dataeval.core.ParityResult¶

class dataeval.core.ParityResult¶

Type definition for parity output.

scores¶

Array of Bias-Corrected Cramér’s V statistics for each factor. Values range from 0.0 (independent/high parity) to 1.0 (perfect association).

Type:¶

NDArray[np.float64]

p_values¶

Array of p-values calculated via the G-test (Log-Likelihood Ratio). Indicates the statistical significance of the calculated association.

Type:¶

NDArray[np.float64]

insufficient_data¶

Dictionary flagging specific data subsets with low sample counts (< 5). Structure: {factor_index: {factor_category_value: {class_label: count}}}.

Type:¶

Mapping[int, Mapping[int, Mapping[int, int]]]