recommenders-team/recommenders
Vedi su GitHubQuestion about Wide and Deep Model for Movie Recommendation
Open
#1206 aperta il 23 set 2020
help wanted
Metriche repository
- Star
- (17.706 star)
- Metriche merge PR
- (Merge medio 6g 16h) (10 PR mergiate in 30 g)
Descrizione
Description
in this example there are some itemID(movie_id) in test data that they don't exists in train data, so during training the model can't build or it is better to say "can't update" embedding for them. in normal situation it's not a problem but "ranking_pool" that was used for evaluation also contains these itemIDs pair with some userIDs therefore they have negative affects on evaluation metric like NDCG because we multiplied some random vectors with learned user vectors. Is there any reason for doing so?