Frank Kane's Taming Big Data with Apache Spark and Python

How does item-based collaborative filtering work?

Well, here's how we're going to do it; remember, this is only one way to do it, there are many techniques, this is just one which is easier to explain. What we do is we find every pair of movies that were watched by the same person. Then we find all the users that watched that pair of movies, and compare how similar their ratings were for each one. So the idea is that if you have two movies that were watched by a bunch of people and they all rated them similarly, those movies are probably similar to each other in some way, shape, or form. After we do that, we then compute the similarity of each movie pair, sort them by the movie and then by the strength of similarities. Then you can slice ...