Mailing List Archive

On 14.11.23 13:05, joe a wrote:
>Low volume home office user and system.
>
>Occasionally when first dealing with a new entity, their
>correspondence gets flagged as SPAM.
>
>When I whitelist these, what should be done with those messages that
>might remain in "flagged SPAM" or "Missed SPAM"?, thinking along lines
>of keeping BAYES "clean and sharp". So to speak.
>
>Leave as is? Delete and re learn?

Simply relearn FPs. Unless you have huge misclassification issue, learning
as few mail as one should fix BAYES issues.

--
Matus UHLAR - fantomas, uhlar@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.
M$ Win's are shit, do not use it !

On Tue, 14 Nov 2023, joe a wrote:

> Low volume home office user and system.
>
> Occasionally when first dealing with a new entity, their correspondence gets
> flagged as SPAM.
>
> When I whitelist these, what should be done with those messages that might
> remain in "flagged SPAM" or "Missed SPAM"?, thinking along lines of keeping
> BAYES "clean and sharp". So to speak.
>
> Leave as is? Delete and re learn?

For a low volume home office user, I would simply NOT autolearn. Set up a
hambox and a spambox and manually feed them and train from them.


--
John Hardin KA7OHZ http://www.impsec.org/~jhardin/
jhardin@impsec.org pgpk -a jhardin@impsec.org
key: 0xB8732E79 -- 2D8C 34F4 6411 F507 136C AF76 D822 E6E6 B873 2E79
-----------------------------------------------------------------------
The reason it took so long to get Bin Laden is that it took the
SEALs five years to swim that far into the desert. -- anon
-----------------------------------------------------------------------
1,263 days since the first private commercial manned orbital mission (SpaceX)

On 11/14/2023 20:48:27, John Hardin wrote:
> On Tue, 14 Nov 2023, joe a wrote:
>
>> Low volume home office user and system.
>>
>> Occasionally when first dealing with a new entity, their
>> correspondence gets flagged as SPAM.
>>
>> When I whitelist these, what should be done with those messages that
>> might remain in "flagged SPAM" or "Missed SPAM"?, thinking along lines
>> of keeping BAYES "clean and sharp".  So to speak.
>>
>> Leave as is?  Delete and re learn?
>
> For a low volume home office user, I would simply NOT autolearn. Set up
> a hambox and a spambox and manually feed them and train from them.
>
>

I have autolearn off and have a spam and ham folder set up and "relearn"
twice daily.

On 11/14/2023 13:46:11, Matus UHLAR - fantomas wrote:
> On 14.11.23 13:05, joe a wrote:
>> Low volume home office user and system.
>>
>> Occasionally when first dealing with a new entity, their
>> correspondence gets flagged as SPAM.
>>
>> When I whitelist these, what should be done with those messages that
>> might remain in "flagged SPAM" or "Missed SPAM"?, thinking along lines
>> of keeping BAYES "clean and sharp".  So to speak.
>>
>> Leave as is?  Delete and re learn?
>
> Simply relearn FPs. Unless you have huge misclassification issue,
> learning as few mail as one should fix BAYES issues.
>

Move previously tagged SPAM into HAM folder and "relearn"?

>>On 14.11.23 13:05, joe a wrote:
>>>Low volume home office user and system.
>>>
>>>Occasionally when first dealing with a new entity, their
>>>correspondence gets flagged as SPAM.
>>>
>>>When I whitelist these, what should be done with those messages
>>>that might remain in "flagged SPAM" or "Missed SPAM"?, thinking
>>>along lines of keeping BAYES "clean and sharp".? So to speak.
>>>
>>>Leave as is?? Delete and re learn?

>On 11/14/2023 13:46:11, Matus UHLAR - fantomas wrote:
>>Simply relearn FPs. Unless you have huge misclassification issue,
>>learning as few mail as one should fix BAYES issues.

On 14.11.23 22:02, joe a wrote:
>Move previously tagged SPAM into HAM folder and "relearn"?

yes.
re-training SA on the same file works as if previous training was not done.

--
Matus UHLAR - fantomas, uhlar@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 just got lost in thought. It was unfamiliar territory.

On Tue, 14 Nov 2023, joe a wrote:

> On 11/14/2023 13:46:11, Matus UHLAR - fantomas wrote:
>> On 14.11.23 13:05, joe a wrote:
>>> Low volume home office user and system.
>>>
>>> Occasionally when first dealing with a new entity, their correspondence
>>> gets flagged as SPAM.
>>>
>>> When I whitelist these, what should be done with those messages that
>>> might remain in "flagged SPAM" or "Missed SPAM"?, thinking along lines of
>>> keeping BAYES "clean and sharp".  So to speak.
>>>
>>> Leave as is?  Delete and re learn?
>>
>> Simply relearn FPs. Unless you have huge misclassification issue, learning
>> as few mail as one should fix BAYES issues.
>>
>
> Move previously tagged SPAM into HAM folder and "relearn"?

Right. Train on misclassifications.

Also if there was a ham in your spam corpus review why it got
misclassified in the first place.


--
John Hardin KA7OHZ http://www.impsec.org/~jhardin/
jhardin@impsec.org pgpk -a jhardin@impsec.org
key: 0xB8732E79 -- 2D8C 34F4 6411 F507 136C AF76 D822 E6E6 B873 2E79
-----------------------------------------------------------------------
Poor planning on your part does not create
an obligation on my part.
-----------------------------------------------------------------------
1,264 days since the first private commercial manned orbital mission (SpaceX)