What algorithm does "you might like" feature use?



I’m a computer science from Thailand and doing research in recommender system field.
Recently, I’m interested in twitter’s you might like feature and would like to know the algorithm that twitter uses.
Sometimes I see my friends’ tweets that they don’t like the tweet from this feature.
So I think it would be great if the algorithm is improved.
I would like to do this research in Master degree.

So anyone can help me about this? Please contact me at plafishy.t@gmail.com
Thank you in advance.


I don’t believe that we have shared information on how our algorithms are structured, sorry. You can certainly assume that we use a variety of signals about how users interact with their followers and other Tweets to drive that feature, but I can’t share specific details.