Parameter “stratify” from method “train_test_split

问题:

I am trying to use train_test_split from package scikit Learn, but I am having trouble with parameter stratify. Hereafter is the code:

from sklearn import cross_validation, datasets 

X = iris.data[:,:2]
y = iris.target

cross_validation.train_test_split(X,y,stratify=y)

However, I keep getting the following problem:

raise TypeError("Invalid parameters passed: %s" % str(options))
TypeError: Invalid parameters passed: {'stratify': array([0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 
0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 
1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 
2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 
2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2])}

Does someone have an idea what is going on? Below is the function documentation.

[...]

stratify : array-like or None (default is None)

If not None, data is split in a stratified fashion, using this as the labels array.

New in version 0.17: stratify splitting

[...]

回答1:

Scikit-Learn is just telling you it doesn't recognise the argument "stratify", not that you're using it incorrectly. This is because the parameter was added in version 0.17 as indicated in the documentation you quoted.

So you just need to update Scikit-Learn.

回答2:

This stratify parameter makes a split so that the proportion of values in the sample produced will be the same as the proportion of values provided to parameter stratify.

For example, if variable y is a binary categorical variable with values 0 and 1 and there are 25% of zeros and 75% of ones, stratify=y will make sure that your random split has 25% of 0's and 75% of 1's.

回答3:

For my future self who comes here via Google:

train_test_split is now in model_selection, hence:

from sklearn.model_selection import train_test_split

# given:
# features: xs
# ground truth: ys

x_train, x_test, y_train, y_test = train_test_split(xs, ys,
                                                    test_size=0.33,
                                                    random_state=0,
                                                    stratify=ys)

is the way to use it. Setting the random_state is desirable for reproducibility.

回答4:

In this context, stratification means that the train_test_split method returns training and test subsets that have the same proportions of class labels as the input dataset.

回答5:

Try running this code, it "just works":

from sklearn import cross_validation, datasets 

iris = datasets.load_iris()

X = iris.data[:,:2]
y = iris.target

x_train, x_test, y_train, y_test = cross_validation.train_test_split(X,y,train_size=.8, stratify=y)

y_test

array([0, 0, 0, 0, 2, 2, 1, 0, 1, 2, 2, 0, 0, 1, 0, 1, 1, 2, 1, 2, 0, 2, 2,
       1, 2, 1, 1, 0, 2, 1])