> I think the Genetic Algorithm (GA) assigns all the scores now.
> GA's are very
> powerful optimization tools, and if the GA lowered those scores, it likely
> raised (compensated) other scores that were more common spam signatures.
>
> The GA is only as good as the population of data it is run on
[EMAIL PROTECTED]
> [mailto:[EMAIL PROTECTED]]On Behalf Of
> CertaintyTech - Ed Henderson
> Sent: Thursday, February 14, 2002 11:28 AM
> To: [EMAIL PROTECTED]
> Subject: [SAtalk] SA 2.01 low scores
>
>
> I have been seeing alot more Spam get thru (false negatives) in v2.01 than
I'll investigate -- it's weird, because both GAs seem to assign low
scores to "HUNZA_DIET_BREAD" even though it obviously only appears in
the spam corpus, and not in non-spam. In fact my GA which allows -ve
scores give it a -ve score! Now having said that, all the
HUNZA_DIET_BREAD messages in th
On 14 February 2002, CertaintyTech - Ed Henderson said:
> I have been seeing alot more Spam get thru (false negatives) in v2.01 than
> with v1.5. I have been comparing the scores of 1.5 with 2.01 to see why.
> Here is an interesting discovery: there are several scores in the
> 50_scores.cf file
I have been seeing alot more Spam get thru (false negatives) in v2.01 than
with v1.5. I have been comparing the scores of 1.5 with 2.01 to see why.
Here is an interesting discovery: there are several scores in the
50_scores.cf file that are 0.01 in value:
50_scores.cf:score A_HREF_TO_UNSUB
