The real words
Types of recommenders
PopularityRecommend the bestsellers. Simple, not personal.
Content-basedRecommend items with similar features to what you liked (same genre, same director).
Collaborative filteringUse other people's behaviour. User-based: find people like you. Item-based: find items liked by the same people (“customers who bought this also bought…”).
HybridA mix, which is what most real systems use.