On Jun 29, 2005, at 12:35 AM, Steve Lamb wrote:
> Chad Leigh -- Shire.Net LLC wrote:
>
>> At least a good portion is sent so widely that one user's spam is
>> similar
>> to another users spam.
>>
>
> True, but you're forgetting that there are two sets to each
> user. Spam
> and ham. While spam is generally going ot be a constant across
> users the
> likelihood of ham being constant is not.
>
No, just that it seems on our systems, ham is no longer all that
important. (After the initial setup 2 years ago where a ton of ham
was fed in...) I feed in 100-300 spams for every ham that gets fed
in, almost all of it from my own receipt across 4-6 accounts I have.
I get some from some of the smarter (technically) customers once in a
while. But feedback from customers is that the spamassassin gets
very very few false positives or negatives. What I do do is feed in
ALL spam back as spam again to reinforce it and only feed in the
false positives as ham and occasionally a few other things while I am
at it and just copy a few random good mails out of my inboxes....
It turns out to work very well
Chad
---
Chad Leigh -- Shire.Net LLC
Your Web App and Email hosting provider
chad@???
