Closed
Description
Describe the issue linked to the documentation
By fitting a LinearRegression
the n_jobs
parameter can set the number of threads only in case of positive=True
or in case of sp.issparse(X)
. In other cases the linalg.lstsq
is called which can handle (M, K) shaped y
and there is no parameter like n_jobs
in this call.
So as i understand in the most of the usecases n_jobs
will not affect the fitting. However i cant find this in documentation.
Suggest a potential alternative/fix
I think this fact should be added to the docs. What do you think?