TypeError：__init __（）为参数'n_splits'获取了多个值

Question

I'm using SKLearn version (0.20.2) following by:

from sklearn.model_selection import StratifiedKFold


grid = GridSearchCV(
    pipeline,  # pipeline from above
    params,  # parameters to tune via cross validation
    refit=True,  # fit using all available data at the end, on the best found param combination
    scoring='accuracy',  # what score are we optimizing?
    cv=StratifiedKFold(label_train, n_splits=5),  # what type of cross validation to use
)

But i don't understand why i will get this error:

---

TypeError                                 Traceback (most recent call last)
<ipython-input-26-03a56044cb82> in <module>()
     10     refit=True,  # fit using all available data at the end, on the best found param combination
     11     scoring='accuracy',  # what score are we optimizing?
---> 12     cv=StratifiedKFold(label_train, n_splits=5),  # what type of cross validation to use
     13 )

TypeError: __init__() got multiple values for argument 'n_splits'

Im already tried n_fold but come with the same error result. And also tired to update my scikit version and my conda. Any idea to fix this ? Thanks a lot!

Answer 1

StratifiedKFold takes exactly 3 arguments when initialized, none of which are the training data:

StratifiedKFold(n_splits='warn', shuffle=False, random_state=None)

So when you call StratifiedKFold(label_train, n_splits=5) it thinks you passed n_splits twice.

Instead, create the object, then use the methods as described in the example on the sklearn docs page for using the object to split your data:

get_n_splits([X, y, groups]) Returns the number of splitting iterations in the cross-validator split(X, y[, groups]) Generate indices to split data into training and test set.

Answer 2

StratifiedKFold takes three arguments but you are passing two arguments. See more in sklearn documentation

Create StratifiedKFold object and pass it to GridSearchCV as below.

skf = StratifiedKFold(n_splits=5)
skf.get_n_splits(X_train, Y_train)

grid = GridSearchCV(
pipeline,  # pipeline from above
params,  # parameters to tune via cross validation
refit=True,  # fit using all available data at the end, on the best found param combination
scoring='accuracy',  # what score are we optimizing?
cv=skf,  # what type of cross validation to use
)