SpamSieve Since 2.0
Posted September 26th, 2003 @ 12:54am by Erik J. Barzeski
Since the 2.0 release, SpamSieve's stats on my system are:
Good Messages.......655 Spam Messages.......2317 False Positives.....25 False Negatives.....3 Correct.............99.1%
Not bad. The highest I got on the previous version was about 93%, which still meant I had to deal with a lot of spam. The false positives number is disconcerting, but I started with a blank whitelist, and as that grows to allow messages from various folks, the accuracy will tend upwards.
Donate Life
Posted 26 Sep 2003 at 9:22am #
I wish I got spam as well so I can try all these nice spam filters... Somehow, the spammers must have missed me.
Well, I hid a honeypot email address in a HTML comment on my site. Let's see how long it takes for it to be discovered.
Posted 26 Sep 2003 at 10:57am #
Erik: the high false positives number is likely due to having a lot more spam messages than good ones in your corpus. In SpamSieve 1.x, this was helpful for catching more spam. In 2.x, it's probably best not to have more than twice as many spam messages as good ones.
Posted 26 Sep 2003 at 2:06pm #
SpamSieve
Erik posted his SpamSieve 2.0 stats, and I'm equally impressed. Still getting more false positives than I like, but I'm sure that will improve as I'm diligent about telling SpamSieve what it's doing wrong through the scripts. What I love...
Posted 11 Oct 2003 at 11:27pm #
Who in the hell opens messages that look like this? This is a screenshot of the subject column of my spam folder. If people didn't...
Posted 20 Jan 2004 at 6:39am #
Ok, I get no spam at all, here's why.
1. I don't post my email address to sites that make it public, including RSS/BLOG/Webcrossing/etc. and Usenet, IM profiles, etc.
2. I do put my email address in the signature that I use in Mail.app, but I only send email to real people from there, not mailing lists.
3. I do subscribe to Yahoo! and MSN Groups regularly, however they have their own webmail that I never bother to check but once each 6 months or so.
4. I use Javascript on my website to hide my email address but still make it link-able. for my Resume I'm switching to PDF exclusively with a Word document available upon request. I tried munging, doesn't work well, newbies don't know how to delete the .removethispart. for some reason in their new mail messages. I tried entity encoding and it actually made me a target for web crawlers, wound up having to get a new shell account because of that.
5. I love freeshell (sdf.lonestar.org) because they offer shell for the life of their service for just $1.00 or €5.00 with web (15MB) and gopher (15MB) hosting, email (15MB), webmail access, and of course really great BSD shell access.