From: Doug T. <dtr...@hf...> - 2008-08-20 19:58:27
|
Running Assp 2.0.0(5.10) on Win2k3 Server running Perl 5.8.8.822. Just copied my corpus (20,000 spam, 20,000 notspam, plus some errors) to my new test server and built spamDB for the first time. It ran movetonum and paired my numbers down to 14,000 and 14,000 (forgot to change the default to my higher number) and then built the file. In the "Last SpamDB Rebuild" listing I have a part I don't understand: ************************************************ Saving rebuilt SPAM database...done Resulting file 'd:\assp/spamdb.rb.tmp' is 6,136,201 bytes *Bayesian Pairs: new - 269,263 : altogether - 1* <<<<<< (I had one record in the spamDB before rebuilding, shouldn't the 'altogether' be a sum of that previous one plus all the new?) cleaning old Spamdb.helo records done - cleaning old Spamdb.helo records HELO Blacklist: new -405 : altogether - 283/512 HELOs Spam Weight: 7,021,080 Not-Spam Weight: 6,633,993 Corpus norm: 1.0583 Total processing time: 770 second(s) ************************************************ Subsequent runs had the processing time drop to 636 seconds but had everything else show up the same. Thanks for looking, Doug Traylor PS. Thanks to all you guys continuing to work on this great project. |