defining parameter ntree (see my code) will influence only the mtry adjustment time, but not the call of original Boruta() function - am I right? Here is my function call:

> library(caret);
# The following command adjusts **mtry** parameter:
> res <- train(My_training_dataset, My_labels, method = "Boruta", tuneLength = 10, ntree=10000);


And could you please make a rough guess about how much time it is going to take for a binary classification problem with very high dimensionality (> 200k of features)? Maybe it is better to end up only with mtry adjustment? Thank you very much beforehand!

-