Gini index

The Gini index is a measure used in decision tree algorithms to evaluate how well a data split separates different groups. It quantifies the likelihood that a randomly chosen item from a set would be incorrectly labeled if it was assigned to a group based on the split. A Gini index of 0 indicates perfect separation (all items in a group are similar), while higher values mean more mixing or impurity. When building decision trees, the goal is to choose splits that result in the lowest Gini index, leading to more homogeneous groups and better classification accuracy.