autolearn=ham

[Robert Swan <rs...@nskinc.com>]

How do I clear, or unlearn the bayes filter it seems that it is picking
up wrong. E-mail that is SPAM has autolearn=ham in the header and this
is wrong.

 

I am Running SPAMASSASSIN 3.0.3 on a Linux Red Hat 9 server. (just
upgraded) did this in version 3.0.2 also, unrelated I know.

 

 

 

Thanks in advance,

 

Robert

 

 

 

 

 

 

Peace he would say instead of goodbye....peace my brother.

 

Re: autolearn=ham

[James R <ja...@trusswood.dyndns.org>]

Robert Swan wrote:
> How do I clear, or unlearn the bayes filter it seems that it is picking 
> up wrong. E-mail that is SPAM has autolearn=ham in the header and this 
> is wrong.
> 
>  
> 
> I am Running SPAMASSASSIN 3.0.3 on a Linux Red Hat 9 server. (just 
> upgraded) did this in version 3.0.2 also, unrelated I know.
> 
>  
> 
>  
> 
>  
> 
> Thanks in advance,
> 
>  
> 
> Robert
> 
>  
> 
>  
> 
>  
> 
>  
> 
>  
> 
>  
> 
> Peace he would say instead of goodbye....peace my brother.
> 
>  
> 
Remove the bayes db. What are you using? File based? SQL based? Need 
more info about that. Also in your case, you may either A) turn off 
autolearn B) change thresholds for spam/ham so this is unlikely to 
happen again.

-- 
Thanks,
James

Re: autolearn=ham

[Matt Kettler <mk...@evi-inc.com>]

Kelson wrote:

> Matt Kettler wrote:
>
>> Robert Swan wrote:
>>
>>> How do I clear, or unlearn the bayes filter it seems that it is picking
>>> up wrong. E-mail that is SPAM has autolearn=ham in the header and this
>>> is wrong.
>>
>>
>> Is it?
>
>
> If it's spam being learned as ham, then yes, it is wrong.  Autolearn
> may be doing what it's supposed to, but it's still a false negative. 
> An expected one, but a misclassification nonetheless.


True. I mis-read Robert's message as implying that the SA autolearn
mechanism was going haywire and randomly learning spam as ham for no
clear reason. Hence my answer.

Sorry for any confusion it may have created.

(The rest of the message is generally correct, albeit topically
misdirected. The facts about how the autolearner works in my message are
correct, albeit some details are omitted for simplicity. Opinions about
the threshold are my personal opinions, but they are my actual opinions.)

Re: autolearn=ham

[Kelson <ke...@speed.net>]

Matt Kettler wrote:
> Robert Swan wrote:
> 
>>How do I clear, or unlearn the bayes filter it seems that it is picking
>>up wrong. E-mail that is SPAM has autolearn=ham in the header and this
>>is wrong.
> 
> Is it?

If it's spam being learned as ham, then yes, it is wrong.  Autolearn may 
be doing what it's supposed to, but it's still a false negative.  An 
expected one, but a misclassification nonetheless.

Robert: just running sa-learn --spam will unlearn the message, then 
re-learn it as spam.

-- 
Kelson Vibber
SpeedGate Communications <www.speed.net>

Re: autolearn=ham

[Matt Kettler <mk...@evi-inc.com>]

Robert Swan wrote:

>How do I clear, or unlearn the bayes filter it seems that it is picking
>up wrong. E-mail that is SPAM has autolearn=ham in the header and this
>is wrong.
>
>  
>
Is it?

The autolearner uses the score the message would have gotten if bayes
was disabled, all userconf (ie: white/blacklist) rules were disabled,
and the AWL was disabled.

Post a X-Spam-Status header for the message in question and we can give
you some more specific advice, but just because the final score
indicated spam it doesn't mean the autolearner can't decide it's ham.
This is particularly true for message that got heavily hit on a
blacklist or AWL rule.

IMHO, the default ham learning threshold in current versions of SA is
begging for problems like this. I keep mine set at a tiny negative
score, but also have a collection of nonspam rules with tiny negative
scores. This way, autolearning as ham must be earned by hitting one of
the negative scoring rules, but the negative scoring rules can't be
abused by spammers as they collectively add up to less than -1.0.

Re: autolearn=ham

[Andy Jezierski <aj...@stepan.com>]

"Robert Swan" <rs...@nskinc.com> wrote on 05/02/2005 02:15:45 PM:

> How do I clear, or unlearn the bayes filter it seems that it is 
> picking up wrong. E-mail that is SPAM has autolearn=ham in the 
> header and this is wrong.
> 
> I am Running SPAMASSASSIN 3.0.3 on a Linux Red Hat 9 server. (just 
> upgraded) did this in version 3.0.2 also, unrelated I know.
> 
> 
> 
> Thanks in advance,
> 
> Robert

If it's a single message try:  sa-learn --forget < 
orginal.message.to.unlearn

If on the other hand you want to clear out the entire bayes db because you 
think it's corrupted then use:  sa-learn --clear

man sa-learn   for more info.

Andy