dataeval.core.nullmodel_accuracy¶

dataeval.core.nullmodel_accuracy(class_prob, model_prob, *, multiclass=False)¶

Compute accuracy from binary classification results.

Parameters:¶

class_prob : Array1D[float]¶

Class-wise probabilities for the test set. Can be a 1D list, or array-like object.

model_prob : Array1D[float]¶

Probability distribution for given null model. Can be a 1D list, or array-like object.

multiclass : bool, default False¶

Whether to compute multiclass accuracy

Returns:¶

Computed accuracy for binary classification

Return type:¶

np.float64