21.1 Introduction

Content filtering is another way to generate recommendations that works quite differently from collaborative filtering. But it can be very effective in situations where you don't have user-item-ratings data yet. For example, if you have just built an app, and you don't have ratings data yet to use for collaborative filtering, you can use the details about users or the details about items to group or cluster similar records together in order to recommend related items or related users. This process is fundamentally different, but provides the same results--recommendations. Let's learn how this process works.