Glossary term
Glossary term
Training and Fine-Tuning
Applying a weight to the downsampled class equal to the factor by which you downsampled.
Created for this library
A fraud team applies upweighting to the rare positive class during training so the gradient signal is not dominated by majority examples.
A medical screening team applies upweighting to rare-disease cases so the model sees their loss contributions proportionally.
A retention team applies upweighting to recently observed churn examples so the model adapts to current behavior more quickly.
Definition source: Google for Developers Machine Learning Glossary | Creative Commons Attribution 4.0 License