Hallo Edvinas, On Tue, 6 Dec 2005 11:17:41 +0200GMT (6-12-2005, 10:17 +0100, where I live), you wrote:
>> Just now, just for the pleasure of testing it for you I deleted the >> whole directory tree where my copy of BayesIt stores its stuff. >> Subsequently I marked some messages as SPAM or HAM and BayesIt created >> not only the files, but also the directory tree.. EM> Well, it is not the case here. I was doing everything absolutely the same as EM> you describing, however BayesIt creates just the following directories and EM> files (output of dir *.* /s): Well, I got the same log entries as you, but when I mark messages as spam or ham I get this in my BayesIt log: : ------------------------------------------- : Training mode... Value is 0, no tokens! : ------------------------------------------- 06.12.2005 08:38:22 Preparing of autotraining: creation of 0000 folder and moving file with collected info into this folder. 06.12.2005 08:38:22 Preparing stage of autotraining is completed. 06.12.2005 08:38:25 This autotraining session will... 06.12.2005 08:38:25 change the 'non-spam' part of dictionary by: collected letters assigned to this part: 483 and the total counter of tokens: 203089 06.12.2005 08:38:25 change the 'spam' part of the dictionary by: collected letters assigned to this part: 940 and the total counter of tokens: 256299 Try pass number: 1 06.12.2005 08:38:26 Recalculation of 'spam' frequency dictionary... 06.12.2005 08:38:26 Recalculation of 'spam' completed. 06.12.2005 08:38:26 Recalculation of 'non-spam' frequency dictionary... 06.12.2005 08:38:26 Recalculation of 'non-spam' completed. 06.12.2005 08:38:26 Final checking of current recalculation... 06.12.2005 08:38:26 checking integrity of new 'non-spam' dictionary... 06.12.2005 08:38:26 Checking of this dictionary is completed. 06.12.2005 08:38:26 checking integrity of new 'non-spam' dictionary... 06.12.2005 08:38:26 Checking of this dictionary is completed. 06.12.2005 08:38:26 Refreshing main 'non-spam' dictionary by recalculated one. 06.12.2005 08:38:26 Refreshing main 'spam' dictionary by recalculated one. 06.12.2005 08:38:26 All autotraining processes completed. You need to train BayesIt, not with one or two messages, but with a couple of hundred, so that it gets a reasonable idea of what you consider spam. BayesIt will not work immediately out of the box and it never will. The EM> 06.12.2005 11:16:27 Dictionary didn't load. Continue in training mode... that you quoted from your log is BayesIt's way to ask for training -- Groetjes, Roelof User - a technical term used by computer pros. See idiot. The Bat! 3.63.06 (Beta) Windows XP 5.1 Build 2600 Service Pack 2 1 pop3 account, server on LAN OTFE enabled P4 3GHz 2 GB RAM
pgpR4xZjhY5Sj.pgp
Description: PGP signature
________________________________________________ Current version is 3.62.14 | 'Using TBUDL' information: http://www.silverstones.com/thebat/TBUDLInfo.html
