In order to build your recommendations, you must first create a Similarity Object. This object is used as the basis for the recommendations. The basic math works as follows
Recommend content that is similar to content that the user as liked

This means that, recommend content similar to the content that the user has liked already. Thus, the importance of a similarity object. Once this similarity object is built following those instructions, you may then configure the recommendations by visiting admin/content/recommendations. There you must select the similarity object you would like to use, the minimum predicted score. This setting will prevent the database from storing recommendations that are to small.