Re: More observation on Bayesit!
Hello Roman Katzer everyone else, on 21-Feb-2005 at 18:25 you (Roman Katzer) wrote: I set up K9 and had it at 95% accuracy in about a week's training. It has improved since. I resetted my K9 stats after 100 days (nice that it supports two stats), stat1, 105 days now, 98.5% accuracy (which includes the initial training were it produced a monster load of false positives during the first mail session, *g*) -- stat2, 5 days now, 100% accuracy. Mr. SANI BAKOYO, the bank Manager Of Africa Continentale Bank Lagos Nigeria ... would not like this. :-) :-) :-) -- Best regards, Alexander (http://www.neurowerx.de - ICQ 238153981) Deliplayer2 is playing: Brain Bokka by The Infinity Project from the album 'Experiments That Identify Change' Current version is 3.0.1.33 | 'Using TBUDL' information: http://www.silverstones.com/thebat/TBUDLInfo.html
Re: More observation on Bayesit!
On Wednesday, February 23, 2005, 13:01:57, Alexander S. Kunz wrote: Mr. SANI BAKOYO, the bank Manager Of Africa Continentale Bank Lagos Nigeria ... would not like this. :-) :-) :-) He doesn't. Would you _believe_ it, he even wrote me under a _false_ _name_ now! The world is bad... Roman -- It could have happened, so it must be true. Current version is 3.0.1.33 | 'Using TBUDL' information: http://www.silverstones.com/thebat/TBUDLInfo.html
Re: More observation on Bayesit!
On Mon, 21 Feb 2005 18:07:21 +, Marck D Pearlstone wrote: Attachments are allowed but only up to a list maximum of 25kb. TBBETA has a higher ceiling of 40kb. Thanks Marck, I stand corrected, and attached is my blacklist of RegEx filters for Bayesit. Feel free to try at your own risk!! -- Best regards Barry barryh'at'kentra'dot'co'dot'uk Using TheBat! version 3.0.1.33 and BayesIt! 0.8.0 Release Current version is 3.0.1.33 | 'Using TBUDL' information: http://www.silverstones.com/thebat/TBUDLInfo.html
Re: More observation on Bayesit!
On Mon, 21 Feb 2005 13:16:45 +0100, Dick H wrote: Hello Barry, On Monday, February 21, 2005 at 09:01:40 GMT + (which was 10:01:40 where I live), Barry wrote and made these valuable points on the subject of More observation on Bayesit!: Hi Well I'm still persevering! I've now transcribed 83 different regex rules over to Bayesit. I extracted them from my copy of MailWasher. Why all these problems and hassle if Spambayes can do all the work for you with just a little bit of training? Hi Dick, I can use Mailwasher Pro VERY successfully. I'm just trying to find an integrated solution that works. So far it hasn't, but I like the challenge. :-) And there's no doubt that for me the RegEx filters have improved performance greatly. I've no doubt that SpamBayes works very well, but it's yet another program that has to be run alongside the multitude of other 'utilities' that seem to clog up my taskbar on a regular basis. So I'll give Bayesit a month and if it doesn't work after that I'll go back to Mailwasher Pro, which does. -- Best regards Barry barryh'at'kentra'dot'co'dot'uk Using TheBat! version 3.0.1.33 and BayesIt! 0.8.0 Release Current version is 3.0.1.33 | 'Using TBUDL' information: http://www.silverstones.com/thebat/TBUDLInfo.html
Re: More observation on Bayesit!
