It was found that a number of user followers on Facebook and Twitter do not follow Benford's distribution as genuine organic accounts do: the first significant digit of numbers is a “1” about 30% of the time - 6 times more often than it’s a 9. Golbeck and others identified a number of accounts that did not follow this pattern and were identified as fake.
Are you aware of other reliable detection methods?
Hm-m-m... Do they mean:
suppose I'm followed by robots, therefore my account is fake?
or what?
Hello Alexander,
Probably yes. Why would a real person's account be followed mainly by bot accounts? The assumption lends itself to further research. Alternatively one could turn the question on its head and focus on a random sample of accounts not following Benford's D to see where that leads in terms of real or fake profiles.
Best,
MM
supervised learning can be used on the text of the micropost (tweet, facebook post...) for troll/cyber-bullying detection. other features can be studied such as geolocation, date and time, language... and of course the relationship with other profiles
- Badges
- Science topic
- Similar topics
- Social Science
- Media
- Social Media