|
From: uwe s. f. <uf...@ph...> - 2004-04-20 14:38:27
|
Marco,
Do I get this right? Bayes has to learn a certain number of spam *and* ham messages, before it start's working? I thought it's enough to feed the spam only. I've done this over quite some time (I'd say with at least 2000 spam messages,) but I've never got a reasonable effect. If it's true what you wrote, it would explain a lot...
Thanks,
Uwe
At 14:40 2004-04-20, Marco Maske wrote:
>Bayes must first learn enough ham/spam, bevor it is _not_ working only
>learning.
> ...
>To be accurate, the Bayes system does not activate until a certain number
>of ham (non-spam) and spam have been learned. The default is 200 of each
>ham and spam, but you can tune these up or down with....
|