On Mon, 21 Feb 2005 12:25:18 -0500, Roman Katzer wrote: Furthermore: I never got BayesIt to work. The fact that there doesn't seem to be documentation doesn't help. Even when the German magazine c't (highly respected) tested several mail programs and their spam filtering capabilities they couln't get it to work. In other tests that I've read BayesIt seemed to always be worse than other Bayes-based competitors. Hi Roman, See my reply to Dick. I used to use K9 with great success, but for me Mailwasher Pro is the top dog! It seems odd to me that some people seem to claim great things of Bayesit, yet for others, like me, it doesn't seem to work. Similarly great claims are made for Bayes Filter, and yet, for me, it just doesn't work. I agree with your comment re the lack of documentation. I'll continue my experiment. -- Best regards Barry barryh'at'kentra'dot'co'dot'uk Using TheBat! version 3.0.1.33 and BayesIt! 0.8.0 Release Current version is 3.0.1.33 | 'Using TBUDL' information: http://www.silverstones.com/thebat/TBUDLInfo.html
Re: More observation on Bayesit! - blacklist.txt (1/1) uuEncoded 8061 bytes
On Tue, 22 Feb 2005 08:37:27 +, Barry wrote: On Mon, 21 Feb 2005 18:07:21 +, Marck D Pearlstone wrote: Attachments are allowed but only up to a list maximum of 25kb. TBBETA has a higher ceiling of 40kb. Thanks Marck, I stand corrected, and attached is my blacklist of RegEx filters for Bayesit. Feel free to try at your own risk!! Hmmm? Well, it was attached when it left me! -- Best regards Barry barryh'at'kentra'dot'co'dot'uk Using TheBat! version 3.0.1.33 and BayesIt! 0.8.0 Release begin 644 blacklist.txt M#0HC6S%=(%9I86=R82!;4UT-E-U8FIE8W0@;6%T8VAER`Z=EM5U]=S`L M,GU;:3%=6UQ77UU[,PR?5MA0%U;7%=?77LP+#)]9UM5U]=S`L,GUR6UQ7 M7UU[,PR?5MA0%T-@T*(ULQ72!!9]3%M($UAFMEB`H4RD-E-U8FIE M8W0@;6%T8VAER`Z061V/SI\7%=!9'95WQ061V7%-@T*(ULR72!!9]3 M%M($UAFMEB`H4RD-E-U8FIE8W0@;6%T8VAER`Z*%QS?%PN*7LU+'TN M*EQW*R0-@T*(ULS72!!9]3%M($UAFMEB`H4RD-E-U8FIE8W0@;6%T M8VAER`Z7'M7'T-@T*(ULT72!!9]3%M($UAFMEB`H4RD-E-U8FIE M8W0@;6%T8VAER`Z*%QS?%PN*7LQ,Q])`T*#0HC6S5=([EMAIL PROTECTED] M:V5R(A3*0T*4W5B:F5C=!M871C:5S(#I-97-S86=E+4ED+BLH36%C:#$P M?%)!1ELP+3E=*0T*#0HC6S9=([EMAIL PROTECTED]:V5R(A3*0T*4W5B:F5C M=!M871C:5S(#I;7'=?%U[,34L?20-@T*(ULQ72!!9'5L=`H4RD-E-U M8FIE8W0@;6%T8VAER`Z6%A8?A!9'5L='Q-871UF4I($-O;G1E;G1\061U M;'1S/R!/;FQY?$YO($UI;F]RWQ5;F-E;G-OF5D#0H-B-;,[EMAIL PROTECTED];'0@ M*%,I#0I3=6)J96-T(UA=-H97,@.BA2;W5G:'Q(;W1\55E;BDN*B!!9'5L M='Q397A\0F%R96QY($QE9V%L(%1E96Y\5V5T(%1E96Y\55E;B`H06-T:6]N M?