Scikit F-score指标错误

Question

我试图使用SciKit的Logistic回归来预测一组标签.我的数据实际上是不平衡的(有更多'0'而不是'1'标签)所以我必须在交叉验证步骤中使用F1得分指标来"平衡"结果.

[Input]
X_training, y_training, X_test, y_test = generate_datasets(df_X, df_y, 0.6)
logistic = LogisticRegressionCV(
    Cs=50,
    cv=4,
    penalty='l2', 
    fit_intercept=True,
    scoring='f1'
)
logistic.fit(X_training, y_training)
print('Predicted: %s' % str(logistic.predict(X_test)))
print('F1-score: %f'% f1_score(y_test, logistic.predict(X_test)))
print('Accuracy score: %f'% logistic.score(X_test, y_test))

[Output]
>> Predicted: [0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0]
>> Actual:    [0 0 0 1 0 0 0 0 0 1 1 0 0 1 0 0 0 0 0 0 0 1 1]
>> F1-score: 0.285714
>> Accuracy score: 0.782609
>> C:\Anaconda3\lib\site-packages\sklearn\metrics\classification.py:958:  
   UndefinedMetricWarning:
   F-score is ill-defined and being set to 0.0 due to no predicted samples.

我当然知道这个问题与我的数据集有关:它太小了(它只是真实的一个样本).但是,任何人都可以解释我所看到的"UndefinedMetricWarning"警告的含义吗？窗帘后面究竟发生了什么？

Answer 1

看来这是一个已知的错误，已修复，我想你应该尝试更新 sklearn。