recommender_pytorch icon indicating copy to clipboard operation
recommender_pytorch copied to clipboard

why convert to implicit scenario

Open stellaywu opened this issue 6 years ago • 2 comments

wondering why the ratings are all set to 1 in the dataset as implicit scenario. Thanks!

stellaywu avatar Oct 21 '19 15:10 stellaywu

In the implicit scenario, you are just deciding weather to recommend a movie or not, it becomes a binary problem. Hence you just use 0 or 1, instead of fine granular rating. This is mostly picked up from the literature on recommender systems, and is not something new for this model.

HarshdeepGupta avatar Oct 25 '19 06:10 HarshdeepGupta

Thank you! Did not realize it was trying to simplify the number of classes.

stellaywu avatar Oct 28 '19 02:10 stellaywu