BayeIT Macros
ok im kinda confused here, I've checked these 3 macros out
1)
Spam Stats, last 7 days (%BAYESITVERSION)
Total Spam Emails: %STATSPAMLETTERS(168)
Total Clean Emails: %STATNONSPAMLETTERS(168)
BayesIT guessed right %calculate="100-(%STATSPAMERLETTERS(168)/%
STATSPAMLETTERS(168))"%% of the time
My email is %calculate="%STATNONSPAMLETTERS(168)/%STATSPAMLETTERS(168)"%%
spam
Spam Stats, last 7 days (BayesIt! 0.5.11)
Total Spam Emails: 270
Total Clean Emails: 246
BayesIT guessed right 99.94% of the time
My email is 0.91% spam
2)
Spam Stats, last 7 days (%BAYESITVERSION)
Total Spam Emails: %STATSPAMLETTERS(168)
Total Clean Emails: %STATNONSPAMLETTERS(168)
BayesIT guessed right %calculate="100-(%STATSPAMERLETTERS(168)/%
STATSPAMLETTERS(168))"%% of the time
My email is %calculate="%STATSPAMLETTERS(168)/%STATNONSPAMLETTERS(168)"%%
spam
Spam Stats, last 7 days (BayesIt! 0.5.11)
Total Spam Emails: 270
Total Clean Emails: 246
BayesIT guessed right 99.94% of the time
My email is 1.097561% spam
3)
Spam Stats, last 7 days (%BAYESITVERSION)
Total Spam Emails: %STATSPAMLETTERS(168)
Total Clean Emails: %STATNONSPAMLETTERS(168)
BayesIT guessed right %calculate="100-(%STATSPAMERLETTERS(168)/%
STATSPAMLETTERS(168))"%% of the time
My email is %_Sum=%Calc("%STATSPAMLETTERS(168)+%STATNONSPAMLETTERS(168)")%-
%_Quot=%Calc("%StatSpamLetters/%_sum")%-
%_Percent=%calc("100*%_quot")%-
%_Percent%% spam
Spam Stats, last 7 days (BayesIt! 0.5.11)
Total Spam Emails: 270
Total Clean Emails: 246
BayesIT guessed right 99.94% of the time
My email is 0% spam
if my total spam is 270 and total clean is 246, shouldn't my spam percentage
be 51% - 52% since my overall spam percentage is greater than clean or is the
270 my complete mails, 246 clean and the 24 left over spam
i cant see how i get 2% and below spam when my spam is greater than clean
mails
anyone want to help me out here?
these macros were posted by "Michael L. Wilson" on Aug 3rd under the subject
"Playing with BayesIT macros"
--
A pedestrian hit me and went under my car.
Current version is 2.12.00 | 'Using TBUDL' information:
http://www.silverstones.com/thebat/TBUDLInfo.html
Re: BayeIT Macros
Hello Drago,
On Tue, 24 Aug 2004, Drago mused about "BayeIT Macros" (at least in
part):
D> if my total spam is 270 and total clean is 246, shouldn't my spam percentage
D> be 51% - 52% since my overall spam percentage is greater than clean or is the
D> 270 my complete mails, 246 clean and the 24 left over spam
D> i cant see how i get 2% and below spam when my spam is greater than clean
D> mails
D> anyone want to help me out here?
My current macro set is
=
Spam Stats, last 24 hours (%BAYESITVERSION)
Total Spam Emails: %STATSPAMLETTERS(24)
Total Clean Emails: %STATNONSPAMLETTERS(24)
BayesIT guessed right %-
%calculate="100-(%STATSPAMERLETTERS(24)/%STATSPAMLETTERS(24))"%% of the time
My email is %-
%Calc("100*%STATSPAMLETTERS(24)/(%STATSPAMLETTERS(24)+%STATNONSPAMLETTERS(24))")%% spam
=
which produces...
=
Spam Stats, last 24 hours (BayesIt! 0.5.11)
Total Spam Emails: 188
Total Clean Emails: 439
BayesIT guessed right 99.989362% of the time
My email is 29.984051% spam
=
--
__ __ _ __
| \/ (_) ___| |__ __ _ ___| |
| |\/| | |/ __| '_ \ / _` |/ _ \ |
| | | | | (__| | | | (_| | __/ |
|_| |_|_|\___|_| |_|\__,_|\___|_|
:einsteinyoyo:
Join the new The Bat! Group - play around with The Bat!
[EMAIL PROTECTED]
Current version is 2.12.00 | 'Using TBUDL' information:
http://www.silverstones.com/thebat/TBUDLInfo.html
Re: BayeIT Macros
On Tue 24 August 2004, 12:30:49 +1000, Drago wrote: > ok im kinda confused here, I've checked these 3 macros out > <...snip...> > if my total spam is 270 and total clean is 246, shouldn't my spam percentage > be 51% - 52% since my overall spam percentage is greater than clean or is the > 270 my complete mails, 246 clean and the 24 left over spam > > i cant see how i get 2% and below spam when my spam is greater than clean > mails > > anyone want to help me out here? All your calculations involve fractions such as: %STATNONSPAMLETTERS(168)/%STATSPAMLETTERS(168) %STATSPAMERLETTERS(168)/%STATSPAMLETTERS(168) You are quoting these as if they were percentages, but to covert them to a percentage you should multiply them by 100 So, for instance, %STATNONSPAMLETTERS(168)/%STATSPAMLETTERS(168) is 246/270 which is 0.9, which is 91.111% Now I am guessing the meaning of the macros is: %STATSPAMLETTERS()- number of msgs classified as spam in time period %STATNONSPAMLETTERS() - number of msgs classified as nonspam in time period %STATSPAMERLETTERS() - number of msgs classified as spam but reclassified as nonspam in time period Lets abbreviate (for simplicity) as follows: %STATSPAMLETTERS()- %s %STATNONSPAMLETTERS() - %n %STATSPAMERLETTERS() - %e Then the correct calculations are: Percentage BayesIT guessed wrong = 100*%e/(%s+%n) Percentage BayesIT guessed right = 100*(1-%e/(%s+%n)) Percentage of email that is spam = 100*(%s/(%s+%n)) -- Robin Using The Bat! v2.12.00 on Windows XP 5.1 Build 2600 Service Pack 1 Current version is 2.12.00 | 'Using TBUDL' information: http://www.silverstones.com/thebat/TBUDLInfo.html
Re: BayeIT Macros
As Drago walked through cyberspace, he tripped on this message from Michael L.
