This adds an alpha parameter to the ALS models - which is the weight each
positive example is multiplied by. While we could previously achieve the
same result before by passsing a sparse matrix to model.fit where the weight
was already multiplied in, having as an explicit parameter makes it easier
to work with hyper parameter tuning frameworks - letting you treat the same
as things like regularization/epochs etc
This adds an alpha parameter to the ALS models - which is the weight each positive example is multiplied by. While we could previously achieve the same result before by passsing a sparse matrix to model.fit where the weight was already multiplied in, having as an explicit parameter makes it easier to work with hyper parameter tuning frameworks - letting you treat the same as things like regularization/epochs etc