I am using caret package in R to train CART model. train function seems to tune only the complexity parameter (which in a way determines depth of the tree and number of terminal nodes). Is this parameter alone enough in general?
In the rpart.control function, I have seen that there are many other parameters that can be tuned as well (minsplit, minbucket..etc). Is cp in a way also covering these? If not, why in the caret package, cp is taken as the main tuning parameter?
Best Answer
The others could make a difference, but the complexity parameter by and large is the fundamental driver for over/under-fitting.
You can tune the others using a custom model if you like.