
Β β with D_1 and D_2 subsets of D, π_π the probability of samples belonging to class π at a given node, and π the number of classes.The lower the Gini Impurity, the higher is the homogeneity of the node. The Gini Impurity of a pure node is zero. To split a decision tree using Gini Impurity, the following steps need to be performed.