bill parducci wrote: > > > http://spambayes.org > > it is the best thing out there'. there are a number of ways > to implement > it. i currently get high 99% accuracy with it myself (and i > am am lazy > with my training).
Yes, its very good as it knows what it doesn't know (to twist a current phrase) and allows you to educate it. Most other Bayesian filters are too absolute and do not deal effectively with the grey area of possible spam. We use a sparse whitelist and selected blacklists within James and SpamBayes on the MUAs (clients). The blacklists prevent most mass spam. SpamBayes allows each user to define as Spam anything that gets through the blacklists and anything they simply do not want to deal with. Hmm. The latter seems to include anything to do with administration, such as progress reports. Also, for anyone wanting to try server side Bayesian filtering within James there is the Praxis Bayesian filter mailet - http://portale.praxis.it/pub/james/readme.htm. -- Steve --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
