python编译报错 ImportError: cannot import name ‘jaccard_similarity_score‘

 

报错如下：

问题代码：

import numpy as np
from sklearn.metrics import jaccard_similarity_score
y_pred = [0, 2, 1, 3]
y_true = [0, 1, 2, 3]
print (jaccard_similarity_score(y_true, y_pred))

一开始以为是scikit-learn安装出了问题，尝试直接定义jaccard，修改sklearn为scikit-learn都不行，找到好久才发现是引入的包的文件名不对。

报错：from sklearn.metrics import jaccard_similarity_score。

原因：新的scikit-learn不再自动修改语法。通过路径找到存放jaccard的py文件，发现其中的jaccard函数名称为jaccard_score

修改：将 from sklearn.metrics import jaccard_similarity_score改为 from sklearn.metrics import jaccard_score

修改后的代码：

import numpy as np
from sklearn.metrics import jaccard_score
y_pred = [0, 2, 1, 3]
y_true = [0, 1, 2, 3]
print (jaccard_score(y_true, y_pred))

仍然有报错：ValueError: Target is multiclass but average='binary'. Please choose another average setting, one of [None, 'micro', 'macro', 'weighted'].

原因：二分类时average参数默认是binary；多分类时，可选参数有micro、macro、weighted和samples。

修改：将jaccard_score(y_true, y_pred)改为jaccard_score(y_true, y_pred,average='micro)

编译后发现能正常运行了

import numpy as np
from sklearn.metrics import jaccard_score
y_pred = [0, 2, 1, 3]
y_true = [0, 1, 2, 3]
print (jaccard_score(y_true, y_pred,average='micro'))

运行结果