Reputation: 4820
Can we use KNNImputer
in parallel ? (i.e n_jobs=-1
like other sklearn
modules) ?
I have a dataframe
with 20 rows (features) and some of them contains nan
.
Instead of using KNNImputer
in sequential way (compute the value of each nan
in row),
can we do it in parallel ? (like n_jobs = -1
) ?
my code for the sequential way looks:
imputer = KNNImputer(n_neighbors=5, weights="uniform", copy=False)
X_new = imputer.fit_transform(X, y)
How can we change it to impute in parallel ?
Upvotes: 2
Views: 1023
Reputation: 47
you couldn't use multiprocessing with KNNImputer and you can check this thread https://github.com/scikit-learn/scikit-learn/issues/18186
Upvotes: 1