Trying to make sense of this Decision Trees question

#1
Hi everyone, I'm trying to figure out how the solution to this decision trees question from college makes sense but there's not much explanation and in the notes there are no similar examples either so I'm lost as to where to begin. I'll attach the question/solution to this post. Thanks!
1610919688014.png 1610919738921.png 1610919788582.png
 
#2
For the first part, it seems you need to consider each point in A and B and find the buckets they belong to. Then, depending on the number of points from each dataset the bucket is A-like or B-like. The goal is to separate the datasets. So, anytime the bucket has a mix of A and B items it is a misclassification. The last part seems open-ended. They want you to create a decision tree for the datasets with no misclassification.
 
Last edited:

hlsmith

Less is more. Stay pure. Stay poor.
#3
Your responses seem fine. Usually percentages are in the terminal nodes when it is a classification tree. So then if you have 70% A's in R1 and 40% A's in R2, R1 would be yes and R2 no.