Identifying SPAM in SmarterMail

[mcguiredesi]

How come when an email comes in and I identify it as SPAM over and over again the SPAM filter does not learn what SPAM is. What good does it do to identify a piece of email as SPAM if SmarterMail does not learn?

[bruce]

We do notuse Baysian filter because it is very ineffective and cause a lot of false positive.


SmarterMail uses SpamAssassin, blacklist, spf to identify spam.


Bruce

DiscountASP.NET
www.DiscountASP.NET (http://www.DiscountASP.NET)

[mcguiredesi]

OK,so does it do anything for me to identify a piece of email as SPAM? OR does deleting it do the same?

[mjp]

Identifying a message as spam will not have an effect, you can simply delete them.

In a system that uses Bayesian filtering to "train" itself, it is necessary to identify spam. But as Bruce suggested, Bayesian filtering is rarely effective anymore. Spammers began including normal valid text in their messages years ago to "poison" Bayesian spam filters.

You have probably received many spam messages that contain large chunks of text that appear to have been lifted from books or articles, but have nothing to do with the message the spammer wants to get across. When those kinds of messages are tagged in a Bayesian system, every word in the message becomes part of that database. So you can see how tagging a few hundred (or a few thousand) such messages effectively ruins anyusefulness the filter may have.

mjp
---
DiscountASP.NET