dspam for MailScanner

[Todd T. Fries]

Penned by R Wahyudi on 20090724 12:55.04, we have:
| Hi Todd,
| 
| How does dpsam performed compared to standard SpamAssassin Bayes ?
| I was trying to replace SpamAssassin Bayes years ago with dspam but
| didn't end up completing it
| 
| Can you give me details on your email statistics ( eg how many email
| you receive / second )
| and what sort of database hardware do you have ?
| 
| Regards,
| Rianto Wahyudi

At the time I switched to dspam, i couldn't figure out how to train
spamassassin.

Memory wise and cpu wise, dspam is much more lightweight.

Given I didn't know how to train spamassassin (I have since been told how,
but am reluctant to switch back even for testing) I found the training of
dspam to take a bit (as advertised) but once trained to get very accurate
very quickly.

In a few years of service, my personal stats are these:

todd at fries.net:
                TP True Positives:         235830
                TN True Negatives:         828592
                FP False Positives:          2665
                FN False Negatives:          1409
                SC Spam Corpusfed:              0
                NC Nonspam Corpusfed:           0
                TL Training Left:               0
                SHR Spam Hit Rate          99.41%
                HSR Ham Strike Rate:        0.32%
                OCA Overall Accuracy:      99.62%

I do have some mailing list archives, and training them is a little more
sporatic:

openbsd at email.fries.net:
                TP True Positives:           5725
                TN True Negatives:         226221
                FP False Positives:          5158
                FN False Negatives:           180
                SC Spam Corpusfed:              0
                NC Nonspam Corpusfed:           0
                TL Training Left:               0
                SHR Spam Hit Rate          96.95%
                HSR Ham Strike Rate:        2.23%
                OCA Overall Accuracy:      97.75%

.. but all in all, I've been very satisfied with dspam.

My father has quite a different set of stats, but it has also helped
him greatly:

tyrone at fries.net:
                TP True Positives:          26294
                TN True Negatives:           3791
                FP False Positives:            85
                FN False Negatives:          2144
                SC Spam Corpusfed:              0
                NC Nonspam Corpusfed:           0
                TL Training Left:               0
                SHR Spam Hit Rate          92.46%
                HSR Ham Strike Rate:        2.19%
                OCA Overall Accuracy:      93.10%

My hardware is old PATA interface disk serving a postgresql database that
has been tweaked a bit to perform well, and I do pruning/reindexing/etc every
two nights, not every night.

If you have the desire to help your spam filter be razor sharp and feel very
gratified by being able to help train it by giving feedback everytime it gets
a false negative (true mail marked as spam) or a false positive (true spam
not marked as spam) then dspam is really a good thing to use.

The rub comes in trying to get people to do it when they don't quite have the
above understanding, desire, or both.

| On Tue, Jul 21, 2009 at 9:38 PM, Glenn Steen<glenn.steen at gmail.com> wrote:
| > 2009/7/19 Todd T. Fries <todd at fries.net>:
| >> I've been using this for a few years now, and keep forgetting to
| >> contribute it back.
| >>
| >> This is my own work, I couldn't be more pleased if MailScanner took it
| >> and made the equivalent or better functionality in the default
| >> distribution.
| >>
| >> If I can polish it or whatever, please let me know, if it saves you
| >> work.
| >>
| > If Jules doesn't decide to include this... then put it all in the
| > wiki;)... After all, that's what it's there for...
| >
| > Cheers
| > --
| > -- Glenn
| > email: glenn < dot > steen < at > gmail < dot > com
| > work: glenn < dot > steen < at > ap1 < dot > se
| > --
| > MailScanner mailing list
| > mailscanner at lists.mailscanner.info
| > http://lists.mailscanner.info/mailman/listinfo/mailscanner
| >
| > Before posting, read http://wiki.mailscanner.info/posting
| >
| > Support MailScanner development - buy the book off the website!
| >
| -- 
| MailScanner mailing list
| mailscanner at lists.mailscanner.info
| http://lists.mailscanner.info/mailman/listinfo/mailscanner
| 
| Before posting, read http://wiki.mailscanner.info/posting
| 
| Support MailScanner development - buy the book off the website! 

-- 
Todd Fries .. todd at fries.net

 _____________________________________________
|                                             \  1.636.410.0632 (voice)
| Free Daemon Consulting, LLC                 \  1.405.227.9094 (voice)
| http://FreeDaemonConsulting.com             \  1.866.792.3418 (FAX)
| "..in support of free software solutions."  \  sip:freedaemon at ekiga.net
|                                             \  sip:4052279094 at ekiga.net
 \\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\
                                                 
              37E7 D3EB 74D0 8D66 A68D  B866 0326 204E 3F42 004A
                        http://todd.fries.net/pgp.txt