Spam Bayes

Post by **Sylvus** » July 19, 2004, 2:33 pm

I recently started using Spam Bayes (a Bayesian Filter for spam) and I have nothing but praise for it. It's a It has a windows client that interfaces with Outlook and, through looking at new incoming messages as well as those you've already received, learns which emails you get are spam and which are not. Once it has a sufficient sample, it will automatically move spam into a junk email folder and possible spam into a suspected junk folder. Out of ~500 new emails on two different machines, I have yet to see it move a good email into the junk folder. I have seen it move one good email into the suspected junk.

I highly recommend it if you get a lot of spam. It has kept my inbox totally clean in the week I've been using it, and I just delete everything in my junk email folder whenever I think about it.

Windows Client

It works quite a bit better if you go through and separate your junk email and real email into separate folders before installing. Then it has quite a nice sampling of what is good and bad and can learn from it right away. Give it a shot if you like, it gets my seal of approval.

Winnow · Post by **Winnow** » July 19, 2004, 2:52 pm

Spam filters have come a long way. Cox recently has started offering a free spam filter. There are options to delete incoming spam or to filter the spam by adding --spam-- to the title of a suspected spam and then directing it to a spam folder. I chose the spam folder as I wanted to see the results. So far I haven't seen a legit email in the spam folder and maybe one spam email a week makes it into my mailbox undetected. Not too shabby.

If Cox hadn't started offering this, I'd be jumping all over your recommendation though.

Post by **Sylvus** » July 19, 2004, 2:56 pm

The only advantage I can see to Spam Bayes, and it's only because I do not know how Cox's filter works, is that it is up to you to decide what is spam and what is not. Say you have a crazy uncle who forwards you all the stupid emails that he gets every day (I do) and you started marking all of his emails as spam, Spam Bayes would learn that emails from crazy_uncle@aol.com (or whatever) was an address that you consider spam and would start moving things accordingly.

If Cox has the flexibility that allows its system to learn from your preferences, then I would recommend just sticking with that.

Winnow · Post by **Winnow** » July 19, 2004, 3:06 pm

Sylvus wrote:The only advantage I can see to Spam Bayes, and it's only because I do not know how Cox's filter works, is that it is up to you to decide what is spam and what is not. Say you have a crazy uncle who forwards you all the stupid emails that he gets every day (I do) and you started marking all of his emails as spam, Spam Bayes would learn that emails from crazy_uncle@aol.com (or whatever) was an address that you consider spam and would start moving things accordingly.

If Cox has the flexibility that allows its system to learn from your preferences, then I would recommend just sticking with that.

I have no crazy uncles! : ) I am probably considered the crazy uncle.

You can submit false positives to ThisIsNotSpam@cox.net and they will "fix the glitch" but it doesn't seem as easy as Spam Bayes.