Machine Learning Concept

Decision Tree

Pick a feature and a threshold — one axis-parallel cut splits the data into two leaves.

split on threshold 5.0

leaf · left

6A / 1B

gini 0.24

leaf · right

1A / 6B

gini 0.24

weighted gini0.24

puritygood split

tip: a tree keeps splitting leaves until they are pure

tree.py