On Sat, 21 Mar 2015 20:51:49 +
RW rwmailli...@googlemail.com wrote:
The two calculations produce the same result when
Ns2/Nh2 = (Ns2-Ns1)/(Nh2-Nh1)
i.e. if spam and ham is being added in the same ratio that it occurs
in the database.
On 21.03.15 22:54, David F. Skoll wrote:
Yup,
Hi,
I think it seldom pays to be too clever with Bayes. If (and this is a
big if) you have a large enough sample of mail, in our experience it's
better just to shovel it all into Bayes than to be selective about
what you present to Bayes. The Bayes algorithms are usually pretty
good at
Am 22.03.2015 um 17:44 schrieb Alex Regan:
Would it be helpful to have something that graphs the data to monitor
the effect of learning changes? Does something already exist?
i am doing something similar recently by one per night iterate through
all ham/spam smaples to get a overview how
On Sun, 22 Mar 2015 12:44:26 -0400
Alex Regan mysqlstud...@gmail.com wrote:
[...]
So instead of trying to figure out the proper expiry period, you just
start over completely every two weeks?
No, we use a two-week sliding window to construct our Bayes DB. We don't learn
for two weeks and
On Fri, 20 Mar 2015 22:08:23 -0400
David F. Skoll wrote:
Bayes expiry is a tricky thing. To do expiry in a way that can be
justified mathematically, you really should expire messages, not
individual tokens. Otherwise, you're skewing the probabilities.
The only token probabilities that can
On Sat, 21 Mar 2015 15:10:19 +
RW rwmailli...@googlemail.com wrote:
The only token probabilities that can be skewed by token expiry are
those than get expired and are then subsequently relearned.
Yup. But they might turn out to be important.
Even then when those tokens are relearned the
On Sat, 21 Mar 2015 13:13:13 -0400
David F. Skoll wrote:
On Sat, 21 Mar 2015 15:10:19 +
RW rwmailli...@googlemail.com wrote:
The only token probabilities that can be skewed by token expiry are
those than get expired and are then subsequently relearned.
Yup. But they might turn out
On Sat, 21 Mar 2015 20:51:49 +
RW rwmailli...@googlemail.com wrote:
The two calculations produce the same result when
Ns2/Nh2 = (Ns2-Ns1)/(Nh2-Nh1)
i.e. if spam and ham is being added in the same ratio that it occurs
in the database.
Yup, that's correct; I got it wrong by extrapolating
On 20.03.15 09:30, Reindl Harald wrote:
why would you want poems or cooking recipes trained as spam?
and why not? they still may contain stuff that helps differ spam from ham,
you never know...
once I have trained spam report as ham, extracted the original spam and
trained that one as spam.
Am 20.03.2015 um 11:40 schrieb Matus UHLAR - fantomas:
On 20.03.15 09:30, Reindl Harald wrote:
why would you want poems or cooking recipes trained as spam?
and why not?
i think i have explained it often enough now
they still may contain stuff that helps differ spam from ham,
you never
Am 19.03.2015 um 23:52 schrieb RW:
On Thu, 19 Mar 2015 20:46:10 +0100
Reindl Harald wrote:
Am 19.03.2015 um 20:35 schrieb RW:
On Thu, 19 Mar 2015 01:12:15 +0100
Reindl Harald wrote:
the last point is easy to prove by having the old, unmodified
corpus and run spamc against the cleaned
On Fri, 20 Mar 2015 17:09:29 -0400
Kevin A. McGrail kmcgr...@pccc.com wrote:
And I've heard arguments for and against removing the poisoning
information. YMMV.
I think it seldom pays to be too clever with Bayes. If (and this is a
big if) you have a large enough sample of mail, in our
Hi,
On 03/20/2015 06:50 AM, Reindl Harald wrote:
Am 20.03.2015 um 11:40 schrieb Matus UHLAR - fantomas:
On 20.03.15 09:30, Reindl Harald wrote:
why would you want poems or cooking recipes trained as spam?
and why not?
i think i have explained it often enough now
I've heard arguments in
On 3/20/2015 4:58 PM, Alex Regan wrote:
Can someone with some authority (ie, those that wrote it or understand
the code) add some input?
Bayesian token systems are probably the basis for quite a number of
master and doctoral level thesiseseses (whatever the plural of thesis).
Understanding
On Thu, 19 Mar 2015 20:46:10 +0100
Reindl Harald wrote:
Am 19.03.2015 um 20:35 schrieb RW:
On Thu, 19 Mar 2015 01:12:15 +0100
Reindl Harald wrote:
the last point is easy to prove by having the old, unmodified
corpus and run spamc against the cleaned bayes database and the
final
Am 19.03.2015 um 09:22 schrieb Matus UHLAR - fantomas:
Am 18.03.2015 um 22:29 schrieb David B Funk:
Just have an internal mail-submission port that isn't routed thru SA
On 18.03.15 22:46, Reindl Harald wrote:
may not be possible if you have hundrets of domains without setup a
internal DNS
Am 18.03.2015 um 22:29 schrieb David B Funk:
Just have an internal mail-submission port that isn't routed thru SA
On 18.03.15 22:46, Reindl Harald wrote:
may not be possible if you have hundrets of domains without setup a
internal DNS view just for a different MX
You apparently mean
On 3/19/15, Reindl Harald h.rei...@thelounge.net wrote:
Am 19.03.2015 um 12:57 schrieb Nick Edwards:
On 3/19/15, Steve Freegard s...@fsl.com wrote:
On 18/03/15 21:46, Reindl Harald wrote:
Wow - you must be fun at parties...
