无法从 sklearn.impute 导入 IterativeImputer

Question

I have a dataset with a column contains missing values which has relatively high correlation with another columns. I want to use IterativeImputer from sklearn.impute but I receive this error:

ImportError: cannot import name 'IterativeImputer'

I upgraded my conda, my scikitlearn package and it is not resolved yet!

Answer 1

I had the same problem even with the 0.21 version of sklearn. You need to add an other import

from sklearn.experimental import enable_iterative_imputer
from sklearn.impute import IterativeImputer

Answer 2

IterativeImputer is only available in scikit-learn 0.21, which as of April 2019 is only released as a developer version, not as stable.