%-W965T:5AG1\3W)G:65S*7Q:71C:'Q%V-OG1S?A?%QS*4-U;[EMAIL PROTECTED] M?%QS?TI?$,P8VM\45N6VDQ;%US?$QEV)I86X-@T*(ULS72!!9'5L=`H M4RD-E-U8FIE8W0@;6%T8VAER`Z4VQU='-\1EMU7I+5]^7EUC:WQ0=7-S M7Q0=7-S:65S?A3=6-K?$)I9WQ-87-S:79E?$QAF=E*2!#;V-K?%9IF=I M;EQ7?$AAF0M/V-OF5\0W5MVAO='Q5T-U;[EMAIL PROTECTED]5S8F\@ M4VET97Q;]W(#]*;V)\4V5X(%9I95OWQ(;W0H=5S=D_(XJ4%MI,6Q= M8W,-@T*(ULT72!!9'5L=`H4RD-E-U8FIE8W0@;6%T8VAER`Z*#\M:2E! M1%5,5'Q#3T-+4PT*#0HC6S5=($%D=6QT(A3*0T*4W5B:F5C=!M871C:5S M(#HH3G5D97Q.86ME9'[EMAIL PROTECTED](XJ4EC?$)AF5L2`H35G86Q\061U;'0I M?A7971\1')I'!I;F=\2]T*2`H55E;G-\35S8FEA;G,I#0H-B-;-ET@ M061U;'[EMAIL PROTECTED],I#0I3=6)J96-T(UA=-H97,@.D%D=6QT($QI;FMS?%-E`H M4VAO=WQ3:71E*7Q0;W)N(A-;W9I97-\1FQI8VLI?$UAG)I960N*T)U=XK M3]N96QY?$AO=!-;VUS?%-E'[EMAIL PROTECTED]AO=]\5V5T(%-E`T*#0HC6S==($%D M=6QT(A3*0T*4W5B:F5C=!M871C:5S(#I(;W)N2XK*$AO=7-E=VEV97-\ M55E;G,I?A.86ME9'Q.=61E*2`H35S8FEA;G-\5V]M96Y\55E;G-\1VER M;',I#0H-B-;[EMAIL PROTECTED];'[EMAIL PROTECTED],I#0I3=6)J96-T(UA=-H97,@.D%N:6UA M;',N*RA':7)LWQ7;VUE;BE\*$=IFQS?%=O;65N*2XK06YI;6%LPT*#0HC M6SE=($%D=6QT(A3*0T*4W5B:F5C=!M871C:5S(#I4:6=H=AEW0I/R`H M4'5SWE\5F%G:6YA*7Q$:6-KWQ66VEL,5UR9UMI;#%=;G-\4%LP0%UR;GQ, M6V\P75ML,5U;:6PQ771A?$=A;F=;(U=/V)A;F=\($1O;F=\2%R9,PF5\ M2#!R;GE\0W5N='Q4=V%T?$=IC%S?%QB5ETS]8GQ!W-H;VQE#0H-B-; M,[EMAIL PROTECTED];'[EMAIL PROTECTED](I#0I;V1Y(UA=-H97,@.BA(;W1\5V5T*2!4965N?$)A MF5L2`H35G86Q\061U;'0I#0H-B-;,[EMAIL PROTECTED];'[EMAIL PROTECTED](I#0I;V1Y(UA M=-H97,@[EMAIL PROTECTED][EMAIL PROTECTED]:2E\4VET92!396-R971S?$%D=6QT(A3 M:71E(D_3EN:W,-@T*(ULS72!!9'5L=`H0BD-D)O9'D@;6%T8VAER`Z M7%=#=6U5WQ(;[EMAIL PROTECTED]@5V]M96Y\1V%N9R`_0F%N9WQ=6-K?$1I;1O MPT*#0HC6S1=($%D=6QT(A*0T*0F]D2!M871C:5S(#I0:]N92!397A\ M36%R[EMAIL PROTECTED]XM74)[EMAIL PROTECTED]XM74QO;F5L0T*#0HC6S5=($%D=6QT(A M*0T*0F]D2!M871C:5S(#I9;W4NS$L,35]719`NS$L,C!]5[EMAIL PROTECTED] M97(-@T*(ULV72!!9'5L=`H0BD-D)O9'D@;6%T8VAER`Z:'1T',_.B\O M+BIX'@N*R0-@T*(ULQ72!860@(D9R;VTB(%M70T*1G)O;2!N;W0@;6%T M8VAER`Z6UQW+BU=*T`H6UQW+5TK7XI*UM!+5I=S(L-'T-@T*(ULQ72! M860@)[EMAIL PROTECTED]($AE861EB!;2%T-DAE861EB!M871C:5S(#I8+4)U;M%;6%I M;'Q8+5)E;6]V86P-@T*(ULR72!860@)[EMAIL PROTECTED]($AE861EB!;2%T-DAE861E MB!M871C:5S(#I8+5-E;F1EBXJ3F5WR!F5A:V5R#0H-B-;,[EMAIL PROTECTED] M(=8)R!(96%D97(@6TA=#0I(96%D97(@;6%T8VAER`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
More observation on Bayesit!
