Hi Maciej Yes, that *train* method is intended to be public, but it is marked as *DeveloperApi*, which means that backward compatibility is not necessarily guaranteed, and that method may change. Having said that, even APIs marked as DeveloperApi do tend to be relatively stable.
As the comment mentions: * :: DeveloperApi :: * An implementation of ALS that supports *generic ID types*, specialized for Int and Long. This is * exposed as a developer API for users who do need other ID types. But it is not recommended * because it increases the shuffle size and memory requirement during training. This *train* method is intended for the use case where user and item ids are not the default Int (e.g. String). As you can see it returns the factor RDDs directly, as opposed to an ALSModel instance, so overall it is a little less user-friendly. The *Float* ratings are to save space and make ALS more efficient overall. That will not change in 2.0+ (especially since the precision of ratings is not very important). Hope that helps. On Tue, 8 Mar 2016 at 08:20 Maciej Szymkiewicz <mszymkiew...@gmail.com> wrote: > Can I ask for a clarifications regarding ml.recommendation.ALS: > > - is train method > ( > https://github.com/apache/spark/blob/master/mllib/src/main/scala/org/apache/spark/ml/recommendation/ALS.scala#L598 > ) > intended to be be public? > - Rating class > ( > https://github.com/apache/spark/blob/master/mllib/src/main/scala/org/apache/spark/ml/recommendation/ALS.scala#L436)is > using float instead of double like its MLLib counterpart. Is it going to > be a default encoding in 2.0+? > > -- > Best, > Maciej Szymkiewicz > > >
