Question

Which best describes how a decision tree performs classification on a given example?

A) Starting at root, evaluate the binary decision to determine which tree branch to go to next. Stop when leaf node is reached.

B) Starting at root, select the tree branch that maximizes information gain. Stop when lead node is reached.

C) Starting at root, evaluate the binary decision to determine which tree branch to go to next. Stop when entropy is 0.

D) Starting at root, evaluate the binary decision to determine which tree branch to go to next. Stop when entropy is 1.

Answer 1

Correct options are (A) and (C), explanation for each option is given below :

Option (A) : Starting at root, evaluate the binary decision to determine which tree branch to go to next. Stop when leaf node is reached.

EXPLANATION :

• This is correct since it perfectly describe how a decision tree performs classification on a given example,Actually the decision tree starts from the root and also evaluates the binary decision to determine which tree branch go to next.

• The stopping decision is taken based on the entropy and it stops when the entropy is 0, i.e when leaf node is reached.

Option(B) : Starting at root, select the tree branch that maximizes information gain. Stop when lead node is reached.

EXPLANATION :

• This is incorrect since Decision tree tree is a top-down tree not the bottom-up tree ,Basically a decision tree starts from the root and evaluates the binary decision to determine which tree branch go to next.

• The stopping decision is taken based on the entropy or the probability and it stops when the entropy is 0, i.e when leaf node is reached, not when the lead node is reached.

Option (C) : Starting at root, evaluate the binary decision to determine which tree branch to go to next. Stop when entropy is 0.

EXPLANATION :

• This option is also correct as it says the same argument as stated in option A but in different manner,Actually the decision tree begins with the root node and computes the expression of the conditions in the binary form to determine which tree branch go to next.

• The stopping decision is taken based on the entropy and ends when the entropy is Zero.

Option (D) : Starting at root, evaluate the binary decision to determine which tree branch to go to next. Stop when entropy is 1.

EXPLANATION :

• This option is wrong since its argument is just opposite as that of the definition of the decision tree and also it does not comply with the rules of operation of the decision tree.

• The decision tree starts from the root and evaluates the expression of the conditions in the binary form to determine which tree branch go to next.

• The stopping decision is taken based on the entropy and it stops when the entropy is Zero.But in this option it is stated that it stops when the entropy is 1 .

So,both the options , (A) and (C) are correct in the given circumstances.