Hi Well I'm still persevering! I've now transcribed 83 different regex rules over to Bayesit. I extracted them from my copy of MailWasher. This seems to have made a dramatic improvement in spam catching. I've now reduced the spam failures from 100% (really!) to below 54% in two days. Looking at the logs it appears that the rules catch the majority of offenders.. If anyone else would like to try my rule set let me know and I'll email the file over. -- Best regards Barry Using TheBat! version 3.0.1.33 and BayesIt! 0.8.0 Release Current version is 3.0.1.33 | 'Using TBUDL' information: http://www.silverstones.com/thebat/TBUDLInfo.html
Re: More observation on Bayesit!
Hello Barry, A reminder of what Barry on TBUDL typed on: 21 February 2005 at 10:05:26 GMT +0100 B If anyone else would like to try my rule set let me know and I'll email the B file over. Yes please Barry. -- Best regards,Tony. _ Message composed on 21/02/2005 at 09:08 UTC 2004 - AWB Using The Bat! v3.0.2.10 on :gentoo: www.gentoo.org PGP Key http://www.theboomclan.com/pgp.txt smime.p7s Description: S/MIME Cryptographic Signature Current version is 3.0.1.33 | 'Using TBUDL' information: http://www.silverstones.com/thebat/TBUDLInfo.html
Re: More observation on Bayesit!
ON Monday, February 21, 2005, 10:01:40 AM, you wrote: B If anyone else would like to try my rule set let me know and I'll email the B file over. Hi Barry, I would love to try them. How much do the slow down bayesIt? -- Best regards, Gerard -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Golf! You hit down to make the ball go up. You swing left and the ball goes right. The lowest score wins. And on top of that, the winner buys the drinks. Using The Bat! v3.0.1.33 on Windows XP 5.1 Build 2600 Service Pack 2 Current version is 3.0.1.33 | 'Using TBUDL' information: http://www.silverstones.com/thebat/TBUDLInfo.html
Re: More observation on Bayesit!
On Mon, 21 Feb 2005 11:47:44 +0100, Gerard wrote: ON Monday, February 21, 2005, 10:01:40 AM, you wrote: B If anyone else would like to try my rule set let me know and I'll email the B file over. Hi Barry, I would love to try them. How much do the slow down bayesIt? Can you confirm your email addresses as the Gmane server confuses them... -- Best regards Barry [EMAIL PROTECTED] Using TheBat! version 3.0.1.33 and BayesIt! 0.8.0 Release Current version is 3.0.1.33 | 'Using TBUDL' information: http://www.silverstones.com/thebat/TBUDLInfo.html
Re: More observation on Bayesit!
Hello Barry! On Monday, February 21, 2005, 3:01 AM, you wrote: Barry Hi Barry I've now transcribed 83 different regex rules over to Bayesit. I extracted Barry them from my copy of MailWasher. Barry If anyone else would like to try my rule set let me know and I'll email the Barry file over. Send it to my email address below please. Thank you. -- IMHO Have A Happy, Ronmailto:[EMAIL PROTECTED] :usflag: :Ron_Cameron: :Flag-US-IA: Using :tbflaga: The Bat! v3.0.2.10 :beta: on Windows XP 5.1 Build on a Pentium III 700 MHz with 768 MB Percentage of spam email in the last 24 hours = 10 Current version is 3.0.1.33 | 'Using TBUDL' information: http://www.silverstones.com/thebat/TBUDLInfo.html
Re: More observation on Bayesit!
On Mon, 21 Feb 2005 05:04:35 -0600, Ron Cameron wrote: Hello Barry! Send it to my email address below please. You'll have to drop me a line Ron, Gmane hides your identity very well! -- Best regards Barry barryh'at'kentra'dot'co'dot'uk Using TheBat! version 3.0.1.33 and BayesIt! 0.8.0 Release Current version is 3.0.1.33 | 'Using TBUDL' information: http://www.silverstones.com/thebat/TBUDLInfo.html
Re: More observation on Bayesit!
Hello Barry, A reminder of what Barry on TBUDL typed on: 21 February 2005 at 12:30:15 GMT +0100 B You'll have to drop me a line Ron, Gmane hides your identity very well! Why not post it to this list, that way everyone who uses Bayesit can try it and you won't have to send it a dozen times. -- Best regards,Tony. _ Message composed on 21/02/2005 at 12:19 UTC 2004 - AWB Using The Bat! v3.0.2.10 on :gentoo: www.gentoo.org PGP Key http://www.theboomclan.com/pgp.txt smime.p7s Description: S/MIME Cryptographic Signature Current version is 3.0.1.33 | 'Using TBUDL' information: http://www.silverstones.com/thebat/TBUDLInfo.html
Re: More observation on Bayesit!
