Glossary term
Glossary term
Foundations
The more common label in a class-imbalanced dataset. For example, given a dataset containing 99% negative labels and 1% positive labels, the negative labels are the majority class.
Contrast with minority class.
See Datasets: Imbalanced datasets in Machine Learning Crash Course for more information.
Created for this library
A fraud team treats non-fraud as the majority class and uses class weights so the model is not driven entirely by majority examples.
A churn team handles the majority class of retained customers by undersampling during training to keep the gradient balanced.
A medical screening team handles the majority class of healthy patients with class weights so rare positive cases still influence training.
Definition source: Google for Developers Machine Learning Glossary | Creative Commons Attribution 4.0 License