uncorfimed spam

Hi, Robert.


Robert LeBlanc wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Janky Jay, III wrote:
>
>
>> Federico Nan wrote:
>>
>>> hi! its my first post, so i dont now if this post is a repost. =)
>>>
>>> Is there currently any way to have emails that are not spam to not appear in the Unconfirmed non-spam cache at all?
>>>
>>     Sure. Just change the 'Should non-spam items be cached?9; setting to
>> 'No9; for any domain you don't want to cache non-spam items for.
>>
>
> While this is correct, please be sure that you understand the
> consequences of doing so.  Usually when people ask about doing this it
> is because they misunderstand the purpose of the unconfirmed non-spam
> cache--they see it as an unnecessary nuisance without recognizing the
> useful purpose it serves.
>
> The unconfirmed non-spam cache is vital for training your Bayes
> database, since without it the Bayes engine will not see examples of
> what non-spam looks like. ; By confirming both spam and non-spam, you
> help your Bayes engine learn much faster and achieve higher rates of
> effectiveness.  If you only show it what spam looks like, the learning
> process will suffer--you'll see more false positives/negatives, and
> learning will take considerably longer.
>
> Maia provides the two spam-related caches (unconfirmed spam and
> unconfirmed non-spam) to let you correct the Bayes learning engine when
> it makes mistakes, and to let you confirm that it learned correctly the
> rest of the time. ; The spam cache lets you tell the Bayes engine about
> false positives--good mail that should not have been blocked; the
> non-spam cache lets you tell the Bayes engine about false
> negatives--spam that got through the filter.  If you disable the
> non-spam cache, you lose the ability to report false negatives, so your
> false negative problems will only get worse over time, since the Bayes
> engine never learns about its mistake (and therefore never corrects itself).
>
> Disable the non-spam cache if you like, in other words, but be aware
> that in doing so you will also be compromising the effectiveness of your
> Bayes database.
>
>


    Indeed. Sorry. I should have provided a bit more information about
it. I was just in a bit of a hurry. But, Robert saved the day. Thanks!


> - --
> Robert LeBlanc < rjlrenaissoft.com" target="_blank" onclick="return top.js.OpenExtLink(window,event,this)">rjlrenaissoft.com>
> Renaissoft, Inc.
> Maia Mailguard <http://www.maiamailguard.com/&gt;
>
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.4.7 (Darwin)
> Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org
>
> iD8DBQFH6t0UGmqOER2NHewRAmeRAKCBboGQR1w5+xUXDUvFDGiGJtx8rQCfcnGL
> /DN4QKNfWacxA9dklZO1470=
> =EvjS
> -----END PGP SIGNATURE-----
> ______________________________________________
>
_______________________________________________
Maia-users mailing list
Maia-usersrenaissoft.com" target="_blank" onclick="return top.js.OpenExtLink(window,event,this)">Maia-usersrenaissoft.com
http://www.renaissoft.com/mailman/listinfo/maia-users