You can tune alpha like any other hyperparam, and measuring whatever metric makes most sense -- AUC, etc. I don't think there's a general guidelines that's more specific than that. I also have not applied this to document retrieval / recommendation before
I don't think you need to modify counts or ratings, and shouldn't, since the formulation is already trying to take care of translating counts into weights as 1 + alpha * r. On Sun, Jul 26, 2015 at 9:35 AM, Debasish Das <debasish.da...@gmail.com> wrote: > In your experience with using implicit factorization for document > clustering, how did you tune alpha ? Using perplexity measures or just > something simple like 1 + rating since the ratings are always positive in > this case.... > --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@spark.apache.org For additional commands, e-mail: dev-h...@spark.apache.org
