Here we are only focusing on the train_test_split attribute of sklearn used in machine learning. In previous version of SciKit Learn the module cross_validation of package sklearn used to have this train_test_split attribute. But there has been some change in the pacakge binding and now the module is associated with the model_selection module.
Error is Occurred in the following scenario:
from sklearn import cross_validation
''' Your additional modules and
'''you can use your X and y values as per your problem'''
X_train, X_test, y_train, y_test = \
cross_validation.train_test_split(X, y, test_size=0.2, random_state=40)
Now in order to perform the cross validaiton with the split of test and train data you need to call cross_validate from module_selection like shown below:
from sklearn.model_selection import cross_validate
If you want to perform the test train splitting than you should implicitly import train_test_split like shown below:
from sklearn.model_selection import train_test_split
'''your other imports
and your code'''
train_test_split(X, y, test_size=0.2, random_state=40)