skactiveml.pool.vote_entropy

skactiveml.pool.vote_entropy(votes, classes)

Calculates the vote entropy for measuring the level of disagreement in QueryByCommittee.

Parameters

votesarray-like, shape (n_samples, n_estimators)

The class predicted by the estimators for each sample.

classesarray-like, shape (n_classes)

A list of all possible classes.

Returns

vote_entropynp.ndarray of shape (n_samples,)

The vote entropy of each row in votes.

References

1

S. P. Engelson and I. Dagan. Minimizing Manual Annotation Cost in Supervised Training from Corpora. In Annu. Meet. Assoc. Comput. Linguist., pages 319–326, 1996.