ertk.classification.class_ratings_to_probs

ertk.classification.class_ratings_to_probs(ratings: Series, classes: List[str] | None = None) → ndarray

Convert annotator ratings into distribution over classes for each instance.

Parameters:

ratings: pd.Series

The Pandas Series containing the ratings.

classes: list of str, optional

Limit to these classes.

Returns:

numpy.ndarray

Matrix of shape (n_instances, n_classes) containing probabilities (relative frequencies) for each class.