Wilson which was posted on 24/08/2004 04:47:10 in gmane.mail.the-bat.user
>
> =
> Spam Stats, last 24 hours (%BAYESITVERSION)
> Total Spam Emails: %STATSPAMLETTERS(24)
> Total Clean Emails: %STATNONSPAMLETTERS(24)
> BayesIT guessed right %-
> %calculate="100-(%STATSPAMERLETTERS(24)/%STATSPAMLETTERS(24))"%% of the time
> My email is %-
> %Calc("100*%STATSPAMLETTERS(24)/(%STATSPAMLETTERS(24)+%STATNONSPAMLETTERS(24))
> ")%% spam =
>
> which produces...
> =
> Spam Stats, last 24 hours (BayesIt! 0.5.11)
> Total Spam Emails: 188
> Total Clean Emails: 439
> BayesIT guessed right 99.989362% of the time
> My email is 29.984051% spam
> =
thanx, that produced more of what i was looking for, now my spam to clean mail
is easier to understand
--
Computers run on faith, not electrons.
XanaNews 1.16.4.2
Current version is 2.12.00 | 'Using TBUDL' information:
http://www.silverstones.com/thebat/TBUDLInfo.html
Re: BayeIT Macros
As Drago walked through cyberspace, he tripped on this message from Robin Anson which was posted on 24/08/2004 04:53:47 in gmane.mail.the-bat.user > > Then the correct calculations are: >Percentage BayesIT guessed wrong = 100*%e/(%s+%n) >Percentage BayesIT guessed right = 100*(1-%e/(%s+%n)) >Percentage of email that is spam = 100*(%s/(%s+%n)) thanx, well that was an interesting read, basically the same as Michael's but defined differently -- The cat that ate the ball of yarnhad mittens! XanaNews 1.16.4.2 Current version is 2.12.00 | 'Using TBUDL' information: http://www.silverstones.com/thebat/TBUDLInfo.html
Re: BayeIT Macros
On Monday, August 23, 2004 it appears that Michael L. Wilson
wrote the following in regards to "BayeIT Macros":
MLW> My current macro set is
MLW> =
MLW> Spam Stats, last 24 hours (%BAYESITVERSION)
MLW> Total Spam Emails: %STATSPAMLETTERS(24)
MLW> Total Clean Emails: %STATNONSPAMLETTERS(24)
MLW> BayesIT guessed right %-
MLW> %calculate="100-(%STATSPAMERLETTERS(24)/%STATSPAMLETTERS(24))"%%
of the time
MLW> My email is %-
MLW>
%Calc("100*%STATSPAMLETTERS(24)/(%STATSPAMLETTERS(24)+%STATNONSPAMLETTERS(24))")%%
spam
MLW> =
Michael, Sorry to be dense but what does this QT macro look at & how
do you use it -- as an "run an external" -- & if so where do you place
it? TIA .
--
Jan Rifkinson
Ridgefield CT USA
TB! v2.13 "Lucky" Beta/7 W2K v5.0 Service Pack 4
Current version is 2.12.00 | 'Using TBUDL' information:
http://www.silverstones.com/thebat/TBUDLInfo.html
Re: BayeIT Macros
Hello Jan, On Tue, 24 Aug 2004, Jan mused about "BayeIT Macros" (at least in part): JR> Michael, Sorry to be dense but what does this QT macro look at & how JR> do you use it -- as an "run an external" -- & if so where do you place JR> it? TIA . This QT macro shows the current status of BayesIT, the anti-spam plug-in for The Bat!. It uses built in macros from BayesIT and can, therefore, only be used if BayesIT is installed. An example of the output is at the bottom of this email I call mine "spamcount" and just insert it where ever I want. Some people here will complain, no doubt, that my signature is to long with this information, so I use it sparingly. -- __ __ _ __ | \/ (_) ___| |__ __ _ ___| | | |\/| | |/ __| '_ \ / _` |/ _ \ | | | | | | (__| | | | (_| | __/ | |_| |_|_|\___|_| |_|\__,_|\___|_| :einsteinyoyo: Join the new The Bat! Group - play around with The Bat! [EMAIL PROTECTED] = Spam Stats, last 24 hours (BayesIt! 0.5.11) Total Spam Emails: 217 Total Clean Emails: 504 BayesIT guessed right 99.990783% of the time My email is 30.097087% spam = Current version is 2.12.00 | 'Using TBUDL' information: http://www.silverstones.com/thebat/TBUDLInfo.html
Re: BayeIT Macros
On Tuesday, August 24, 2004 it appears that Michael L. Wilson wrote the following in reply to my comments re: "BayeIT Macros": MLW> JR> Michael, Sorry to be dense but what does this QT macro look at & how MLW> JR> do you use it -- as an "run an external" -- & if so where do you place MLW> JR> it? TIA . MLW> This QT macro shows the current status of BayesIT, the MLW> anti-spam plug-in for The Bat!. It uses built in macros from BayesIT MLW> and can, therefore, only be used if BayesIT is installed. Understand that. But what I'm trying to find out is what is the macro looking at? Incoming mail? Is it part of a filter action? Thanks for taking the time to reply to my question(s). -- Jan Rifkinson Ridgefield CT USA TB! v2.13 "Lucky" Beta/7 W2K v5.0 Service Pack 4 Current version is 2.12.00 | 'Using TBUDL' information: http://www.silverstones.com/thebat/TBUDLInfo.html
Re: BayeIT Macros
Hello Jan, On Tue, 24 Aug 2004, Jan mused about "BayeIT Macros" (at least in part): JR> Understand that. But what I'm trying to find out is what is the JR> macro looking at? Incoming mail? Is it part of a filter action? JR> Thanks for taking the time to reply to my question(s). There is no filter action, except the filters built into BayesIT. As BayesIT works, it is constantly updating its records and database to be current. At any time, you can query this data through the macros. To see detailed BayesIT stats, go to options->preferences->Plugins. Highlight BayesIT and click information. -- __ __ _ __ | \/ (_) ___| |__ __ _ ___| | | |\/| | |/ __| '_ \ / _` |/ _ \ | | | | | | (__| | | | (_| | __/ | |_| |_|_|\___|_| |_|\__,_|\___|_| :einsteinyoyo: Join the new The Bat! Group - play around with The Bat! [EMAIL PROTECTED] = Spam Stats, last 24 hours (BayesIt! 0.5.11) Total Spam Emails: 217 Total Clean Emails: 502 BayesIT guessed right 99.990783% of the time My email is 30.180807% spam = Current version is 2.12.00 | 'Using TBUDL' information: http://www.silverstones.com/thebat/TBUDLInfo.html
