data mining - How is frequent itemsets compared with item-based collaborative filtering in recommender systems? -
what difference between data mining approaches: frequent itemsets , item-based collaborative filtering in area of recommender systems?
frequent itemsets use unary data: know person has items in basket have no idea if knows other items in universe exist. is, not have item z in basket because knows , doesn't it, or because doesn't know (and might have included if did).
item-based collaborative filtering, on other hand, includes rating data. e.g., star ratings on scale of 0.5 5.0 (in case of movielens). here can better calculate similarity between items because have both positive , negative opinions, , can make more informed guess if user doesn't rate item, doesn't know exists.
Comments
Post a Comment