From the course: Building Recommender Systems with Machine Learning and AI

Unlock this course with a free trial

Join today to access over 22,400 courses taught by industry experts.

Similarity metrics

Similarity metrics

data sparsity out of the way, let's talk about some other ways to compute similarity. One is the adjusted cosine metric. And it's applicable mostly to measuring the similarity between users based on their ratings. It's based on the idea that different people might have different base lines that they are working from. What Bob considers a three star movie may be different While Alice tries to be nice and rates things five stars unless she really didn't like them. This is a real effect you'll see not just across different individuals but across different cultures too. Some countries are more brutal with their ratings than others. So adjusted cosine attempts to normalize these differences. Instead of measuring similarities between people based on their raw rating values, we instead measure similarity based on the difference between a user's rating for an item and their average rating for all items. So if you look at this equation, we've replaced X with X sub I minus X bar. And replaced Y…

Contents