Make a Recommender System with Collaborative Filtering Techniques

On this journey, I want to share how to make a recommender system using Collaborative Filtering techniques. Don’t be confused if you don’t know what is recommender system. I will explain a bit about it. So, recommender system is a system that predicted user preferences on a list of items based on your interaction with other item or user. This system try to recommend you item or product that you might be interested. Have you used Netflix, Youtube or Spotify? If yes, then you already using recommender system on your daily. Netflix’s algorithm give you some of recommended movie you like based on what kind of movies you watch. It can save your time to not looking all of the movies in Netflix and also it gives you new experience to trying new recommendation from them.

Maybe some of you already know that is also other techniques you can use to build your recommender system, like Content-based Filtering. But for now, let us focus on Collaborative Filtering. Collaborative filtering works by filter the items and give recommendation based on similarity to the other user. If your interest is similar with bunch of users, then maybe you also like the items that other user consumed.

Source : https://miro.medium.com/max/2656/1*6_NlX6CJYhtxzRM-t6ywkQ.png

In this tutorial, I will use Matrix Factorization with Global Average calculation for give top-5 item to each user. You can read more from : http://www.quuxlabs.com/blog/2010/09/matrix-factorization-a-simple-tutorial-and-implementation-in-python/

So, enjoy the video!