Submission + - Flushing out suspicious social media activity using Benford's Law (technologyreview.com)
Taco Cowboy writes: Benford's Law (See
http://en.wikipedia.org/wiki/B...
) refers to the frequency distribution of digits in many types of real-life data in which the digit "1" occurs as the leading digit about 30% of the time
Jennifer Golbeck at the University of Maryland in College Park applied Benford's Law on data of users from five major social networks, and in tweeter only 170 people out of the 21,000 that she investigated had a correlation lower than 0.5
When she investigated further, she discovered that only 2 accounts out of the 170 seem to belong to legitimate users, with the rest (168) are part of a Russian botnet
That’s interesting work that has important implications for social network forensics. In recent years, it has become increasingly difficult to spot accounts on social networks that are engaged in suspicious activity. Comparing a large number of these against Benford’s law is a quick and simple way to find ones that require further investigation
Of course, this process will not find all suspicious accounts. Any account that grows in the same way as a conventional one would remain hidden and it’s possible that maleficent users could employ simple techniques to make their accounts less identifiable now that this method has been revealed
But for the time being, Benford’s law looks to be a valuable tool in the war against fraud and suspicious activity on social networks. “The applicability of Benford’s Law to social media is a new tool for analyzing user behavior, understanding when and why natural deviations may occur, and ultimately detecting when abnormal forces are at work,” concludes Ms. Golbeck
Jennifer Golbeck at the University of Maryland in College Park applied Benford's Law on data of users from five major social networks, and in tweeter only 170 people out of the 21,000 that she investigated had a correlation lower than 0.5
When she investigated further, she discovered that only 2 accounts out of the 170 seem to belong to legitimate users, with the rest (168) are part of a Russian botnet
That’s interesting work that has important implications for social network forensics. In recent years, it has become increasingly difficult to spot accounts on social networks that are engaged in suspicious activity. Comparing a large number of these against Benford’s law is a quick and simple way to find ones that require further investigation
Of course, this process will not find all suspicious accounts. Any account that grows in the same way as a conventional one would remain hidden and it’s possible that maleficent users could employ simple techniques to make their accounts less identifiable now that this method has been revealed
But for the time being, Benford’s law looks to be a valuable tool in the war against fraud and suspicious activity on social networks. “The applicability of Benford’s Law to social media is a new tool for analyzing user behavior, understanding when and why natural deviations may occur, and ultimately detecting when abnormal forces are at work,” concludes Ms. Golbeck
