dscience.ml.accuracy_frame

Module Contents

class dscience.ml.accuracy_frame.AccuracyFrame

Bases: OrganizingFrame

Has columns ‘label’, ‘score’, ‘prediction’, and ‘score_for_prediction’, with one row per prediction.

classmethod required_columns(cls)

counts(self)

means(self)

with_label(self, label: Union[str, Iterable[str]])

boot_mean(self, b: int, q: float = 0.95)

Calculates a confidence interval of the mean from bootstrap over the rows. :param b: The number of bootstrap samples :param q: The high quantile, between 0 and 1. :return: A DataFrame with columns ‘label’, ‘lower’, and ‘upper’.

class dscience.ml.accuracy_frame.AccuracyFrames

classmethod concat(cls, *views: Sequence[AccuracyFrame])