zlacker

>>sharps+(OP)
I wonder if trolls can be categorized automatically. Caveats and all, trolls are characterized by their participation in negative-karma two-person conversations and down-voting of their comments by a diverse and changing set of users. A simple learning algorithm should zap the predictable ones. Trolls are nourished by attention and early detection and removal should nip that behavior in the bud.

>>Chaita+72
I think the hardest problem with that approach would be avoiding a false-positive of detecting people who hold controversial opinions as trolls.

The moderation graph between users and comments would be really interesting. There's a lot of information in there to be mined.