Hybrid Recommender
A weighted blend of content-based and collaborative approaches that covers each one's weaknesses: with .
When to use
Worked example from the lecture: a content-based prediction of 2.8 and an item-CF prediction of 2.5, with weights 0.4 and 0.6, give 2.62. Raise the content-based weight for new users and the CF weight for active users. The hybrid had the best results in the Seret+ case.
See Lec 6.