The effect or randomly restricting the splitting variables is again illustrated by means of four bootstrap samples drawn from the smoking data: In addition to growing a large tree on each bootstrap sample, as in bagging, now the variable selection is limited to mtry=2 randomly preselected candidates in each split. The resulting trees are displayed in Figure 7: We find that, due to the random restriction, the trees have become even more diverse; for example the strong predictor variable friends_smoke is no longer chosen for the first split in every single tree.