HAHAHA reindl doesnt go to parties because , he has no friends
I am getting tired of all of the personal attacks. End it, everyone. Stick to
facts or take it offlist. I don't care which.
Regards,
KAM
Am 19.03.2015 um 12:57 schrieb Nick Edwards:
On 3/19/15, Steve Freegard s...@fsl.com wrote:
On 18/03/15 21:46, Reindl Harald wrote:
Wow - you must be fun at parties...
HAHAHA reindl doesnt go to parties because , he has no friends and no
one would have him, he can start a fight when he's
On 18/03/15 21:46, Reindl Harald wrote:
Am 18.03.2015 um 22:29 schrieb David B Funk:
Just have an internal mail-submission port that isn't routed thru SA
may not be possible if you have hundrets of domains without setup a
internal DNS view just for a different MX
In general you don't want
On 3/19/15, Steve Freegard s...@fsl.com wrote:
On 18/03/15 21:46, Reindl Harald wrote:
Wow - you must be fun at parties...
HAHAHA reindl doesnt go to parties because , he has no friends and no
one would have him, he can start a fight when he's the only tosser in
the room.
hahahaha come on reindl, do you think people here will bother to click
your links when all they only need do is google reindl harald to
get an unbiased view of who you really are and the sort of offensive
abusive creep you really are.
fedora users BANNED
roundcube MODERATED
postifx BANNED
Am 19.03.2015 um 12:30 schrieb Steve Freegard:
On 18/03/15 21:46, Reindl Harald wrote:
Am 18.03.2015 um 22:29 schrieb David B Funk:
Just have an internal mail-submission port that isn't routed thru SA
may not be possible if you have hundrets of domains without setup a
internal DNS view
Am 19.03.2015 um 13:09 schrieb Nick Edwards:
On 3/19/15, Reindl Harald h.rei...@thelounge.net wrote:
Am 19.03.2015 um 12:57 schrieb Nick Edwards:
On 3/19/15, Steve Freegard s...@fsl.com wrote:
On 18/03/15 21:46, Reindl Harald wrote:
Wow - you must be fun at parties...
HAHAHA reindl
well, i never pretended to be the nicest person on the world but the
difference between both of us is that i *never ever* go out and seek
postings from a specific person and call him/her names without any
context to the thread and technical background just because it's he/her
and in fact i
Last warning. Offlist with these emails or next step is list removal.
Regards,
KAM
On 3/19/15, Reindl Harald h.rei...@thelounge.net wrote:
and in fact i was *never ever* that abusive as you are all the time
beause there is a difference in get heatet in a technical discussion or
like you do absue for the sake of abuse
hahahaha thankfulyl google shows otherwise
so you
Am 18.03.2015 um 22:29 schrieb David B Funk:
Just have an internal mail-submission port that isn't routed thru SA
On 18.03.15 22:46, Reindl Harald wrote:
may not be possible if you have hundrets of domains without setup a
internal DNS view just for a different MX
Am 19.03.2015 um 09:22
On Thu, 19 Mar 2015 01:12:15 +0100
Reindl Harald wrote:
Am 19.03.2015 um 00:54 schrieb RW:
This is nothing to do with auto-learning. There is a difference
between miss-training and training with spam that contains
so-called Bayes poison. Bayes is best trained on what is in
real-world
Am 19.03.2015 um 20:35 schrieb RW:
On Thu, 19 Mar 2015 01:12:15 +0100
Reindl Harald wrote:
Am 19.03.2015 um 00:54 schrieb RW:
This is nothing to do with auto-learning. There is a difference
between miss-training and training with spam that contains
so-called Bayes poison. Bayes is best
Am 19.03.2015 um 13:53 schrieb Matus UHLAR - fantomas:
Am 18.03.2015 um 22:29 schrieb David B Funk:
Just have an internal mail-submission port that isn't routed thru SA
On 18.03.15 22:46, Reindl Harald wrote:
may not be possible if you have hundrets of domains without setup a
internal DNS
On 18 Mar 2015, at 17:24, Axb wrote:
Why is Internal or local mail sent thru SA?
If the MTA handles outbound mail from Windows users, passing their mail
through SA is prudent. There are spamming trojans that figure out how to
use the victim's legitimate submission config including
Am 19.03.2015 um 19:48 schrieb Bill Cole:
On 18 Mar 2015, at 17:24, Axb wrote:
Why is Internal or local mail sent thru SA?
