Background of regression trees

Regression trees divide the data into subsets, that is, branches, nodes, and leaves. Like decision trees, regression trees select splits that decrease the dispersion of target attribute values. Thus, the target attribute values can be predicted from their mean values in the leaves.

The resulting model is piecewise-constant. It has fixed predicted values that are assigned to regions for which the domain is split by the tree structure.

Creating and using regression tree models comprises the following algorithmic subtasks for regression trees:

Growing
Pruning
Prediction