On Mon, 21 Feb 2005 at 09:01:40 +, Barry wrote: B Hi B Well I'm still persevering! B I've now transcribed 83 different regex rules over to Bayesit. I extracted B them from my copy of MailWasher. B This seems to have made a dramatic improvement in spam catching. I've now B reduced the spam failures from 100% (really!) to below 54% in two days. B Looking at the logs it appears that the rules catch the majority of B offenders.. B If anyone else would like to try my rule set let me know and I'll email the B file over. I'd be happy of a copy of the ruleset. Might see if I can apply it to Bayes Filter as well. -- Best regards, David See kludges for my pgp key. Using The Bat! v3.0.2.10 on Windows XP 5.1 Build 2600 Service Pack 2 Current version is 3.0.1.33 | 'Using TBUDL' information: http://www.silverstones.com/thebat/TBUDLInfo.html
Re: More observation on Bayesit!
On Mon, 21 Feb 2005 12:20:31 +, Tony Boom wrote: Hello Barry, A reminder of what Barry on TBUDL typed on: 21 February 2005 at 12:30:15 GMT +0100 B You'll have to drop me a line Ron, Gmane hides your identity very well! Why not post it to this list, that way everyone who uses Bayesit can try it and you won't have to send it a dozen times. I thought attachments were not allowed. Marck's slapped my wrist before! -- Best regards Barry barryh'at'kentra'dot'co'dot'uk Using TheBat! version 3.0.1.33 and BayesIt! 0.8.0 Release Current version is 3.0.1.33 | 'Using TBUDL' information: http://www.silverstones.com/thebat/TBUDLInfo.html
Re: More observation on Bayesit!
Hello Barry, On Monday, February 21, 2005 at 09:01:40 GMT + (which was 10:01:40 where I live), Barry wrote and made these valuable points on the subject of More observation on Bayesit!: Hi Well I'm still persevering! I've now transcribed 83 different regex rules over to Bayesit. I extracted them from my copy of MailWasher. Why all these problems and hassle if Spambayes can do all the work for you with just a little bit of training? -- Best regards, Dick ___ Those are my principles. If you don't like them I have others.-- Groucho Marx Using The Bat! v3.0.2.10 on Windows XP 5.1 Build 2600 Service Pack 1 Current version is 3.0.1.33 | 'Using TBUDL' information: http://www.silverstones.com/thebat/TBUDLInfo.html
Re: More observation on Bayesit!
Hello Barry, I've now transcribed 83 different regex rules over to Bayesit. I extracted them from my copy of MailWasher. This seems to have made a dramatic improvement in spam catching. I've now reduced the spam failures from 100% (really!) to below 54% in two days. this seems to be a great deal of work just to filter spam. I set up K9 and had it at 95% accuracy in about a week's training. It has improved since. Furthermore: I never got BayesIt to work. The fact that there doesn't seem to be documentation doesn't help. Even when the German magazine c't (highly respected) tested several mail programs and their spam filtering capabilities they couln't get it to work. In other tests that I've read BayesIt seemed to always be worse than other Bayes-based competitors. The upshot of this: No BayesIt fo me!!11 Roman -- We are on the verge: Today our program proved Fermat's next-to-last theorem. Current version is 3.0.1.33 | 'Using TBUDL' information: http://www.silverstones.com/thebat/TBUDLInfo.html
Re: More observation on Bayesit!
Dear Barry, @21-Feb-2005, 17:01 Barry [B] in mid:[EMAIL PROTECTED] said: Why not post it to this list, that way everyone who uses Bayesit can try it and you won't have to send it a dozen times. B I thought attachments were not allowed. Marck's slapped my wrist before! Attachments are allowed but only up to a list maximum of 25kb. TBBETA has a higher ceiling of 40kb. -- Cheers -- //.arck D Pearlstone -- List moderator and fellow end user TB! v3.0.2.10 on Windows NT Clone 5.2.3790 ' pgpGRhF4XcbiY.pgp Description: PGP signature Current version is 3.0.1.33 | 'Using TBUDL' information: http://www.silverstones.com/thebat/TBUDLInfo.html