python sklearn features selection mutual_info_regression

Question

In mutual_info_regression pre-installed n_neighbors=3 This code works for n_neighbors=3:

selector = SelectKBest(mutual_info_regression, k='all').fit(X, y)

That request for n_neighbors=2 in mutual_info_regression?

Don't work variants:

selector = SelectKBest(mutual_info_regression, k='all').fit(X, y,**{'n_neighbors':2})
selector = SelectKBest(mutual_info_regression(**{'n_neighbors':2}), k='all').fit(X, y)
selector = SelectKBest(mutual_info_regression(n_neighbors=2), k='all').fit(X, y)
selector = SelectKBest(mutual_info_regression,n_neighbors=2, k='all').fit(X, y)

scoring = make_scorer(mutual_info_regression, greater_is_better=True, n_neighbors = 2)

selector = SelectKBest(scoring, k='all').fit(feat, targ)

Answer 1

You can use pythons partial function to create a scorer with non-default values:

from functools import partial
scorer_function = partial(mutual_info_regression, n_neighbors=2)
selector = SelectKBest(scorer_function, k='all').fit(X, y)