Find users with similar tastes and a high compatibility rating
kind of like the "neighbours" feature on last.fm, you get suggestions on who to add as a friend (people with similar tv show/movie taste to you) based on the tv shows, episodes and movies you marked as seen.
We’re in the planning stages for this. It’s a complicated calculation with such a large community, but we’ll figure it out.
where is the compatibility? does it still exist?
Maybe you should also take into account that rating a movie 1 or 4 is not REALLY VERY different, both users mean to say "this movie sucks", while rating a movie 5 or 8 is something completely different.
Hi, I see that there is a "compatibility"-rating when I open a profile now, but is there a way to list people with high compatibility?
@Justin Nemeth, if you want, I am currently working on a recommendation engine for music. Shouldn't be too difficult to port. Do you have a dataset with ~1000 users data that I can use? (Sensitive data removed of course).
All my votes. Not only would it be a great way to find new stuff to watch, but it would also be a great way to facilitate more interaction among users.
Shikhar Mall commented
Shouldn't be very difficult, you can use kNN to make this work. Only question is how good profile features are / can be made available in case of your website.
Rob Verboven commented
When you search for a user you see the compatibility percentage already, so should be possible to search for users with same taste between a certain percentage.
We could find people that has, for example, 80% or more compatibility with us.
with an option to see only weekly, monthly, yearly and all