New answers tagged

1

This is referred to as side information. This is used to enhance the recommender system. A good library for collaborative filtering (and beginner friendly) is turicreate. Have a look at this link. o summarise, the traditional, basic matrix factorisation will encode user i and items j respectively as vectors $u_i$ and $v_j$ so that the predicted score that a ...


0

Clearly defining the purpose of your project will determine the proper method to be used. If you have user playlists or listening histories you can use a clustering algorithm to group these users into categories based on the similarity of their prefer songs' sonic characteristics. That way, you can analyze the music that a future user has on their account ...


1

The nDCG depends on the relevance of each document as you can see on the Wikipedia definition. I guess you could use 0 and 1 as relevance scores, but then all relevant documents would have the same score of 1, and then it wouldn't make much sense to apply the nDCG penalty discounts. A similar measure often used with binary relevance scores is the mean ...


1

Is there a reason you cannot filter NSFW articles as a post-processing step? For instance, if you need to recommend 10 articles, you find the top 50 recommendations and then apply the filter if it is needed and truncate down to top 10? In my experience, it is pretty common to apply business rules before/after recommendations to account for things like this. ...


Top 50 recent answers are included