module 'sklearn' has no attribute 'cross_validation'

问题

I am trying to split my dataset into training and testing dataset, but I am getting this error:

X_train,X_test,Y_train,Y_test = sklearn.cross_validation.train_test_split(X,df1['ENTRIESn_hourly'])

---

AttributeError                            Traceback (most recent call last)
<ipython-input-53-5445dab94861> in <module>()
----> 1 X_train,X_test,Y_train,Y_test = sklearn.cross_validation.train_test_split(X,df1['ENTRIESn_hourly'])

AttributeError: module 'sklearn' has no attribute 'cross_validation'

How can I handle this?

回答1:

sklearn does not automatically import its subpackages. If you only imported via: import sklearn, then it wont work. Import with import sklearn.cross_validation instead.

Furhter, sklearn.cross_validation will be deprecated in version 0.20. Use sklearn.model_selection.train_test_split instead.

回答2:

Try this:

from sklearn.model_selection import train_test_split
X_train, X_test, Y_train, Y_test = train_test_split(X, Y, test_size=0.33, random_state=101)

回答3:

The equivalent to cross_validation in sklearn is:

  sklearn.model_selection

回答4:

you can try this

X_train,X_test,Y_train,Y_test = 
    sklearn.model_selection.train_test_split(X,boston_df.price)

回答5:

"cross_validation" name is now deprecated and was replaced by "model_selection" inside the new anaconda versions. So you can use

from sklearn.model_selection import train_test_split

来源：https://stackoverflow.com/questions/46572475/module-sklearn-has-no-attribute-cross-validation