PROBLEM $1[30$ points$]$

A $[20$pt$].$ Use information gain to build a decision tree that predicts the value of the target feature PLAY based on the values of other input

features such as ACE, TEN, and FIRST$\_$MOVE. Please use the training data provided below and shows the steps of your calculations.

Suppose that a tree is considered not optimal if there is another tree that achieves the same classification error on the training data but has

smaller depth.

B $[5$pt$].$ Is the tree found in the example above optimal? Explain why or why not.

C $[5$pt$].$ If it is not optimal, draw the optimal tree as well. You can draw it by hand on a piece of paper, $L,$

Note that: if the link is not publicly accessible, the TA will not be able to load it and grade this part of your submission.

ANSWER:

PROBLEM points$]$

Express the concept PLAY$=$Hit learned by all the trees found in PROBLEM $1$ in terms of logical if$-$then rules.