Re: Twitter
[email protected] Tue, 02 Oct 2012 10:35:13 -0400
| Newsgroups | gmane.mail.spam.crm114 |
|---|---|
| Message-ID | <[email protected]> |
Chris Babcock <[email protected]> writes: > Here... this is better: > > http://pastebin.com/wmdq9bLW > < sorry for the delay. Root canal failure. > Hmmm... looks not bad. Possible double-training, which might be harmful in the long term. But unnecessarily convoluted... I think you can just do the classify, and then test and act soley on the :stats: result. (to do that, you have to wrap the classify in {} to suppress any fail-skip actions). So, change your pastebin of: # SSTTT { # Are we following user? match [:bio:] <nomultiline> /Do you follow this user?.+true/ { # Should test as ham { classify <hyperspace> [:tweets:] ( ham | spam ) (:stats:) match (:: :pR:) [:stats:] /Best match.*: (.*) \nTotal/ } alius { learn <hyperspace> [:bio:] ( ham ) goto /:NEXT:/ } eval /:@::*:pR: < 0.5:/ learn <hyperspace> [:bio:] ( ham ) } to just: # SSTTT { # Are we following user? match [:bio:] <nomultiline> /Do you follow this user?.+true/ { # Find out just how hammy we are classify <hyperspace> [:tweets:] ( ham | spam ) (:stats:) } match (:: :pR:) [:stats:] /Best match.*: (.*) \nTotal/ { eval /:@::*:pR: < 0.5:/ # Are we hammy enough? learn <hyperspace> [:bio:] ( ham ) } goto /:NEXT:/ } ... and a similar stanza with "ham"<->"spam". You might also want to NOT hardcode the pR threshold for SSTTT to 0.5; that value is just a good starting point for the experiments I ran, which are not tweets but much longer texts. Your mileage will almost certainly vary. - Bill ------------------------------------------------------------------------------ Don't let slow site performance ruin your business. Deploy New Relic APM Deploy New Relic app performance management and know exactly what is happening inside your Ruby, Python, PHP, Java, and .NET app Try New Relic at no cost today and get our sweet Data Nerd shirt too! http://p.sf.net/sfu/newrelic-dev2dev