News for hackers. Stuff that matters.

A comment on YouTube monetization analysis. Parent comment

Re:YouTube monetization analysis by syrrim on Thursday November 30, 2017 @04:14PM (#15819603)
Another hypothesis I've heard that would explain the same effect is that youtube groups uploaders with the preferences of their viewers. If someone has lots of viewers who also watch videos of undesirable topics, that uploader will be considered undesirable as well.
[ Parent | on HN ]
Re:YouTube monetization analysis by dopamean on Thursday November 30, 2017 @04:39PM (#15819798)
Is there a name for this type of "guilty by association" algorithm?
[ Parent | on HN ]
Re:YouTube monetization analysis by rainbowmverse on Thursday November 30, 2017 @06:49PM (#15820746)
K-means clustering.
[ Parent | on HN ]
Re:YouTube monetization analysis by loeber on Thursday November 30, 2017 @10:26PM (#15821677)
This is one popular clustering algorithm. I think "clustering algorithms in general" would be a more accurate answer.
[ Parent | on HN ]
Re:YouTube monetization analysis by rainbowmverse on Saturday December 02, 2017 @07:00AM (#15831538)
I didn't know there were others. I heard about this from a journal entry somewhere on DeviantArt about 10 years ago.
[ Parent | on HN ]
Re:YouTube monetization analysis by londons_explore on Thursday November 30, 2017 @05:51PM (#15820390)
The algorithm does not have viewer's other watched videos as an input. Perhaps it should, but it does not.
[ Parent | on HN ]