Re: BayeIT Macros
On Tuesday, August 24, 2004 it appears that Michael L. Wilson wrote the following in reply to my comments re: "BayeIT Macros": MLW> There is no filter action, except the filters built into BayesIT. As MLW> BayesIT works, it is constantly updating its records and database to MLW> be current. At any time, you can query this data through the macros. Hi Michael. OK, I think it's sinking in. But here's my problem or what started me down this road: In my junk file I currently have 95 msgs from today Yet here's the stat output Spam Stats, last 24 hours (BayesIt! 0.5.11) Total Spam Emails: 2 Total Clean Emails: 182 BayesIT guessed right 99.5% of the time My email is 1.086957% spam So there's a problem somewhere unless I'm totally off the wall. I do use other filters to move obvious spam to my spam folder, thereby bypassing the BayesIT filter. Could that explain the difference between the real #s & the reported #s? MLW> To see detailed BayesIT stats, go to options->preferences->Plugins. MLW> Highlight BayesIT and click information. Thanks, this was interesting. -- Jan Rifkinson Ridgefield CT USA TB! v2.13 "Lucky" Beta/7 W2K v5.0 Service Pack 4 Current version is 2.12.00 | 'Using TBUDL' information: http://www.silverstones.com/thebat/TBUDLInfo.html
Re: BayeIT Macros
Hello Jan, On Tue, 24 Aug 2004, Jan mused about "BayeIT Macros" (at least in part): JR> I do use other filters to move obvious spam to my spam folder, JR> thereby bypassing the BayesIT filter. Could that explain the JR> difference between the real #s & the reported #s? I let BayesIT filter everything that is spam. I use no external notifier, and no internal spam filters. Therefore, that could be the problem -- __ __ _ __ | \/ (_) ___| |__ __ _ ___| | | |\/| | |/ __| '_ \ / _` |/ _ \ | | | | | | (__| | | | (_| | __/ | |_| |_|_|\___|_| |_|\__,_|\___|_| :einsteinyoyo: Join the new The Bat! Group - play around with The Bat! [EMAIL PROTECTED] = Spam Stats, last 24 hours (BayesIt! 0.5.11) Total Spam Emails: 219 Total Clean Emails: 505 BayesIT guessed right 99.990868% of the time My email is 30.248619% spam = Current version is 2.12.00 | 'Using TBUDL' information: http://www.silverstones.com/thebat/TBUDLInfo.html
Re: BayeIT Macros
ON Tuesday, August 24, 2004, 6:10:24 PM, you wrote: JR> I do use other filters to move obvious spam to my spam folder, JR> thereby bypassing the BayesIT filter. Could that explain the JR> difference between the real #s & the reported #s? Hi Jan, I don't think you are bypassing BasesIT. Baseyit works before TB! get its hands on it, so it is BayesIT first and than your filters. -- Best regards, Gerard -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= People by and large become what they think about themselves. Using The Bat! v2.12.00 on Windows XP 5.1 Build 2600 Service Pack 1 Current version is 2.12.00 | 'Using TBUDL' information: http://www.silverstones.com/thebat/TBUDLInfo.html
Re: BayeIT Macros
On Tuesday, August 24, 2004 it appears that Michael L. Wilson wrote the following in regards to "BayeIT Macros": MLW> I let BayesIT filter everything that is spam. I use no external MLW> notifier, and no internal spam filters. Therefore, that could be the MLW> problem And I see from your stats, BayesIT is running at nearly 100% accuracy. I guess a user couldn't ask for much more than that. Something for me to think about although I don't know if there's a difference in processing one way or the other but it would certainly free up the TB! filtering system to deal with other housekeeping matters. Thanks. -- Jan Rifkinson Ridgefield CT USA TB! v2.13 "Lucky" Beta/7 W2K v5.0 Service Pack 4 Current version is 2.12.00 | 'Using TBUDL' information: http://www.silverstones.com/thebat/TBUDLInfo.html
Re: BayeIT Macros
On Tuesday, August 24, 2004 it appears that Zonnet wrote the following in reply to my comments re: "BayeIT Macros": Z> JR> I do use other filters to move obvious spam to my spam folder, Z> JR> thereby bypassing the BayesIT filter. Could that explain the Z> JR> difference between the real #s & the reported #s? Z> I don't think you are bypassing BasesIT. Baseyit works before TB! get its Z> hands on it, so it is BayesIT first and than your filters. Hi Gerard. If that's the case, then do you have any ideas to explain the difference between the BayesIT stats & the # of junk emails I have in my current file? TIA -- Jan Rifkinson Ridgefield CT USA TB! v2.13 "Lucky" Beta/7 W2K v5.0 Service Pack 4 Current version is 2.12.00 | 'Using TBUDL' information: http://www.silverstones.com/thebat/TBUDLInfo.html
Re: BayeIT Macros
