PR

class interpret.perf.PR(model, feature_names=None, feature_types=None)

Produces precision-recall curves.

Initializes class.

Parameters:

• model – model or prediction function of model (predict_proba for classification or predict for regression)

• feature_names – List of feature names.

• feature_types – List of feature types.

explain_perf(X, y, name=None)

Produce precision-recall curves.

Parameters:

• X – Numpy array for X to compare predict function against.

• y – Numpy vector for y to compare predict function against.

• name – User-defined explanation name.

Returns:

An explanation object.