Information gain Machine Learning
noun phrase
Definition: A measure of the reduction in entropy (uncertainty) achieved by splitting data according to a feature, commonly used in decision-tree learning [Google ML Glossary].
Example in context: “We now employ our cost-aware acquisition function, which makes use of an information gain based on objective space entropy.” [Iqbal et al. 2023]
Related terms: entropy reduction, mutual information (related but broader), Gini decrease