Why do we use a weighted average of child entropies when we calculate information gain?

Asked May 13 '20 at 15:24

Active May 13 '20 at 20:24

Viewed 120 times

In the decision tree algorithm, why do we use a weighted average of child entropies when we calculate information gain? What is wrong about using the arithmetic mean of entropies?

edited May 13 '20 at 20:22

nbro

42,615
12
119
217

asked May 13 '20 at 15:24

Krushe

Why do we use a weighted average of child entropies when we calculate information gain?

0 Answers0