What is a branching criterion in trees?

The splitting criterion in decision trees is a metric or rule used to select the feature and value for dividing data into subsets at each node of the tree. The goal is to maximize the homogeneity (purity) of the resulting subsets to improve the quality of classification or regression.

Examples of splitting criteria:

• Gini Index (Gini impurity) — measures the probability of incorrectly classifying a randomly chosen object from the set.
• Information Gain — measures the reduction in uncertainty after the split.
• Misclassification Error — the proportion of incorrectly classified objects.

The choice of criterion affects the structure of the tree and the quality of the model. For example, the CART algorithm typically uses the Gini index, while ID3 uses information entropy.