Hi all
I've disabled autolearn for a week and trained SA with tons of spams.
I still receive same spam messages with "random" scores. I've attached the
zip
file that contains same template with different sender addresses. Scores
seem
randomly to me; from 0 to 2.6, etc. Sample spam messages attached:
https://www.dropbox.com/s/h9kennmy0kerjc3/_spams.zip?dl=0
We have got 500+ inboxes but a few addresses are the victims of mass spams
these days;
same users got 20+ very same spams per day, maybe more There are a few
different
spam templates. I search for specific words like "profile is", "Ginger",
"hot date",
"Appswarm" in files less than 5KB.
I also found a weird break point; one message has score of 1.6 and the very
next same
message has score of -1.1. The two files are here to examine:
https://www.dropbox.com/s/kiv81gclyaaqmsq/_high-and-then-low.zip?dl=0
I think I'm gonna install a fresh new SA.
On Wed, Jan 18, 2017 at 9:43 AM, Emin Akbulut <eminakbu...@gmail.com> wrote:
> Hi all
>
> I've disabled autolearn for a week and trained SA with tons of spams.
> I still receive same spam messages with "random" scores. I've attached the
> zip
> file that contains same template with different sender addresses. Scores
> seem
> randomly to me; from 0 to 2.6, etc. Sample spam messages attached:
> https://www.dropbox.com/s/h9kennmy0kerjc3/_spams.zip?dl=0
>
> We have got 500+ inboxes but a few addresses are the victims of mass spams
> these days;
> same users got 20+ very same spams per day, maybe more There are a few
> different
> spam templates. I search for specific words like "profile is", "Ginger",
> "hot date",
> "Appswarm" in files less than 5KB.
>
> I also found a weird break point; one message has score of 1.6 and the
> very next same
> message has score of -1.1. The two files are here to examine:
> https://www.dropbox.com/s/kiv81gclyaaqmsq/_high-and-then-low.zip?dl=0
>
>
> I think I'm gonna install a fresh new SA.
>
>
> On Tue, Jan 10, 2017 at 7:22 PM, John Hardin <jhar...@impsec.org> wrote:
>
>> On Tue, 10 Jan 2017, Emin Akbulut wrote:
>>
>> I've trained the SA and it worked for a while but now it's useless.
>>>
>>>
>>> How can I prevent those spams? They look like poems
>>>
>>> * * * * * * * * * * * * * * *
>>>
>>> I am a tender and passionate girl-student.
>>>
>>> I assure satisfaction and all the pleasures to my lover!
>>>
>>> I will be alone on Christmas holidays((
>>>
>>> Would you like to keep company?
>>>
>>> Myphotos and questionnaire are waiting for you! <http://link.removed>
>>>
>>> * * * * * * * * * * * * * * *
>>>
>>
>> That looks like it contains some good Bayes fodder. Question: how are you
>> training? Is this a large install where you allow users to directly fee the
>> corpus, or do you hand-review every message in your training corpus?
>>
>> You might want to do the following:
>>
>> (1) review your entire ham and spam corpora again.
>> (2) wipe your bayes database and re-learn from your saved corpora
>> (3) disable autolearn if you have it enabled.
>>
>> --
>> John Hardin KA7OHZ http://www.impsec.org/~jhardin/
>> jhar...@impsec.org FALaholic #11174 pgpk -a jhar...@impsec.org
>> key: 0xB8732E79 -- 2D8C 34F4 6411 F507 136C AF76 D822 E6E6 B873 2E79
>> -----------------------------------------------------------------------
>> Markley's Law (variant of Godwin's Law): As an online discussion
>> of gun owners' rights grows longer, the probability of an ad hominem
>> attack involving penis size approaches 1.
>> -----------------------------------------------------------------------
>> 7 days until Benjamin Franklin's 311th Birthday
>>
>
>