If the MTA handles outbound mail from Windows users, passing their mail
through SA is prudent. There are spamming trojans that figure out how to
use the victim's
I noticed that some of the Zimbra auto-generated emails (reports on various
bits) are getting hit with RBL scoring for some customers. This appears to
be because they are (quite reasonably) using private IPs on some of thier
internal Zimbra servers. However, when it goes through the MTA, it
Am 19.03.2015 um 00:54 schrieb RW:
On Wed, 18 Mar 2015 23:57:13 +0100
Reindl Harald wrote:
Am 18.03.2015 um 23:34 schrieb RW:
On Wed, 18 Mar 2015 22:46:14 +0100
Reindl Harald wrote:
frankly i trained over months with *hand chosen* mail smaples and
spent nearly two weeks day and night to
On Wed, 18 Mar 2015 22:46:14 +0100
Reindl Harald wrote:
frankly i trained over months with *hand chosen* mail smaples and
spent nearly two weeks day and night to remove bayes-posioning from
the samples and rebuild bayes from scratch leading in reduce the
ntokens from 170 to 150
Am 18.03.2015 um 23:34 schrieb RW:
On Wed, 18 Mar 2015 22:46:14 +0100
Reindl Harald wrote:
frankly i trained over months with *hand chosen* mail smaples and
spent nearly two weeks day and night to remove bayes-posioning from
the samples and rebuild bayes from scratch leading in reduce the
On Wed, 18 Mar 2015 23:57:13 +0100
Reindl Harald wrote:
Am 18.03.2015 um 23:34 schrieb RW:
On Wed, 18 Mar 2015 22:46:14 +0100
Reindl Harald wrote:
frankly i trained over months with *hand chosen* mail smaples and
spent nearly two weeks day and night to remove bayes-posioning from
On 03/18/2015 10:39 PM, Reindl Harald wrote:
Am 18.03.2015 um 22:24 schrieb Axb:
On 03/18/2015 09:48 PM, Quanah Gibson-Mount wrote:
I noticed that some of the Zimbra auto-generated emails (reports on
various bits) are getting hit with RBL scoring for some customers. This
appears to be
On 03/18/2015 10:46 PM, Reindl Harald wrote:
Am 18.03.2015 um 22:29 schrieb David B Funk:
Just have an internal mail-submission port that isn't routed thru SA
may not be possible if you have hundrets of domains without setup a
internal DNS view just for a different MX
In general you don't
Am 18.03.2015 um 22:58 schrieb Axb:
On 03/18/2015 10:46 PM, Reindl Harald wrote:
Am 18.03.2015 um 22:29 schrieb David B Funk:
Just have an internal mail-submission port that isn't routed thru SA
may not be possible if you have hundrets of domains without setup a
internal DNS view just for
Am 18.03.2015 um 21:48 schrieb Quanah Gibson-Mount:
I noticed that some of the Zimbra auto-generated emails (reports on
various bits) are getting hit with RBL scoring for some customers. This
appears to be because they are (quite reasonably) using private IPs on
some of thier internal Zimbra
Am 18.03.2015 um 22:11 schrieb Reindl Harald:
Am 18.03.2015 um 21:48 schrieb Quanah Gibson-Mount:
I noticed that some of the Zimbra auto-generated emails (reports on
various bits) are getting hit with RBL scoring for some customers. This
appears to be because they are (quite reasonably) using
On 03/18/2015 09:48 PM, Quanah Gibson-Mount wrote:
I noticed that some of the Zimbra auto-generated emails (reports on
various bits) are getting hit with RBL scoring for some customers. This
appears to be because they are (quite reasonably) using private IPs on
some of thier internal Zimbra
Am 18.03.2015 um 22:24 schrieb Axb:
On 03/18/2015 09:48 PM, Quanah Gibson-Mount wrote:
I noticed that some of the Zimbra auto-generated emails (reports on
various bits) are getting hit with RBL scoring for some customers. This
appears to be because they are (quite reasonably) using private
Hi Quanah,
X-Spam-Status: Yes, score=10.297 tagged_above=-10 required=10
tests=[ALL_TRUSTED=-1, BAYES_00=-0.5, T_RP_MATCHES_RCVD=-0.01,
URIBL_BLACK=3.25, URIBL_DBL_SPAM=2.5, URIBL_JP_SURBL=1.25,
URIBL_RHS_DOB=1.514, URIBL_SBL_A=0.1, URIBL_WS_SURBL=1.608,
--On Wednesday, March 18, 2015 11:11 PM +0100 Reindl Harald
h.rei...@thelounge.net wrote:
The IP is clearly listed in trusted_networks
your problem are not RBL's
your problem are URIBL's and so mail content
ask yourself why autogenerated mails contains crap URLs listed on
URIBL_BLACK,
On Wed, 18 Mar 2015, Reindl Harald wrote:
Am 18.03.2015 um 21:48 schrieb Quanah Gibson-Mount:
I noticed that some of the Zimbra auto-generated emails (reports on
various bits) are getting hit with RBL scoring for some customers. This
appears to be because they are (quite reasonably) using
Am 18.03.2015 um 22:29 schrieb David B Funk:
Just have an internal mail-submission port that isn't routed thru SA
may not be possible if you have hundrets of domains without setup a
internal DNS view just for a different MX
In general you don't want auto-mail running thru SA for this
50 matches
Mail list logo