kanzi icon indicating copy to clipboard operation
kanzi copied to clipboard

Improve Recommendations

Open jingai opened this issue 7 years ago • 0 comments

Currently, recommendations are a random pick from the top-rated items in a user's library, optionally filtered on a user-specified genre. We can probably get more clever than that, though.

If it's at all possible, it'd be nice if a user requested a movie suggestion, the skill looked at the last N movies watched, and picked 'similar' movies. Probably simply picking based on the mostly commonly-watched genre is the way to go, but perhaps there are other things we can do too.

Further, it'd be nice to be able to say something along the lines of recommend a movie like ghostbusters and have it choose a random movie that is 'similar'.

Ultimately we need to really define all of this, but it's something worth discussing.

jingai avatar Sep 01 '17 20:09 jingai