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.

Hybrid recommenders and exercise

Hybrid recommenders and exercise

- [Instructor] As we've seen with Netflix, there's no need to choose a single algorithm for your recommender system. Each algorithm has its own strengths and weaknesses and you may end up with a better system by combining many algorithms together in what we call a hybrid approach. We touched on it earlier, but these hybrid or ensemble approaches can make a real difference. The winner of the Netflix prize came from a group called KorBell and they won by creating an ensemble of 107 different algorithms that work together. The top two were SVD++ and RBM, both of which ended up in use by Netflix itself. Another example of where an ensemble approach has been shown to work is with session based recommendations. Earlier, we covered using Recurrent Neural Networks or RNNs to the problem of recommending items as part of a sequence of session data. This was called groove or rec. Its results were okay, but actually not as good as a simpler K-Nearest Neighbor approach applied to the items in a…

Contents