Am 21.09.2016 um 18:47 schrieb Bowie Bailey:
That is ridiculous. The more training bayes gets the better it works.
And manual training is better than autolearning because autolearning can
automatically learn false positives and false negatives and cause
problems for the database.
correct according to my experience.
On 22.09.16 10:16, Thomas Barth wrote:
And what about filter poisening? In the last 10 hours my company
address got 43 mails classified as spam (even a virus mail detected
today). And there was one mail classified as spam due to my rule (bad
X-Spam-Status: Yes, score=7.474 tag=2 tag2=6.31 kill=6.31
tests=[MESSAGEID_LOCAL=3, RDNS_NONE=1.274, RELAYCOUNTRY_BAD=3.2]
there's no poisoning, unless you count two your rules with indcredibly high
score (which is why rules should not have too big scores).
Lower scores of those two...
according to your previous mail you have:
- rule RP_MATCHES_RCVD scoring -3.096
that should be increased to -0.001 (already recommended by li...@rhsoft.net)
or killed/zeroed (recommended by me)
- rule URIBL_BLOCKED indicating you use DNS server used by too many clients.
Set up your own recursing nameserver, BIND or unbound and don't configure it
to forward queries to upstream.
There is no spam content, am I right? Normal words and content that a
normal person can use.
spammers typically use "normal words and content that a normal person can
use", that's why it's so hard to catch spam. BAYES helps you find the
differencies between spam and ham and you can trust us it works great.
I dont need spam learning for all the mails
already classified as spam with high score. Spam with low score are
interesting for spam learning like this one. But when I use these
mails for spam learning there is a risk of false positive some day,
because it has learned that normal mails are also spam?
you must of course train ham mail, especially false positives, bayes needs
to be trained with ham too, because it needs to see the differencies.
if you train ham with big bayes score, it will help you much.
since you already got false positive without using BAYES, I think it's
useless to be reluctant about it.
Matus UHLAR - fantomas, uh...@fantomas.sk ; http://www.fantomas.sk/
Warning: I wish NOT to receive e-mail advertising to this address.
Varovanie: na tuto adresu chcem NEDOSTAVAT akukolvek reklamnu postu.
I intend to live forever - so far so good.