Collaborative future event recommendation

摘要: We demonstrate a method for collaborative ranking of future events. Previous work on recommender systems typically relies feedback particular item, such as movie, and generalizes this to other items or people. In contrast, we examine setting where no exists the item. Because direct does not exist events that have taken place, recommend them based individuals' preferences past events, combined collaboratively with peoples' likes dislikes. topic unseen item recommendation through user study academic (scientific) talk recommendation, aim correctly estimate function each user, predicting which talks would be most interest them. Then by decomposing parameters into shared individual dimensions, induce similarity metric between users degree they share these dimensions. show predictions are more effective than pure content-based recommendation. Finally, further reduce need explicit feedback, suggest an active learning approach eliciting incorporating available implicit cues.