Final answer:
The 'gain' used to build a tree in XGBoost is Split Gain, which is dependent on either Information Gain or Gini Gain.
Step-by-step explanation:
The 'gain' used to build a tree in XGBoost is Split Gain. XGBoost stands for Extreme Gradient Boosting, which is a machine learning algorithm that uses gradient boosting to train decision trees. Split Gain refers to the improvement in the model's performance that is obtained from splitting a node in the tree.
Split Gain is calculated based on the reduction in the impurity of the target variable that is achieved by performing a split. In XGBoost, impurity is typically measured using either the Information Gain or the Gini Gain. Information Gain measures the reduction in entropy, while Gini Gain measures the reduction in the Gini index.
Therefore, Split Gain in XGBoost is dependent on either Information Gain or Gini Gain, as these are used to assess the quality of a split.