使用scikit-learn计算精度的ValueError

Question

from sklearn.metrics import precision_score
precision_score(expected, predicted)

期望为array([ 4., 3.])

，预测为array([ 2., 4.])

我得到了报酬。error: *** ValueError: pos_label=1 is not a valid label: array([ 2., 3., 4.])

如何解决这个问题?

Answer 1

多类标签需要average参数

否则，您需要将pos_label设置为两个数组中的类标签之一，即2,3或4:

>>> # score for all classes
>>> precision_score(expected, predicted, average=None)
array([ 0.,  0.,  0.])
>>> # score for each class
>>> precision_score(expected, predicted, pos_label=2)
0.0
>>> precision_score(expected, predicted, pos_label=3)
0.0
>>> precision_score(expected, predicted, pos_label=4)
0.0

参考:sklearn.metrics.precision_score