AI engineering glossary · Time series recsys
What is hybrid recommender?
Hybrid recommender: A recommender that combines several approaches, typically collaborative and content-based, to offset the weaknesses of each. Combination can be by blending scores, switching by situation, or feeding both signals to one model.
Hybrid recommender: an example
Use content features for brand-new items and collaborative signals once enough interactions exist.
Related terms
More Time series recsys terms
- Intermittent demand
- Inverse propensity scoring
- Item-based collaborative filtering
- Kalman filter
- KPSS test
- LambdaMART
- Latent factors
- Learning to rank
Where you learn it
Hybrid recommender is part of the AI Engineering Program, which teaches it with notes, a hands-on lab and an exam. The first lesson is free.