Hello Michael L. Wilson, can you please cut down the size of your signature? 3 lines of your text: > I let BayesIT filter everything that is spam. I use no external > notifier, and no internal spam filters. Therefore, that could be the > problem followed by 20 lines of signature is a bit out of balance IMHO. Its nice that you like to play with TB's macros, but please forgive my ignorance, I'm not interested to get your personal spam statistics with each and every mail you send to this list. Thanks. -- Best regards, Alexander (http://www.neurowerx.de - ICQ 238153981) Whatever it is... I'm against it. -- Groucho Marx Current version is 2.12.00 | 'Using TBUDL' information: http://www.silverstones.com/thebat/TBUDLInfo.html
Re: BayeIT Macros
Hello Jan Rifkinson, 24-Aug-2004 18:33, you wrote: > Hi Gerard. If that's the case, then do you have any ideas to explain > the difference between the BayesIT stats & the # of junk emails I > have in my current file? Do you have your own filters that deal with junk email, too? If those aren't caught by BayesIt and you sort them to the same junk email folder, that would explain the difference. For a start, use your junk email to train BayesIt with them: mark all of the junk mails, right-click, specials, mark as junk. That should at least make a significant difference on the false negatives statistics: the percentage value that says "...guessed right 99.5% of the time" should drop a lot if BayesIt only detected 2 mails while in reality 95 mails were junk, ahem... :-) -- Best regards, Alexander (http://www.neurowerx.de - ICQ 238153981) Science may be described as the art of systematic oversimplification - the art of discerning what we may with advantage omit. -- Karl Popper Current version is 2.12.00 | 'Using TBUDL' information: http://www.silverstones.com/thebat/TBUDLInfo.html
Re: BayeIT Macros
On Tuesday, August 24, 2004 it appears that Alexander S. Kunz wrote the following in reply to my comments re: "BayeIT Macros": ASK> 24-Aug-2004 18:33, you wrote: >> Hi Gerard. If that's the case, then do you have any ideas to explain >> the difference between the BayesIT stats & the # of junk emails I >> have in my current file? ASK> Do you have your own filters that deal with junk email, too? Yes ASK> If those ASK> aren't caught by BayesIt and you sort them to the same junk email folder, ASK> that would explain the difference. Doesn't this depend on whether BayesIT screens before or after TB! filters -- that's one thing I'm trying to find out. If it's before, then the stats are dead wrong. If it's after that might explain some of the difference but see below. ASK> For a start, use your junk email to train BayesIt with them: mark all of ASK> the junk mails [snip] should at least ASK> make a significant difference on the false negatives statistics: the ASK> percentage value that says "...guessed right 99.5% of the time" should drop ASK> a lot if BayesIt only detected 2 mails while in reality 95 mails were junk, ASK> ahem... :-) OK, did as you suggested. Here are "new" stats: Spam Stats, last 24 hours (BayesIt! 0.5.11) Total Spam Emails: 2 Total Clean Emails: 186 BayesIT guessed right 99.5% of the time My email is 1.06383% spam Now what? TIA -- Jan Rifkinson Ridgefield CT USA TB! v2.13 "Lucky" Beta/7 W2K v5.0 Service Pack 4 Current version is 2.12.00 | 'Using TBUDL' information: http://www.silverstones.com/thebat/TBUDLInfo.html
Re: BayeIT Macros
ON Tuesday, August 24, 2004, 7:25:53 PM, you wrote: JR> Spam Stats, last 24 hours (BayesIt! 0.5.11) JR> Total Spam Emails: 2 JR> Total Clean Emails: 186 JR> BayesIT guessed right 99.5% of the time JR> My email is 1.06383% spam JR> Now what? Well, (2/(186+2))*100%= 1.0638% so that is correct. I am not sure how BayesIT calculates the "guessed right" statistic but I would assume that it looks at how many Spam or Ham msg you correct by using the "mark as (NOT) junk " function. The question is did you receive 188 email within the last 24 hours? Could it be 48 hours because you are using the wrong functions? -- Best regards, Gerard -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= Analyst - Psychiatric specialist who treats individuals suffering from the delusion that playing golf is a form of pleasure. Using The Bat! v2.12.00 on Windows XP 5.1 Build 2600 Service Pack 1 Current version is 2.12.00 | 'Using TBUDL' information: http://www.silverstones.com/thebat/TBUDLInfo.html
Re: BayeIT Macros
Hello Michael,
On 24 Aug 2004 at 20:47:10 -0700 GMT [05:47 CEST] you wrote:
MLW> My current macro set is
MLW> =
MLW> Spam Stats, last 24 hours (%BAYESITVERSION)
MLW> Total Spam Emails: %STATSPAMLETTERS(24)
MLW> Total Clean Emails: %STATNONSPAMLETTERS(24)
MLW> BayesIT guessed right %-
MLW> %calculate="100-(%STATSPAMERLETTERS(24)/%STATSPAMLETTERS(24))"%% of the time
You forgot to multiply by 100. Also, you are only looking at spam. But
bayesit is guessing for all mails not just spam.
%calculate="100-(100*((%STATSPAMERLETTERS(24)+%STATNONSPAMERLETTERS(24))/(%STATSPAMLETTERS(24)+%STATNONSPAMLETTERS(24"%%
MLW> My email is %-
MLW>
%Calc("100*%STATSPAMLETTERS(24)/(%STATSPAMLETTERS(24)+%STATNONSPAMLETTERS(24))")%% spam
MLW> =
--
Cheers,
Andre
"I am returning this otherwise good typing paper to you
because someone has printed gibberish all over it
and put your name at the top."
