changing spamassassin points configuration

[Pete]

Julian Field wrote:

> Stuff that isn't spam.
>
> At 11:41 03/03/2004, you wrote:
>
>> err...what's "ham"?
>>
>> At 07:09 PM 3/3/2004, you wrote:
>>
>>> kfliong wrote:
>>>
>>>> Hi,
>>>>
>>>> I have this email which is not spam but have a score of 5.642 which is
>>>> high
>>>> as default of more than 5 is considered spam.
>>>>
>>>> Can I know how I can reduce the score?
>>>>
>>>> spam, SpamAssassin (score=5.642, required 5, BAYES_90 2.10,
>>>> DATE_IN_PAST_12_24 0.75, DEAR_SOMETHING 2.30, HTML_FONTCOLOR_BLUE
>>>> 0.10,
>>>> HTML_FONTCOLOR_UNSAFE 0.10, HTML_MESSAGE 0.10, HTML_TAG_BALANCE_A
>>>> 0.20)
>>>>
>>>> Also, the scores mainly comes from BAYES_90 2.10 and DEAR_SOMETHING
>>>> 2.30....where can i get more details on what those score means? Does
>>>> mailscanner uses a different config file for controlling spamassassin?
>>>>
>>>> thanks in advance
>>>>
>>>>
>>>> thanks
>>>>
>>> ISnt this a situation for learning as ham? I am NO expert, but if you
>>> have no other method maybe turn on archiving till you get a copy of
>>> this
>>> message, then sa-learn it as ham?:
>>
>>
>> thanks
>
>
> --
> Julian Field
> www.MailScanner.info
> MailScanner thanks transtec Computers for their support
>
> PGP footprint: EE81 D763 3DB0 0BFD E1DC 7222 11F6 5947 1415 B654
>
>
>
Since i think Julian's comment is confirmation - this is the sort of
thing that using Bayesian Learning (Bayes) with spama assassin will fix.

I am not well versed enought o try and explain it, so have a search
through the list archives, or google, its works plenty good with
mailscanner and spam assassin.