I sent a message the other day about the Statistical Filtering Settings and what is the best way to set this up. Someone said I have to train it for every user. I don t see how that is done in email. That would be a look of work for over 300+ users.
That's just one of the drawbacks of statistical filtering. The legitimate E-mail that you get it probably very different than the legitimate E-mail that other users on your server get. And, there's a good chance that the spam that you get is different than the spam that other users get.
As a simple example, say that one of your users works at a bank or is buying a home. They will likely get a lot of E-mail mentioning words like "mortgage". If their E-mail is lumped together with yours, that makes it more likely that you will get spam with the word "mortgage" in it.
That's why we never put any statistical filtering into Declude JunkMail (we were one of the first to experiment with the concept) -- for a server-based product, it can work reasonably well without full training, but in many cases the CPU usage outweighs the benefits. With full training, it would normally be effective enough to outweigh the CPU usage -- but few people do the full training.
Then, with intentional misspellings, using foreign characters that resemble English characters, and adding "good" words to the spam, it gets harder and harder to use statistical filtering to catch these.
-Scott
---
Declude JunkMail: The advanced anti-spam solution for IMail mailservers since 2000.
Declude Virus: Ultra reliable virus detection and the leader in mailserver vulnerability detection.
Find out what you've been missing: Ask for a free 30-day evaluation.
--- [This E-mail was scanned for viruses by Declude Virus (http://www.declude.com)]
To Unsubscribe: http://www.ipswitch.com/support/mailing-lists.html List Archive: http://www.mail-archive.com/imail_forum%40list.ipswitch.com/ Knowledge Base/FAQ: http://www.ipswitch.com/support/IMail/
