Skip to content
ML Academy

Introduced on day 12

Early stopping

Ending iterative training when the score on a validation set has not improved for a set number of rounds, keeping the best round.

In detail

The stopping round is a hyperparameter chosen with validation data, so the validation score at that round is optimistic and cannot be the final report; use a separate untouched block. In boosting libraries the patience (early stopping rounds) and the evaluation metric are explicit parameters; if several evaluation sets or metrics are supplied, check which one drives the decision in your library's documentation.