mahout 读书

People tend to like things that are similar to other things they like.

Mahout contains a recommender engine – several types, in fact, beginning with conventional user-based
and item-based recommenders. It also includes implementations based on “slope-one” techniques, a
new and efficient approach. You will also find experimental, preliminary implementations based on the
singular value decomposition (SVD) and more.

 This takes the form of “preferences” in Mahout-speak.

 A preference consists of a user ID and an item
ID, and usually a number expressing the strength of the user's preference for the item. IDs in Mahout
are always numbers, integers in fact. The preference value could be anything, as long as larger values
mean stronger positive preferences. For instance, these values might be ratings on a scale of 1 to 5,
where the user has assigned “1” to items she can't stand, and “5” to her favorites.

Mahout doesn’t literally require that the users and items be
named with numbers!