The maximum gini impurity of a decision tree leaf occurs when a leaf is correct only 50% of the time. This means that the maximum gini impurity is 0.5. Conversely, the minimum gini impurity is 0. This happens when the leaf is correct 100% of the time.

Example:

flowchart TD A[Loves popcorn] A --> |True| B[Loves the song\nCorrect predictions: 100\nIncorrect predictions: 0\nGini impurity: 0] A --> |False| C[Deos not love the song\nCorrect predictions: 50\nIncorrect predictions: 50\nGini impurity: 0.5]

References⌗