5. Bagging grows a large number of trees by

A. drawing random subsamples of the training set and then growing an unpruned tree on each subsample

B. drawing random subsamples of the test set and then growing an unpruned tree on each subsample

C. drawing random subsamples of the predictors and then growing an unpruned tree on each subsample

D. growing one unpruned tree and then randomly pruning it

E. growing one unpruned tree and then randomly changing the split values at each node

F. growing one unpruned tree and then randomly perturbing the predicted values at each terminal node

I know that it is not F.