On Mon, 12 Apr 2021, [email protected] wrote:
John Hardin writes:
> From: John Hardin <[email protected]>
> Date: Mon, 12 Apr 2021 07:29:03 -0700 (PDT)
>
> On Sun, 11 Apr 2021, Loren Wilton wrote:
>
> >> 3.5 BAYES_99 BODY: Bayes spam probability is 99 to 100%
> >> [score: 1.0000]
> >> 0.5 BAYES_999 BODY: Bayes spam probability is 99.9 to 100%
> >> [score: 1.0000]
> >
> > I have
> > 5.0 BAYES_99 BODY: Bayes spam probability is 99 to 100%
> > [score: 1.0000]
> > 0.2 BAYES_999 BODY: Bayes spam probability is 99.9 to 100%
> > [score: 1.0000]
> >
> > I suggest raising BAYES_99 to at least 5.
>
> It'd be better to instead boost BAYES_999 to Poison Pill status, as the
> confidence is higher.
Increasing the score for BAYES_99 and BAYES_999 is a fine idea as long
as bayes is accurately trained and well maintained with sufficient
email and any mistakes corrected. People with that sort of trained
bayes tend to know it. Doing a general suggestion to increase the
BAYES scores seems rather misguided.
I'm suggesting that *only* BAYES_999 should be increased. I agree that you
should only do so if your Bayes training is reliable (i.e. *not*
end-user-driven without review).
--
John Hardin KA7OHZ http://www.impsec.org/~jhardin/
[email protected] pgpk -a [email protected]
key: 0xB8732E79 -- 2D8C 34F4 6411 F507 136C AF76 D822 E6E6 B873 2E79
-----------------------------------------------------------------------
What the hell is an "Aluminum Falcon"?? -- Emperor Palpatine
-----------------------------------------------------------------------
Tomorrow: Thomas Jefferson's 278th Birthday
