Well, candidate shuffling is something I just made up – although we will in fact build recommender systems that just recommend random, shuffled stuff as a baseline to compare against. Everything else is something we have to do in a top-N recommender – we need to generate recommendation candidates based on the user’s interests, we need to rank those candidates, and then filter them down to the N results we want.

And that’s it for this quiz, and for this initial section of the course! You’ve already learned a lot about the concepts, challenges, and architecture of recommender systems. Keep on going, as there’s a lot more detail and hands-on practice ahead for you.