News for hackers. Stuff that matters.

A comment on Botrnot: an R package for detecting Twitter bots.

Re:Botrnot: an R package for detecting Twitter bots by andrew-lucker on Wednesday March 14, 2018 @11:46PM (#16590292)
I tried putting in verified users and they were all "probably bots". By definition is that not the only type of user publicly acknowledged as "not a bot"?
[ Parent | on HN ]
Re:Botrnot: an R package for detecting Twitter bots by nl on Thursday March 15, 2018 @12:14AM (#16590369)
Yeah it's not very good. It marks the author's account[1] as "not a bot" when it appears to be just reposting Instagram posts.. ie, it is a bot.

Plus it said I was a bot (85% chance anyway), and I'm pretty sure I'm not.

[1] https://twitter.com/mkearney (I assume this is his account?)

[ Parent | on HN ]
Re:Botrnot: an R package for detecting Twitter bots by opsiprogram on Thursday March 15, 2018 @02:21AM (#16590795)
It correctly classified me (I am a bot) w/ 96% accuracy.
[ Parent | on HN ]
Re:Botrnot: an R package for detecting Twitter bots by mrkoot on Thursday March 15, 2018 @02:26AM (#16590808)
It incorrectly classified me (@mrkoot) as a bot w/.923 probability.
[ Parent | on HN ]