Comments on: Bayesian blocking software http://www.kryogenix.org/days/2003/08/24/bayesian scratched tallies on the prison wall Thu, 04 Dec 2008 00:13:55 +0000 http://wordpress.org/?v=2.6.5 By: Eugene http://www.kryogenix.org/days/2003/08/24/bayesian#comment-2301 Eugene Thu, 01 Jan 1970 01:00:00 +0000 http://www.kryogenix.org/adpb/2003/08/24/bayesian/#comment-2301 <p>Classyfying web pages using Bayesian filtering is <em>hard</em>—it’s very different from classyfying email. I’ve tried it. You need lots and lots of sample pages. But otherwise, yeah, Bayesian filtering is quite effective once you’ve trained it.</p> Classyfying web pages using Bayesian filtering is hard—it’s very different from classyfying email. I’ve tried it. You need lots and lots of sample pages. But otherwise, yeah, Bayesian filtering is quite effective once you’ve trained it.

]]> By: Dave Farquhar http://www.kryogenix.org/days/2003/08/24/bayesian#comment-2302 Dave Farquhar Thu, 01 Jan 1970 01:00:00 +0000 http://www.kryogenix.org/adpb/2003/08/24/bayesian/#comment-2302 <p>I’ve often wondered the same thing. I know it won’t be perfect, but before spammers got hold of Bayesian filters, false positives were very rare. I’ll bet it’d be more effective than the blockers we have now.</p><p>I think it would have ship with a database of keywords though; I don’t think most workplaces would be too keen on someone training it manually. :)</p> I’ve often wondered the same thing. I know it won’t be perfect, but before spammers got hold of Bayesian filters, false positives were very rare. I’ll bet it’d be more effective than the blockers we have now.

I think it would have ship with a database of keywords though; I don’t think most workplaces would be too keen on someone training it manually. :)

]]>