|
I am currently running an eval version of the Spam Filter. I must say, so far it is working OK. I know there is alot of tweaking I need to do still, though.
Before I invest too much time, I would like to know the maximum number of domains the filter can successfully monitor. We currently have 191 domains, and growing. Is this too much of a burden for the filter?
Also, I have a question about the beyesian filter... It seems like the stats for some of the words are a little different than what I expected. (See below)
*Token ,Good, Spam, ProbSpam, ModDate
*!!!,12,0,9.99999974737875E-5,2/22/2005
*!!!!,6,0,9.99999974737875E-5,2/20/2005
*!!!!!,3,0,0.400000005960464,2/16/2005
*!!!!!!,1,0,0.400000005960464,2/16/2005
*$$$,14,0,9.99999974737875E-5,2/23/2005
*porn,23,3,0.279009759426117,2/23/2005
In this example, it stats the word "porn" was used in 23 legit emails? Currently, I dont have any word filters setup, so is the beyesian filter not smart enough yet? I thought the beyesian filter would "learn" from emails that were blocked by MAPS. If that were the case, I would assume the porn entry should be more like (*porn,3,23, etc...) Does that make sense?
Also, if you have a list of keywords that you could email, I would very much appreciate it. It woudl save me a ALOT of time. Thanks again! My email is jomaits(at)ogdennews.com
|