New beta 2.0.256 question... |
Post Reply 
|
| Author | ||||||||||||||||||
Kurt 
Guest Group 
|
 Post Options
 Thanks(0)
 Quote Reply
 Topic: New beta 2.0.256 question...Posted: 11 November 2003 at 4:43pm |
|||||||||||||||||
|
Hello, We installed the new beta on Sunday and have received 1108 spam and 734 good messages (according to the corpus.ini file). When does the new filter start working, I do not see any messages in the tblquarantine with the reject code of 14 (new bayesian filter)? Also just FYI our db.dat file is about 500k now. Thanks, Kurt |
||||||||||||||||||
 |
||||||||||||||||||
|
Desperado
Senior Member 
Joined: 27 January 2005 Location: United States Status: Offline Points: 1143 |
Post Options
Thanks(0)
Quote Reply
Posted: 11 November 2003 at 7:27pm |
|||||||||||||||||
|
The previous Beta (255) "Kicked In" at 500 / 500. I will assume this does also. Dan S. |
||||||||||||||||||
|
||||||||||||||||||
|
Logik!
Guest Group
|
Post Options
Thanks(0)
Quote Reply
Posted: 11 November 2003 at 9:52pm |
|||||||||||||||||
|
Well, mine hasn't kicked in yet either. I was just going to ask the same question: I'm at 1363/1281 and only the old filters are doing anything. |
||||||||||||||||||
|
||||||||||||||||||
|
LogSat
Admin Group 
Joined: 25 January 2005 Location: United States Status: Offline Points: 4104 |
Post Options
Thanks(0)
Quote Reply
Posted: 11 November 2003 at 11:28pm |
|||||||||||||||||
|
Kurt, The Bayesian filter begins to kick in after 500 good + 500 spam emails. However it's accuracy increases with time and database size. 500K is still quite small, it'll need to reach a few MBs in size to be more effective. Roberto F. |
||||||||||||||||||
|
||||||||||||||||||
|
Desperado
Senior Member
Joined: 27 January 2005 Location: United States Status: Offline Points: 1143 |
Post Options
Thanks(0)
Quote Reply
Posted: 12 November 2003 at 1:45am |
|||||||||||||||||
|
With a "Fresh Corpus, After an hour and 15 minute, I have 2049 Spam / 644 Good. My db.dat is at 1268 KB and I have started blocking. Example below:
Dan S.
|
||||||||||||||||||
|
||||||||||||||||||
|
Kurt
Guest Group
|
Post Options
Thanks(0)
Quote Reply
Posted: 12 November 2003 at 10:02am |
|||||||||||||||||
|
Thanks. I did notice this morning that 2 messages have been flagged as matching the bayesian filter. We are currently at 1516 SPAM, and 952 GOOD... so we just need to give the system more time to build the corpus db.dat file (only 568k now). From looking at the db.dat file it appears that each item counts about .4% so we would currently need about 225 matches in the file to get to the default catch rate of 90%. Is this logic correct, or how does spamfilter use the db.dat file? Kurt |
||||||||||||||||||
|
||||||||||||||||||
|
LogSat
Admin Group
Joined: 25 January 2005 Location: United States Status: Offline Points: 4104 |
Post Options
Thanks(0)
Quote Reply
Posted: 12 November 2003 at 10:52pm |
|||||||||||||||||
|
Kurt, We'd rather not divulge details on the format of the db.dat file and how we calculate the probabilities. We use Bayesian statistics, but can't say more... sorry! Robero F. |
||||||||||||||||||
|
||||||||||||||||||
Topic Options|
Post Reply
|
|
Tweet
|
| Forum Jump | Forum Permissions You cannot post new topics in this forum You cannot reply to topics in this forum You cannot delete your posts in this forum You cannot edit your posts in this forum You cannot create polls in this forum You cannot vote in polls in this forum |
This page was generated in 0.152 seconds.