Anti Spam-Bot Bot
Posted: Wed Jun 27, 2007 8:40 pm
A couple of weeks ago I finished my A-Level exams and since then I've been attempting to get some extensive coding done in PB - but for some reason I've been unable to get enthused by a project. Anyway, I was searching for a problem which I could actually solved and then I thought about the (occasional) Spam Bots we get here.
I've managed to write a Python script which can read the forums and get information from it (topic names, poster, replies...). I hope to be able to use this to combat spam and perhaps even abusive users.
As for spam bots they're usually quite easy to spot. Their first post will be full of links and/or certain keywords - so finding them will be easy. I'm less sure of how to detect abusive users (basic things like excessive swearing/large image posting is, again, very easy) or what to do once my bot encounters either.
I'd quite like to get this bot an account and allow it to have a certian level of interaction with the forums i.e. deleting suspicious posts and then PMing the poster with an explaination just incase. However, I know that giving a bot such privilidges may not be possible or practical.
Even so I'm still interested in approches to the problems I've outlined. Any suggestions?
I've managed to write a Python script which can read the forums and get information from it (topic names, poster, replies...). I hope to be able to use this to combat spam and perhaps even abusive users.
As for spam bots they're usually quite easy to spot. Their first post will be full of links and/or certain keywords - so finding them will be easy. I'm less sure of how to detect abusive users (basic things like excessive swearing/large image posting is, again, very easy) or what to do once my bot encounters either.
I'd quite like to get this bot an account and allow it to have a certian level of interaction with the forums i.e. deleting suspicious posts and then PMing the poster with an explaination just incase. However, I know that giving a bot such privilidges may not be possible or practical.
Even so I'm still interested in approches to the problems I've outlined. Any suggestions?