model_eval.py 文件源码

def calculate_binary_classification_metrics(trained_sklearn_estimator, x_test, y_test):
    """
    Given a trained estimator, calculate metrics.

    Args:
        trained_sklearn_estimator (sklearn.base.BaseEstimator): a scikit-learn estimator that has been `.fit()`
        x_test (numpy.ndarray): A 2d numpy array of the x_test set (features)
        y_test (numpy.ndarray): A 1d numpy array of the y_test set (predictions)

    Returns:
        dict: A dictionary of metrics objects
    """
    # Squeeze down y_test to 1D
    y_test = np.squeeze(y_test)

    _validate_predictions_and_labels_are_equal_length(x_test, y_test)

    # Get binary and probability classification predictions
    binary_predictions = np.squeeze(trained_sklearn_estimator.predict(x_test))
    probability_predictions = np.squeeze(trained_sklearn_estimator.predict_proba(x_test)[:, 1])

    # Calculate accuracy
    accuracy = skmetrics.accuracy_score(y_test, binary_predictions)
    roc = compute_roc(y_test, probability_predictions)
    pr = compute_pr(y_test, probability_predictions)

    # Unpack the roc and pr dictionaries so the metric lookup is easier for plot and ensemble methods
    return {'accuracy': accuracy, **roc, **pr}