snorkel.utils.probs_to_preds¶

snorkel.utils.probs_to_preds(probs, tie_break_policy='random', tol=1e-05)[source]¶

Convert an array of probabilistic labels into an array of predictions.

Policies to break ties include: “abstain”: return an abstain vote (-1) “true-random”: randomly choose among the tied options “random”: randomly choose among tied option using deterministic hash

NOTE: if tie_break_policy=”true-random”, repeated runs may have slightly different results due to difference in broken ties

Parameters

prob – A [num_datapoints, num_classes] array of probabilistic labels such that each row sums to 1.
tie_break_policy (str) – Policy to break ties when converting probabilistic labels to predictions
tol (float) – The minimum difference among probabilities to be considered a tie

Returns

A [n] array of predictions (integers in [0, …, num_classes - 1])

Return type

np.ndarray

Examples

>>> probs_to_preds(np.array([[0.5, 0.5, 0.5]]), tie_break_policy="abstain")
array([-1])
>>> probs_to_preds(np.array([[0.8, 0.1, 0.1]]))
array([0])