SGDClassifier與class_weight =自動上失敗scikit學習0.15而不是0.14

Question

當我訓練的scikit-learn v0.15 SGDClassifier這些選項：SGDClassifier(loss='log', class_weight=None, penalty='l2')，訓練，並沒有錯誤完成。 然而，當我在訓練這個分類與class_weight='auto' scikit學習v0.15，我得到這個錯誤：

return self.model.fit(X, y) 
    File "/home/rose/.local/lib/python2.7/site-packages/scikit_learn-0.15.0b1-py2.7-linux-x86_64.egg/sklearn/linear_model/stochastic_gradient.py", line 485, in fit 
    sample_weight=sample_weight) 
    File "/home/rose/.local/lib/python2.7/site-packages/scikit_learn-0.15.0b1-py2.7-linux-x86_64.egg/sklearn/linear_model/stochastic_gradient.py", line 389, in _fit 
    classes, sample_weight, coef_init, intercept_init) 
    File "/home/rose/.local/lib/python2.7/site-packages/scikit_learn-0.15.0b1-py2.7-linux-x86_64.egg/sklearn/linear_model/stochastic_gradient.py", line 336, in _partial_fit 
    y_ind) 
    File "/home/rose/.local/lib/python2.7/site-packages/scikit_learn-0.15.0b1-py2.7-linux-x86_64.egg/sklearn/utils/class_weight.py", line 43, in compute_class_weight 
    raise ValueError("classes should have valid labels that are in y") 
ValueError: classes should have valid labels that are in y 

這是什麼原因呢？

僅供參考，這裏的一對class_weight文檔：

Preset for the class_weight fit parameter. Weights associated with classes. If not given, all classes are supposed to have weight one. The 「auto」 mode uses the values of y to automatically adjust weights inversely proportional to class frequencies.

Answer 1

我想這可能是內scikit學習中的錯誤。作爲變通，請嘗試以下操作：

from sklearn.preprocessing import LabelEncoder 
le = LabelEncoder() 
y_encoded = le.fit_transform(y) 
self.model.fit(X, y_encoded) 
pred = le.inverse_transform(self.model.predict(X)) 

Answer 2

我在這個公關努力解決這個問題：

https://github.com/scikit-learn/scikit-learn/pull/3515

請隨意測試和報告，如果它解決了問題你。