Glossary term
Glossary term
Foundations
While training a decision tree, the routine (and algorithm) responsible for finding the best condition at each node.
Created for this library
A research engineer uses a learned splitter to decide which feature and threshold to use for a decision tree split.
A risk modeling team uses an axis-aligned splitter in its production decision trees because each split is easy to audit.
A research team experiments with oblique splitters in research prototypes when single-feature splits cannot separate classes cleanly.
Definition source: Google for Developers Machine Learning Glossary | Creative Commons Attribution 4.0 License