collaborative filtering - What are some ways for a recommendation engine to deal with one time, novel and potentially important content? -


say built recommendation engine recommend live tv shows watch. regular shows, pretty job using collaborative filtering , like. 1969 moon landing. it's important event, want recommendation engine handle case. can't rely on past behavior since value of recommendation drops 0 once show over.

what effective methods deal problem in recommendation space?

the problem in cf opposite: new items no clicks / ratings yet can't recommended cf algorithm , have trouble getting in front of users. old, famous item ought recommendable.

there's opposite problem: recommender system algorithms tend favor famous items knows rather more long-tail, lesser-known items may better recommendations in sense.

sounds have notion item extra-good in sense. that's side information include crudely boosting estimated rating value amount. think effective approach that.


Comments

Popular posts from this blog

unicode - Are email addresses allowed to contain non-alphanumeric characters? -

C#: Application without a window or taskbar item (background app) that can still use Console.WriteLine() -

c++ - Convert big endian to little endian when reading from a binary file -