You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Twitter (mostly, but not only) is full of spam accounts managed by bots/scripts. We can detect these by a set of rules:
Look at the profile picture, if it's still the default...that's a strike against the account
Look at number of followers to following ratio (in the case of twitter)
Look at what is being posted (run your typical spam filter over the message - or maybe atypical, train naive bayesian classifier for this too)
Look at the name, I've seen things like "Bot" in the name for people announcing the fact that the account is automated (not always going to be the case, but does happen)
Then put another field on the message series (and probably shared links, if not all series) that flags possible robot. We'll keep the data, but it would be nice to be able to filter out suspicious accounts for greater accuracy.
The text was updated successfully, but these errors were encountered:
Twitter (mostly, but not only) is full of spam accounts managed by bots/scripts. We can detect these by a set of rules:
Then put another field on the message series (and probably shared links, if not all series) that flags possible robot. We'll keep the data, but it would be nice to be able to filter out suspicious accounts for greater accuracy.
The text was updated successfully, but these errors were encountered: