[Spambayes] Personal statistics

[Ryan Gerry]

I have been using SpamBayes version 1.0a2 through procmail and cron on 
Solaris since May 12th and have been pleased with it.  I would like to 
contribute my statistics to the project.

Spam classification effectiveness from 5/12/2003 to present:
Total number of legitimate message received: 186
Total number of spam message received: 48
Total False Positives: 5
Total False Negatives: 2
Percent ham accuratly identified: 97.3100
Percent spam accuratly identified: 95.8300

Here is how large my ~/.hammiedb.dir and ~/.hammiedb.dat files were for 
a range of days.

DATE            .dat    .dir
05/19/03        8192    384
05/20/03        16384   392
05/20/03        16544   400
05/21/03        16640   408
05/22/03        16824   416
05/23/03        16960   424
05/24/03        17168   432
05/25/03        17304   432
05/26/03        17336   440
05/27/03        17568   448
05/28/03        17616   456
05/29/03        17864   472
05/30/03        22008   472
05/31/03        22008   480
06/01/03        10056   480
06/02/03        22008   488
06/03/03        22008   528

These file sizes are in kilobytes produced by du(1).  One thing to note 
is that I append to the file size file (above) in the same crontab right 
after I do the mboxtrain.py.  I think that mboxtrain.py may inflate the 
database size becuase if I do a:
$ du -k ~/.hammiedb.dat
right now I get a disk usage of only 11000 KB.

Hope this helps!
Ryan Gerry