Assume that a training data is randomly split into the following two subsets: growing set and pruning set.

Assume that the following decision tree is learned from the growing set (where the attribute at each node is likely randomly selected during tree construction):

What is the classification error rate of the above tree on the pruning set?

(2) Assume that you would like to use reduced-error pruning method. What will be the tree after the first iteration of the pruning method? That is, which part of the tree you would remove first in the reduced-error pruning method?

Growing set: A BC Class I O I 4 011 5011 6100+ 1110 12 10 Pruning set: A B C Class 120+ 131 0+ 1410- 1500+ (6,6) 0 (3,2) B B 3+,4-) 0