Current version is 2.12.00 | 'Using TBUDL' information:
http://www.silverstones.com/thebat/TBUDLInfo.html
Re: BayeIT Macros
On Tuesday, August 24, 2004 it appears that Zonnet wrote the following in regards to "BayeIT Macros": Z> ON Tuesday, August 24, 2004, 7:25:53 PM, you wrote: Z> JR> Spam Stats, last 24 hours (BayesIt! 0.5.11) Z> JR> Total Spam Emails: 2 Z> JR> Total Clean Emails: 186 Z> JR> BayesIT guessed right 99.5% of the time Z> JR> My email is 1.06383% spam Z> Well, (2/(186+2))*100%= 1.0638% so that is correct. Z> I am not sure how BayesIT calculates the "guessed right" statistic but I Z> would assume that it looks at how many Spam or Ham msg you correct by Z> using the "mark as (NOT) junk " function. Geez, why would I do that? I'd be busy most of my day. Or am I missing something? Z> The question is did you receive 188 email within the last 24 hours? Z> Could it be 48 hours because you are using the wrong functions? Actually I've rcvd 97 pieces of Junk Mail within the last 24 hours so the calculation seems to be double the actual #. And I don't know what wrong function you refer to? What do you mean? TIA -- Jan Rifkinson Ridgefield CT USA TB! v2.13 "Lucky" Beta/7 W2K v5.0 Service Pack 4 Current version is 2.12.00 | 'Using TBUDL' information: http://www.silverstones.com/thebat/TBUDLInfo.html
Re: BayeIT Macros
Hello Jan Rifkinson, 24-Aug-2004 19:25, you wrote: > Spam Stats, last 24 hours (BayesIt! 0.5.11) > Total Spam Emails: 2 > Total Clean Emails: 186 > BayesIT guessed right 99.5% of the time > My email is 1.06383% spam > Now what? Apparently, something is pretty wrong with the Bayesit stats. :-} [my opinion, the more I read about it here: BayesIt is a strange thing... apparently, it works... but do its programmers know why it works? No offense intended:-] -- Best regards, Alexander (http://www.neurowerx.de - ICQ 238153981) Whistler's Law: You never know who's right, but you always know who's in charge. Current version is 2.12.00 | 'Using TBUDL' information: http://www.silverstones.com/thebat/TBUDLInfo.html
Re: BayeIT Macros
ON Tuesday, August 24, 2004, 9:09:51 PM, you wrote: JR> And I don't know what wrong function you refer to? What do you JR> mean? Hi Jan, I have been playing with BayesIT functions myself. Here is a list I known is available: ,- [ ] | Spam Letters:%STATSPAMLETTERS(48) | Spam traffic: %STATSPAMTRAFFIC(48) | Stat no spam leters: %STATNONSPAMLETTERS(48) | Stat no spam Trafic: %STATNONSPAMTRAFFIC(48) | Stat spam letters: %STATSPAMERLETTERS(48) | Stat spam Traffic: %STATSPAMERTRAFFIC(48) | Stat no spam letters: %STATNONSPAMERLETTERS(48) | Stat no spam traffic: %STATNONSPAMERTRAFFIC(48) `- Do you know the difference between all these? I have not figured it out and if you use the wrong one you could get strange results. That is what I meant. At the moment non of these functions seem to work on this account? They all come up zero. -- Best regards, Gerard -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= You don't know what pressure is until you play for 5 bucks with only 2 bucks in your pocket. Using The Bat! v2.12.00 on Windows XP 5.1 Build 2600 Service Pack 1 Current version is 2.12.00 | 'Using TBUDL' information: http://www.silverstones.com/thebat/TBUDLInfo.html
Re: BayeIT Macros
Hello Jan Rifkinson, 24-Aug-2004 21:09, you wrote: Z>> I am not sure how BayesIT calculates the "guessed right" statistic but Z>> I would assume that it looks at how many Spam or Ham msg you correct by Z>> using the "mark as (NOT) junk " function. > Geez, why would I do that? I'd be busy most of my day. Or am I missing > something? If I were you, I'd disable your own filters that catch spam temporarily, and only train BayesIt for a while. After all, it is a *learning* filter that increases in precision the more you teach it (by telling it what is spam and what is not), and not a statistics tool. :-) -- Best regards, Alexander (http://www.neurowerx.de - ICQ 238153981) Our political leaders are unenlightened and corrupt, but with rare exceptions, political leaders have always been unenlightened and corrupt. (Alobar) Current version is 2.12.00 | 'Using TBUDL' information: http://www.silverstones.com/thebat/TBUDLInfo.html
Re: BayeIT Macros
On Tuesday, August 24, 2004 it appears that Zonnet wrote the following in regards to "BayeIT Macros": Z> ,- [ ] Z> | Spam Letters:%STATSPAMLETTERS(48) Z> | Spam traffic: %STATSPAMTRAFFIC(48) Z> | Stat no spam leters: %STATNONSPAMLETTERS(48) Z> | Stat no spam Trafic: %STATNONSPAMTRAFFIC(48) Z> | Stat spam letters: %STATSPAMERLETTERS(48) Z> | Stat spam Traffic: %STATSPAMERTRAFFIC(48) Z> | Stat no spam letters: %STATNONSPAMERLETTERS(48) Z> | Stat no spam traffic: %STATNONSPAMERTRAFFIC(48) Z> `- Z> Do you know the difference between all these? Nope :-( Z> I have not figured it out Z> and if you use the wrong one you could get strange results. Z> That is what I meant. Take a look @ Menu -> Options | Preferences | Plug-Ins Click plug-in click Information Scroll down to the bottom That's information that is made available by BayesIT via QT macro. I haven't tried them all, just the one that was posted to see what it would do -- not too much as it turns out. Sorry, I wish I knew what I was doing -- or even talking about for that matter. It just doesn't make sense to me unless TB! filters are second to BaeyesIT filters but no one has confirmed that. And even then, the numbers wouldn't make sense but at least I'd have a starting point. -- Jan Rifkinson Ridgefield CT USA TB! v2.13 "Lucky" Beta/7 W2K v5.0 Service Pack 4 Current version is 2.12.00 | 'Using TBUDL' information: http://www.silverstones.com/thebat/TBUDLInfo.html
