The recommendations from collaborative filtering are based on the analysis of the historical buying patterns of users. The basic assumption is that if two users show interest in mostly the same items, we can classify both users as similar. In other words, we can assume the following:

• If the overlap in the buying history of two users exceeds a threshold, we can classify them as similar users.
• Looking at the history of similar users, the items that do not overlap in the buying history become the basis of future recommendations through collaborative filtering.  

For example, let's look at a specific example. We have two users, Mike and Elena, as shown in the following diagram:

Note the following:

• Both Mike and Elena have shown interest in exactly the same documents, Doc1 and Doc2. 
• Based on their similar historical patterns, we can classify both of them as similar users.
• If Elena now reads Doc3, then we can suggest Doc3 to Mike as well.

Let's assume that Elena and Mike showed interest in Doc1, which was about photography (because they share a love of photography). Also, both of them showed interest in Doc2, which was about cloud computing, again, because both of them have an interest in the subject. Based on collaborative filtering, we classified them as similar users. Now Elena starts reading Doc3, which is a magazine on women's fashion. If we follow the collaborative filtering algorithm, we will suggest Mike read it, who may not have much interest in it.

Note that the collaborative filtering algorithm does not depend on any other information and is a standalone algorithm, based on the changing behaviors of users and collaborative recommendations.