Decision trees, whether they are used for a continuous (regression) or discrete (classification) target variable, implement an algorithm called Recursive Binary Splitting.
- Search over the entire feature space, looking for the feature j and the value within that feature $v_j$