We’ve all known trolls who love to dwell in dank forum posts and shadowy comment sections, ready to spread ill will and spark flame wars just to see the world burn. But what if a tool could be devised to identify such miscreants before they could do much damage?
That is one possibility that has arisen from a new Cornell University project in which researchers studied online communities (including IGN) and created an algorithm that can predict which posters had the highest likelihood of being banned in the future. The algorithm isn’t perfect (it misclassifies one out of five users), but the team claims that it is able to spot a troll in as few as 10 posts.
According to the study, most banned accounts “began their commenting life at a lower perceived standard of literacy and/or clarity than the median for their host groups.” The project also said that communities may be partially to blame for creating an environment that could “incubate antisocial behavior.”
[Source: The Stack via Slashdot. Thanks, Hagu.]
