Ultimate access to all questions.
Explanation:
Gini impurity measures variance across classes and is used for determining the best split point between regions of a decision tree. The highest information gain is the segmentation that creates the largest Gini measure reduction.
Gini = $1 - \left[\left(\frac{25}{30}\right)^2 + \left(\frac{5}{30}\right)^2\right] = 1 - (0.6944 + 0.0278) = 0.28$
(Book 2, Module 26.2, LO 26.e)
Question 29
Assume a portfolio consists of 30 large-cap value stocks, and that 25 of these companies currently pay dividends while 5 companies do not. Given a machine learning decision tree model, what is the closest computed Gini measure for this scenario?
A
0.45.
B
0.30.
C
0.55.
D
0.70.
No comments yet.