scikit-learn 在多标签分类中计算 F1

Question

我正在尝试在 multi-label classification

中使用 scikit 计算 macro-F1

from sklearn.metrics import f1_score

y_true = [[1,2,3]]
y_pred = [[1,2,3]]

print f1_score(y_true, y_pred, average='macro')

但是失败并显示错误消息

ValueError: multiclass-multioutput is not supported

如何计算多标签分类的macro-F1？

Answer 1

在当前的 scikit-learn 版本中，您的代码会导致以下警告：

DeprecationWarning: Direct support for sequence of sequences multilabel
    representation will be unavailable from version 0.17. Use
    sklearn.preprocessing.MultiLabelBinarizer to convert to a label
    indicator representation.

按照此建议，您可以使用 sklearn.preprocessing.MultiLabelBinarizer 将此多标签 class 转换为 f1_score 接受的格式。例如：

from sklearn.preprocessing import MultiLabelBinarizer
from sklearn.metrics import f1_score

y_true = [[1,2,3]]
y_pred = [[1,2,3]]

m = MultiLabelBinarizer().fit(y_true)

f1_score(m.transform(y_true),
         m.transform(y_pred),
         average='macro')
# 1.0