Alright, let's talk about how item-based collaborative filtering works. It's very similar to user-based collaborative filtering, but instead of users, we're looking at items.

So, let's go back to the example of movie recommendations. The first thing we would do is find every pair of movies that is watched by the same person. So, we go through and find every movie that was watched by identical people, and then we measure the similarity of all those people who viewed that movie to each other. So, by this means we can compute similarities between two different movies, based on the ratings of the people who watched both of those movies.

So, let's presume I have a movie pair, okay? Maybe Star Wars and The Empire Strikes Back. I find a list of everyone who watched both of those movies, then I compare their ratings to each other, and if they're similar then I can say these two movies are similar, because they were rated similarly by people who watched both of them. That's the general idea here. That's one way to do it, there's more than one way to do it!

And then I can just sort everything by the movie, and then by the similarity strength of all the similar movies to it, and there's my results for people who liked also liked, or people who rated this highly also rated this highly and so on and so forth. And like I said, that's just one way of doing it.

That's step one of item-based collaborative filtering-first I find relationships between movies based on the relationships of the people who watched every given pair of movies. It'll make more sense when we go through the following example:

For example, let's say that our nice young lady in the preceding image watched Star Wars and The Empire Strikes Back and liked both of them, so rated them both five stars or something. Now, along comes Mr. Edgy Mohawk Man who also watched Star Wars and The Empire Strikes Back and also liked both of them. So, at this point we can say there's a relationship, there is a similarity between Star Wars and The Empire Strikes Back based on these two users who liked both movies.

What we're going to do is look at each pair of movies. We have a pair of Star Wars and Empire Strikes Back, and then we look at all the users that watched both of them, which are these two guys, and if they both liked them, then we can say that they're similar to each other. Or, if they both disliked them we can also say they're similar to each other, right? So, we're just looking at the similarity score of these two users' behavior related to these two movies in this movie pair.

So, along comes Mr. Moustachy Lumberjack Hipster Man and he watches The Empire Strikes Back and he lives in some strange world where he watched The Empire Strikes Back, but had no idea that Star Wars the first movie existed.

Well that's fine, we computed a relationship between The Empire Strikes Back and Star Wars based on the behavior of these two people, so we know that these two movies are similar to each other. So, given that Mr. Hipster Man liked The Empire Strikes Back, we can say with good confidence that he would also like Star Wars, and we can then recommend that back to him as his top movie recommendation. Something like the following illustration:

You can see that you end up with very similar results in the end, but we've kind of flipped the whole thing on its head. So, instead of focusing the system on relationships between people, we're focusing them on relationships between items, and those relationships are still based on the aggregate behavior of all the people that watch them. But fundamentally, we're looking at relationships between items and not relationships between people. Got it?