cec517350d259a4f49a58a769db784c248cfffa8,mlxtend/evaluate/accuracy.py,,accuracy_score,#Any#Any#Any#Any#Any#,17

Before Change

    elif method == "average":
        return sum([_compute_metric(np.where(target_temp != l, 1, 0),
                    np.where(predicted_temp != l, 1, 0))
                    for l in unique_labels]) / float(unique_labels.shape[0])

    else:
        raise ValueError("`method` must be "standard", "average""

After Change

    elif method == "average":
        all_class_acc = []
        for c in np.unique(y_target):
            positive_labels = (y_target == c)
            class_acc = np.mean((y_predicted == c)[positive_labels])
            all_class_acc.append(class_acc)

        return np.mean(all_class_acc)

    else:
        raise ValueError("`method` must be "standard", "average""

Instances