Re: BayeIT Macros
On Tuesday, August 24, 2004 it appears that Alexander S. Kunz wrote the following in regards to "BayeIT Macros": ASK> If I were you, I'd disable your own filters that catch spam temporarily, and ASK> only train BayesIt for a while. After all, it is a *learning* filter that ASK> increases in precision the more you teach it (by telling it what is spam ASK> and what is not), and not a statistics tool. :-) Hi Alexander. Yes, of course you are right about your second poing but, just to be stubborn for one more moment, If BayesIT filters see msgs before TB! filters then this would not be necessary, right? And I can't get a definite answer to this question yet. -- Jan Rifkinson Ridgefield CT USA TB! v2.13 "Lucky" Beta/7 W2K v5.0 Service Pack 4 Current version is 2.12.00 | 'Using TBUDL' information: http://www.silverstones.com/thebat/TBUDLInfo.html
Re: BayeIT Macros
Hello Jan, On 24 Aug 2004 at 16:01:13 -0400 GMT [22:01 CEST] you wrote: JR> Yes, of course you are right about your second poing but, just to JR> be stubborn for one more moment, If BayesIT filters see msgs JR> before TB! filters then this would not be necessary, right? And I JR> can't get a definite answer to this question yet. The plugins come before the filters. -- Cheers, Andre "I don't suffer from insanity. I enjoy every minute of it." Current version is 2.12.00 | 'Using TBUDL' information: http://www.silverstones.com/thebat/TBUDLInfo.html
Re: BayeIT Macros
Hello Zonnet, On 24 Aug 2004 at 21:52:57 +0200 GMT [21:52 CEST] you wrote: Z> ,- [ ] Z> | Spam Letters:%STATSPAMLETTERS(48) Z> | Spam traffic: %STATSPAMTRAFFIC(48) Z> | Stat no spam leters: %STATNONSPAMLETTERS(48) Z> | Stat no spam Trafic: %STATNONSPAMTRAFFIC(48) Z> | Stat spam letters: %STATSPAMERLETTERS(48) Z> | Stat spam Traffic: %STATSPAMERTRAFFIC(48) Z> | Stat no spam letters: %STATNONSPAMERLETTERS(48) Z> | Stat no spam traffic: %STATNONSPAMERTRAFFIC(48) Z> `- Z> Do you know the difference between all these? I have not figured it out Z> and if you use the wrong one you could get strange results. Z> That is what I meant. Z> At the moment non of these functions seem to work on this account? Z> They all come up zero. They do what the names say. ER stands for error. -- Cheers, Andre "Wise men talk because they have something to say; fools, because they have to say something." Current version is 2.12.00 | 'Using TBUDL' information: http://www.silverstones.com/thebat/TBUDLInfo.html
Re: BayeIT Macros
On Tuesday, August 24, 2004 it appears that Andre Wichartz wrote the following in reply to my comments re: to "BayeIT Macros": AW> [snip] If BayesIT filters see msgs AW> JR> before TB! filters then this would not be necessary, right? And I AW> JR> can't get a definite answer to this question yet. AW> The plugins come before the filters. OK, Thank you. So, tonight, as an experiment, I am going to empty my junk file and de-activate all my active spam filters to see what happens over the next 24hrs. -- Jan Rifkinson Ridgefield CT USA TB! v2.13 "Lucky" Beta/7 W2K v5.0 Service Pack 4 Current version is 2.12.00 | 'Using TBUDL' information: http://www.silverstones.com/thebat/TBUDLInfo.html
Re: BayeIT Macros
ON Tuesday, August 24, 2004, 9:57:43 PM, you wrote: JR> Take a look @ Menu -> Options | Preferences | Plug-Ins JR> Click plug-in JR> click Information JR> Scroll down to the bottom Hi Jan, I finally found that the difference between spam letters and traffic. Letters is obvious but traffic is measured in bytes. That seems to me to be a rather arbitrary number. It should count all the spam words measured in bytes. I will stick with letters ;-) -- Best regards, Gerard -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= Before playing any shot, a golfer must lock her eyes and mind into the smallest possible target. Using The Bat! v2.12.00 on Windows XP 5.1 Build 2600 Service Pack 1 Current version is 2.12.00 | 'Using TBUDL' information: http://www.silverstones.com/thebat/TBUDLInfo.html
Re: BayeIT Macros
Hello Andre, Just checking my self following macro combination: Spam Stats, last 24 hours (%BAYESITVERSION) Total Spam Emails: %STATSPAMLETTERS(24) Total Clean Emails: %STATNONSPAMLETTERS(24) BayesIT guessed right %- %calculate="100-(%STATSPAMERLETTERS(24)/%STATSPAMLETTERS(24))"%% of the time My email is %- %calculate="100-(100*((%STATSPAMERLETTERS(24)+%STATNONSPAMERLETTERS(24))/(%STATSPAMLETTERS(24)+%STATNONSPAMLETTERS(24"%% spam Result: Spam Stats, last 24 hours (BayesIt! 0.5.11) Total Spam E-mail: 119 Total Clean Mails: 8 BayesIt guessed right 100% of the time My e-mail is 96.850394% spam That's impossible. I was getting around 300 Mails in the last 24 hours. Most of them Mailinglist messages, which are whitelisted in BayesIt. Isn't BayesIt counting whitelisted e-mail as good e-mail in the statistic. -- Ciao Thomas Using: TheBat! 2.12.00 and BayesIt! 0.5.11 System: Windows XP Build 2600 Service Pack 1 PGP:Key-ID: 0xD1ED3F0B Fingerprint: C1C8 7CEE 47A1 07A4 D5FD 8A7D AFB7 EC83 D1ED 3F0B Visit: TheBat! World on http://www.thebatworld.de pgpTpvcxkMLK5.pgp Description: PGP signature Current version is 2.12.00 | 'Using TBUDL' information: http://www.silverstones.com/thebat/TBUDLInfo.html
