04 April 2014 1 6K Report

I'm building a recommendation systems that suggests items to a user based on items chosen by similar users. It's similar to collaborative filtering, although I am using multiple dimensions to describe the similarity between two users (excluding the similarity of their previous choices).

I am looking for advice/references regarding how this problem has been solved in the past. Specifically, I am interested in strategies of how to combine multiple similarity features to form a single recommendation.

Similar questions and discussions