On Thu, 13 Jan 2011 13:51:14 +0000 RW <rwmailli...@googlemail.com> wrote:
> Is there anything to prevent spammers signing up and using your > databases to autogenerate spam? Not really, but then we only make our database available to customers using our commercial product, so the cost would probably deter spammers. > It sounds like it may be the sort of technique that works until > spammers take it seriously. Bayes is quite resilient, actually. It would take a lot of work for spammers to craft messages that work around it. (Because of the way our Bayes calculation works, it's not enough just to add a whole lot of hammy tokens to your message.) > Training from slowly changing public corpora has no advantage to set > against the loss of local information, although it should be OK for > testing purposes. Our product also has local Bayes databases trained by each user; these local databases augment the central database. Regards, David.
