can anyone explain difference between Alternating Least Squares(ALS) and the recommendation systems?
it will be helpful if you give me an example.
Recommendation systems eg. Netflix movie recommendations, I assume its pretty straight forward to get. Recommendations systems can be in either of content based filtering, collaborative filtering and combination of these two.
Now, we can use matrix factorization for solve this learning problem (problem here is the recommending movies). cut short, it turns out to be an optimization problem and there are method available such as gradient decent (slow and costly in this case) and another is ALS algorithm.
So, ALS is an algorithm which you can use to solve the learning problem a.k.a build recommendation systems.
Please read here more-
Aiming to do a Recommendation system you can use ALS method - you cannot compare these things (aim & method to realize it). There're 2 main approaches for minimization task (or -fx maximization task): Direct & Iterative (ALS using factorization can be considered to be the first one - as is speedy).
for using Direct methods choice of Factorization is meaningfull
for using Iterative methods (like IRLS) it is important to select proper Preconditioning to achieve min with less quantity of iterations in Gradient Descent
can see algo in Alternating_least_squares_for_personalized_ranking
In this paper, proposed RankALS, a computationally effective approach for the direct minimization of a ranking objective function without sampling, that is able to cope with the implicit feedback case.
P.S. some dev.hint, e.g. TF top-K recommended, another py-package for Collaborative Filtering for Implicit Feedback Datasets