The basic idea of the content-based recommendation engine is to suggest items similar to those in which the user has preceding shown interest. The effectiveness of content-based recommendation engines is dependent on our ability to quantify the similarity of an item to others.
Let's look into the following diagram. If User 1 has read Doc 1, then we can recommend Doc 2 to the user, which is similar to Doc 1:
Now, the problem is how to determine which items are similar to each other. Let's look into a couple of methods of finding similarities between different items.
