Hello Who, Friday, January 23, 2004, 5:08:57 AM, you wrote:
WK> Anthony Martinez wrote: >> I got a spam today where the X-Originating-IP header wasn't a number. Hotmail >> always puts the dotted quad in the header. WK> I have been receiving a good many of these lately. I am hestant to add WK> any rules for them yet because all the ones I have been receiving seem WK> to also contain a list of words that can only be there to spoil baysian WK> tracking. WK> Is there anyway to avoid adding the spoilers to the spam baysian list? IMO, the spoiler seem to improve SA's Bayes activity. I used to exclude these from my sa-learn, but for the past two or three months I've fed all of the so-called spam poison into the database, and bayes has gotten very good at not only a) not treating this as ham-sign, but b) apparently recognizing it as spam-sign. Almost all "spam poison" emails now hits Bayes scores 80% or higher, most at 99%. Sylvain Robitaille's SR_hxo_OrigIPNotIP rule posted here this week is excellent for this purpose (hit over 3k spam, 0 ham, on my corpus). Bob Menschel ------------------------------------------------------- The SF.Net email is sponsored by EclipseCon 2004 Premiere Conference on Open Tools Development and Integration See the breadth of Eclipse activity. February 3-5 in Anaheim, CA. http://www.eclipsecon.org/osdn _______________________________________________ Spamassassin-talk mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/spamassassin-talk
