Back to Glossary Index
Dagster Data Engineering Glossary:
Early Stopping
A form of regularization used to avoid overfitting when training a learner with an iterative method, such as gradient descent, by stopping the training process before it completes all iterations.