Gini Change
Explore the concept of Gini change within classification trees to understand how the CART algorithm quantifies and minimizes impurity for better splits. Learn the step-by-step calculation process, including parent node impurity, weighted splits, and why larger observations are prioritized, enhancing your grasp of decision tree optimization.
We'll cover the following...
Minimizing Gini impurity
Gini impurity is the math behind how the CART classification tree algorithm quantifies the impurity for a collection of labels. The overall objective of the algorithm is to minimize impurity across all the training data.
With an intuitive understanding of how the algorithm splits data at each stage of tree building to minimize impurity, it’s time to learn the math of how the CART algorithm builds entire trees with the lowest impurity possible via the Gini change calculation.
Gini change
The Gini change calculation is defined as follows:
The math of Gini change isn’t complicated. The best way to learn how this math works ...