Re: BayeIT Macros
Hello Jan Rifkinson, 24-Aug-2004 22:01, you wrote: > Yes, of course you are right about your second poing but, just to be > stubborn for one more moment, If BayesIT filters see msgs before TB! > filters then this would not be necessary, right? And I can't get a > definite answer to this question yet. It doesn't matter wether BayesIt comes first or TB's filters come first when processing messages. If a message passes BayesIt as not spam (while it actually is spam) and your filter picks it up as being spam, how would BayesIt ever know the message was spam? Or what do you mean? -- Best regards, Alexander (http://www.neurowerx.de - ICQ 238153981) Is it not that Her Majesty's government infored our people of what they know about U.F.O.'s? I think it is time our people were told the truth. -- Brinsley Le Poer Trench, Earl of Clancarty Current version is 2.12.00 | 'Using TBUDL' information: http://www.silverstones.com/thebat/TBUDLInfo.html
Re: BayeIT Macros
On Wednesday, August 25, 2004 it appears that Alexander S. Kunz wrote the following in reply to my comments re: to "BayeIT Macros": >> Yes, of course you are right about your second poing but, just to be >> stubborn for one more moment, If BayesIT filters see msgs before TB! >> filters then this would not be necessary, right? And I can't get a >> definite answer to this question yet. ASK> It doesn't matter wether BayesIt comes first or TB's filters come first ASK> when processing messages. If a message passes BayesIt as not spam (while it ASK> actually is spam) and your filter picks it up as being spam, how would ASK> BayesIt ever know the message was spam? Or what do you mean? Hi Alexander. What I mean is if BayesIT filters come first & I de-activated all my own TB! filters, then I can see how accurate BayesIT is. -- Jan Rifkinson Ridgefield CT USA TB! v2.13 "Lucky" Beta/7 W2K v5.0 Service Pack 4 Current version is 2.12.00 | 'Using TBUDL' information: http://www.silverstones.com/thebat/TBUDLInfo.html
Re: BayeIT Macros
On Tuesday, August 24, 2004 it appears that Alexander S. Kunz wrote the following in regards to "BayeIT Macros": ASK> 24-Aug-2004 19:25, you wrote: >> Spam Stats, last 24 hours (BayesIt! 0.5.11) >> Total Spam Emails: 2 >> Total Clean Emails: 186 >> BayesIT guessed right 99.5% of the time >> My email is 1.06383% spam >> Now what? ASK> Apparently, something is pretty wrong with the Bayesit stats. :-} ASK> [my opinion, the more I read about it here: BayesIt is a strange thing... ASK> apparently, it works... but do its programmers know why it works? No ASK> offense intended:-] Hi Alexander. So as advertised, I de-activated all my TB! filters even though I understand BayesIT filters are supposed to do their work before the TB! filters. So, tonight I have 376 msgs in my junk folder. I marked none of them as junk, i.e. they were all junked automatically -- presumably by BayesIT. So here are my stats Spam Stats, last 24 hours (BayesIt! 0.5.11) Total Spam Emails: 5 Total Clean Emails: 155 BayesIT guessed right 99.4% of the time My email is 3.125% spam Anyone have any ideas beyond the notion that the macros don't work. Or put it another way, has anyone found that the macros do work & represent some real numbers? TIA -- Jan Rifkinson Ridgefield CT USA TB! v2.13 "Lucky" Beta/7 W2K v5.0 Service Pack 4 Current version is 2.12.00 | 'Using TBUDL' information: http://www.silverstones.com/thebat/TBUDLInfo.html
Re: BayeIT Macros
Hello Jan, On 26 Aug 2004 at 19:18:06 -0400 GMT [01:18 CEST] you wrote: JR> Anyone have any ideas beyond the notion that the macros don't JR> work. Or put it another way, has anyone found that the macros do JR> work & represent some real numbers? My junk folder shows 57 mails from the last 24 hours. Spam Stats, last day (BayesIt! 0.5.10) Total Spam: 67 Total Ham: 110 Error Spam: 0 Error Ham: 0 Accuracy: 100% My email is 37.853107% spam There is a difference of 10 mails but I get lots of dupes in my junk folder because of the duplicate mails on the server bug and I don't know how those count in the statistic. The macros give me the same information as under preferences. -- Cheers, Andre "Fliegen ist gar nicht so schwierig, wie man denkt. Man muß sich nur auf den Boden schmeißen und vergessen aufzuschlagen." Current version is 2.12.00 | 'Using TBUDL' information: http://www.silverstones.com/thebat/TBUDLInfo.html
Re: BayeIT Macros
Hello Jan Rifkinson, 26-Aug-2004 01:18, you wrote: ASK>> Apparently, something is pretty wrong with the Bayesit stats. :-} > So, tonight I have 376 msgs in my junk folder. I marked none of them as > junk, i.e. they were all junked automatically -- presumably by BayesIT. > So here are my stats Those 376 junk mails are a "total", correct? I mean... > Spam Stats, last 24 hours (BayesIt! 0.5.11) ... :-) > Anyone have any ideas beyond the notion that the macros don't work. Or > put it another way, has anyone found that the macros do work & represent > some real numbers? Try replacing the "(24)" hours in the macro call with a number large enough to give a "totals" value maybe? -- Best regards, Alexander (http://www.neurowerx.de - ICQ 238153981) A cookie store is a bad idea. Besides, the market research reports say America likes crispy cookies, not soft and chewy cookies like you make. -- Response to Debbi Fields' idea of starting Mrs. Fields' Cookies. Current version is 2.12.00 | 'Using TBUDL' information: http://www.silverstones.com/thebat/TBUDLInfo.html
