User talk:ClueBot Commons
This page is for comments on the bot or questions about the bot.
Report false positives on the false positives page, not here.
Praise should go on the praise page. Barnstars and other awards should go on the awards page.
Use the "new section" button at the top of this page to add a new section. Use the [edit] link above each header to edit that header.
ClueBot's owner or someone else who knows the answer to your question will reply on this page unless you request otherwise.
 | If you are here for a question about ClueBot III, please see this notice. |
| ClueBots | |
|---|---|
| ClueBot NG/Anti-vandalism · ClueBot II/ClueBot Script |  |
| ClueBot III/Archive · Talk page for all ClueBots |
|
Wrong username reported in edit summary
Here are two entries from ClueBot NG's user contributions:
- (del/undel) 20:04, 4 May 2011 (diff | hist) N User talk:Monkeynutt$32 (Warning Monkeynutt$32 - #1)
- (del/undel) 20:04, 4 May 2011 (diff | hist) m Aberdeen (Reverting possible vandalism by Monkeynutt4272406582 to version by 2.102.76.190. False positive? Report it. Thanks, ClueBot NG. (388897) (Bot))
Looks as if $32 at the end of the username may have been expanded into 4272406582 ? Hope this helps. —SMALLJIM 19:53, 4 May 2011 (UTC)
- I'm looking into this at the moment - the bot should handle usernames as a string so nothing about it should change. DamianZaremba (talk • contribs) 20:07, 4 May 2011 (UTC)
- Thanks for the quick reply. I hope it's fixable, or at least just a cosmetic issue. It's good to have CBNG running again, BTW. —SMALLJIM 22:41, 4 May 2011 (UTC)
CBNG as a vandalism firewall
On a completely different tack, I was wondering if CBNG could eventually be tacked on as a sort of firewall to MediaWiki so that vandalism could be rejected before it reached the database. There'd have to be some sort of "edit refused" message returned to the vandal, and a process for dealing with false positives, of course, but wouldn't it save a lot of hassle! —SMALLJIM 22:41, 4 May 2011 (UTC)
- That thing about the edit refusing should not be done, because CBNG still has false positives. How would they be dealt with? --43?9enter ★☭ 06:33, 5 May 2011 (UTC)
- Firstly, I'm not talking about doing this now - it's known that the ANN in CBNG will continue to improve as it gets further training, and the threshold can be set high enough to make false positives (fp) arbitrarily rare; the stats say it catches around 40% of vandalism now, with an fp rate of below 0.1%. So one way of working would be to set a very well trained CBNG firewall to silently reject just the most obvious vandalism, with an vanishingly small fp rate. In this case, the existing CBNG would still run as it does now, probably using the same dataset as the firewall but with a lower threshold. Alternatively, if the firewall replaced the existing bot - with an attendant higher fp rate - there would seem to be several possibilities, but I think the basic principle would be to replace vandalism edits with messages to the user's own talk page along the lines of the ones CBNG posts now (reworded appropriately, of course). Maybe only one such message per day would be issued to avoid giving positive feedback to the vandal. —SMALLJIM 10:58, 5 May 2011 (UTC)
- But still, this idea is interesting. How would recognition be obtained if a vandalism edit doesn't even get through? It would just be cluttering the talk page with "We rejected your vandalism edits, the ones which never went live" repeatedly. In fact, it would only show that the vandal couldn't get past the hypothetical CBNG filters. --43?9enter ★☭ 06:40, 7 May 2011 (UTC)
- Firstly, I'm not talking about doing this now - it's known that the ANN in CBNG will continue to improve as it gets further training, and the threshold can be set high enough to make false positives (fp) arbitrarily rare; the stats say it catches around 40% of vandalism now, with an fp rate of below 0.1%. So one way of working would be to set a very well trained CBNG firewall to silently reject just the most obvious vandalism, with an vanishingly small fp rate. In this case, the existing CBNG would still run as it does now, probably using the same dataset as the firewall but with a lower threshold. Alternatively, if the firewall replaced the existing bot - with an attendant higher fp rate - there would seem to be several possibilities, but I think the basic principle would be to replace vandalism edits with messages to the user's own talk page along the lines of the ones CBNG posts now (reworded appropriately, of course). Maybe only one such message per day would be issued to avoid giving positive feedback to the vandal. —SMALLJIM 10:58, 5 May 2011 (UTC)
Cluebot NG API (or Logs)
Hi, this is a copy of my message already posted on Crispy1989's user page, but I may rather get an answer here.
I hope you can help me out: I'm a PhD student doing research on Wikipedia, and I wonder if there is also an API for Cluebot NG that you can send a RevID to and get back the vandalism score (or -if not available - maybe just a log of the vandalism scores already computed in the past). I couldn't find anything in that regard on the user:Cluebot NG page or elsewhere. The background is that for my research, I analyze a local dump of the English Wikipedia and I need to filter out revisions that are most probably vandalism (and need to do so also for the past time when Cluebot NG was not active yet and for revisions where he wasn't fast enough and got 'beat' by others ). And Cluebot NG seems to do the best job so far marking such revisions with high (and proven) accuracy. I would set up ClueBot NG myself on my machine to let it go trough the dumps. But this seems fairly complicated and I don't have the proficiency to do that. So it would be nice if you could give me some hints. 129.13.72.198 (talk) 12:38, 6 May 2011 (UTC) --> main account: Fmoo (talk) 13:05, 6 May 2011 (UTC)
- There is no API for checking edit scores against - it is something I am looking into but need to talk to Cobi about. The easiest way to get edit scores is to put a bot in #cluebotng-spam and #wikipedia-van on irc.cluenet.org - these channels are where reverted and not reverted edit info gets thrown. As for existing scores please see the list here (I just dumped them out from logs). DamianZaremba (talk • contribs) 13:15, 6 May 2011 (UTC)
- On a side note that file is around 100M so don't open it in a browser ;) DamianZaremba (talk • contribs) 13:23, 6 May 2011 (UTC)
unfortunately my son has taken to not believing cluebot not because he doesnt like them but the stick from other kids who tell him he is gay as they are telling him it stands for gay and proud kids are cruel and no matter what you are when they say things like this to you you do anything to avoid it so part of the problem i feel is that as soon as kids start something like this it spirals. I had never heard the expression before and still try to encourage him to wear what he wants and not because of what people say to him but its hard when they are at the growing up stage and want to fit in and not stand out. I'm proud of my kids no matter who or what they are but will listen to them when they make choices about what makes them happy and comfortable in their environment