Hello, I just saw your message and here is my expirience with Bayes and JAMES. You need to do everything that is explained in wiki and when you finis with these settings (from wiki) when you start JAMES it will create all necessary tables for Bayes. Then, when mails start to come into your system you will have to monitor postmaster account (if you set that postmaster receive maiIs marked as spam - I strongly recomend you set this at least for a motnh) and redirect all mails that are not spam to ham. From my expirience, Bayes filter will recognize mail as ham if you redirect it few times (it will not work for just one redirection :) - I do not know why). After few months (in my case 2-3) you can stop sending spams to postmaster. I hope this helps. Cheers
On Thu, Jun 18, 2009 at 9:18 AM, SUL19 <[email protected]> wrote: > > I'm i clear? or not > i need your help please, i have 6 days to do this > > > SUL19 wrote: > > > > Hi all, > > > > I'm a student and i want to do something using the james server, i'm > using > > the version 2.3.1. First, i want to run the bayesian algorithm and the > > blacklist algorithm, from reading the wiki i knew how to run the bayesian > > but you know it needs a training, i want to train it and the training > > needs a dataset, i will try to find a dataset, but can you help me in > > finding a code that automatically trains the algorithm, that was > regarding > > to the bayesian. Regarding to the blacklist i didn't know how to run it, > > and i didn't know how to add the blocked lists. Second, after running the > > previous algorithms i want to make a comparasion between the results of > > the two algorithms. so, i need your help. please help me . > > > > regards, > > > > > > -- > View this message in context: > http://www.nabble.com/doing-a-comparasion-between-bayesian-and-blacklist-tp24071049p24087671.html > Sent from the James - Users mailing list archive at Nabble.com. > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [email protected] > For additional commands, e-mail: [email protected] > >
