Recently on Steem I was exploring the followers of an account I really like (it posts good content and I was hoping to find similar content among its followers/followings), when I kept running into accounts that resteemed the exact same set of posts, in the exact same order.
Being a social platform nerd for my day job, a couple possibilities jumped out:
- These accounts are a voting ring, boosting each other/or are sock puppets for single entity
- These accounts are not sock puppets, but in a "white hat" resteem ring (maybe via handing credentials over to well-intentioned app that is resteeming the same posts across all accounts in same order)
I then started to wonder how Steem uses community detection.
Does it follow other commercial platforms with similar "upvoting" mechanisms (which shall remain nameless :D)? These commercial sites shadow-ignore upvotes that come from voting rings (they allow the votes, and the total votes is reflected in the tally, but they don't contribute any weight to algorithm that ranks the story for things like "hot" or "trending" or "popular" or "front page").
(Side note: even the SEC uses community detection to help track down insider trading).
But community detection isn't only used to find fraud and bad actors. It can also be used for other stuff, like recommendation systems. Say... to suggest an account to follow or a piece of content you may want to check out.
Has anybody done Steem research with this? (Official or third party/independent stuff.)
Would love to check out any links you can recommend.
If you don't have any recommendations for me, but now are curious about the field, here's a video I like about topic-based community detection on Twitter.
PS- I probably didn't make it clear, but community detection has plenty of non-online-social network application, including important physical science stuff. I just didn't bring it up coz that's not my lane.