Re: BayeIT Macros
On Thursday, August 26, 2004 it appears that Alexander S. Kunz wrote the following in regards to "BayeIT Macros": >> So, tonight I have 376 msgs in my junk folder. I marked none of them as >> junk, i.e. they were all junked automatically -- presumably by BayesIT. >> So here are my stats ASK> Those 376 junk mails are a "total", correct? Hi Alexander. yes. However, at this point I have 651. ASK> I mean... >> Spam Stats, last 24 hours (BayesIt! 0.5.11) ASK> ... :-) >> Anyone have any ideas beyond the notion that the macros don't work. Or >> put it another way, has anyone found that the macros do work & represent >> some real numbers? ASK> Try replacing the "(24)" hours in the macro call with a number large enough ASK> to give a "totals" value maybe? Changed to 48 hrs. Stats: Spam Stats, last 48 hours (BayesIt! 0.5.11) Total Spam Emails: 69 Total Clean Emails: 286 BayesIT guessed right 99.478261% of the time My email is 19.43662% spam Then for kicks I selected all msgs in junk file <651> and marked them as junk manually. BayesIT went thru training. Stats below. So I'll have to go with my conclusion until someone comes up with a definitive answer about BayesIT stats. Spam Stats, last 48 hours (BayesIt! 0.5.11) Total Spam Emails: 70 Total Clean Emails: 288 BayesIT guessed right 99.471429% of the time My email is 19.553073% spam Thanks, -- Jan Rifkinson Ridgefield CT USA TB! v2.13 "Lucky" Beta/7 W2K v5.0 Service Pack 4 Current version is 2.12.00 | 'Using TBUDL' information: http://www.silverstones.com/thebat/TBUDLInfo.html
Re: BayeIT Macros
Hello Jan Rifkinson, 26-Aug-2004 19:12, you wrote: ASK>> Those 376 junk mails are a "total", correct? > yes. However, at this point I have 651. Nearly 300 junk mails in 1 day? Well... even without correct statistics from BayesIt the other conclusion for me would be dropping that mail account. :-} -- Best regards, Alexander (http://www.neurowerx.de - ICQ 238153981) The release of atom power has changed everything except our way of thinking... the solution to this problem lies in the heart of mankind. If only I had known, I should have become a watchmaker. -- Albert Einstein Current version is 2.12.00 | 'Using TBUDL' information: http://www.silverstones.com/thebat/TBUDLInfo.html
Re: BayeIT Macros
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Hello Jan, on Wednesday, 25. August 2004, at 11:18:20 [GMT -0400] you wrote regarding "BayesIt Macros": > What I mean is if BayesIT filters come first & I de-activated all my > own TB! filters, then I can see how accurate BayesIT is. Following order is proofed: 1. AntiVirus Plugin 2. AntiSpam Plugin 3. Filters of the Filtermanager But there seams to be a other problem with the statistics. E-mail which BayesIt classifies as good Mail because of a match in the own whitelist are not counted as good mails in the statistic. They don't appear there. Other experiences? But how ever, the plugin works incredible good for what it is designed. Killing Spam. Good Job. But a good and correct statistic would be fine. - -- Ciao Thomas Using: TheBat! 2.13 "Lucky" Beta/6 and BayesIt! 0.5.11 System: Windows XP Build 2600 Service Pack 1 PGP:Key-ID: 0xD1ED3F0B Fingerprint: C1C8 7CEE 47A1 07A4 D5FD 8A7D AFB7 EC83 D1ED 3F0B Visit: TheBat! World on http://www.thebatworld.de -BEGIN PGP SIGNATURE- Version: PGP SDK 3.2.2 Comment: "Space to rent" iQA/AwUBQSz8LK+37IPR7T8LEQLlBACeMi9yUPSjjoyHVB7xo1Aq3H4zHFoAn3Go KzYqoPOyl99WG2H2MFkk9KM9 =wRVe -END PGP SIGNATURE- Current version is 2.12.00 | 'Using TBUDL' information: http://www.silverstones.com/thebat/TBUDLInfo.html
Re: BayeIT Macros
On Wednesday, August 25, 2004 it appears that Thomas Martin wrote the following in regards to "BayeIT Macros": TM> Following order is proofed: TM> 1. AntiVirus Plugin TM> 2. AntiSpam Plugin TM> 3. Filters of the Filtermanager Hi Thomas. Thank you for proofing. That answers that question. TM> But there seams to be a other problem with the statistics. E-mail TM> which BayesIt classifies as good Mail because of a match in the own TM> whitelist are not counted as good mails in the statistic. They don't TM> appear there. Other experiences? I don't think it correctly counts the spam mails that are caught by black list either. It may be counting every piece that is marked manually but otherwise I don't know where to go with this. But that's my suspicion. So maybe the problem is *both* blacklist *&* whitelist. That might explain the wide discrepancies in the figures. TM> But how ever, the plugin works incredible good for what it is TM> designed. Killing Spam. Good Job. But a good and correct statistic TM> would be fine. Agreed. Seems to be working ok. I suppose I should be grateful for little favors. -- Jan Rifkinson Ridgefield CT USA TB! v2.13 "Lucky" Beta/8 W2K v5.0 Service Pack 4 Current version is 2.12.00 | 'Using TBUDL' information: http://www.silverstones.com/thebat/TBUDLInfo.html
Re[2]: BayeIT Macros
On Tue 24 August 2004, 12:30:49 +1000, Drago wrote: >> if my total spam is 270 and total clean is 246, shouldn't my spam percentage >> be 51% - 52% since my overall spam percentage is greater than clean or is the >> 270 my complete mails, 246 clean and the 24 left over spam Might be a newbie question: Where can one find these statistics within TB? -- Gabor Current version is 2.12.00 | 'Using TBUDL' information: http://www.silverstones.com/thebat/TBUDLInfo.html
