Splitting Let's say that you further split the left...

Splitting Let's say that you further split the left partition on a1 = 1.5, 4.5. You get the following tree on splitting Yes a2 < 20 (15, 15) No Yes a1 = 1.5, 4.5 (13,5) (1,5) No (12, 0) (2, 10) What is the approximate Gini index of the partitions obtained after splitting on a1 = 1.5, 4.5? 0.04 0.09 0.14 0.19

Labs

Want to see this concept in action?

NEW

Explore this concept interactively to see how it behaves as you change inputs.

View Labs

Transcript

00:01 Hello students, so the decision tree that has been split on the attributes a1 at the root node and then split again on the attributes a2 at the left child node.

00:11 So the number of the data point in each leaf node is shown in the parentheses.

00:16 So here written in the parentheses number.

00:19 So here to calculate the gini index of the partition obtained after splitting a1 equal to 1 .5, 4 .5, we need to consider the distribution of the target variable in each leaf node.

00:32 So the leaf node is here 1 ,5.

00:36 So this is a leaf node.

00:39 So here the gini index of the leaf node is calculated as follows.

00:44 So here to calculate the gini index 1 plus 5 equal to 6.

00:49 So here 1 minus 1 upon 6 whole square minus 5 by 6 whole square which comes 5 by 18.

00:57 After that node number 2.

01:00 So the node number 2 is 12, 0.

01:03 So this is this is also the leaf node 12, 0.

01:07 So to calculate the gini index is 12 plus 0 is 12.

01:12 So 1 minus 12 by 12 raised to power 2, 0 by 12 raised to power 2.

01:16 So it will come 0.

01:18 So this is the gini index for the second leaf node and the leaf node 3, 15, 15.

01:24 So here the leaf node 15, 15 we will calculate as follows...

Question

Please give Ace some feedback