Print Page | Close Window

Bayesian filter question

Printed From: LogSat Software
Category: Spam Filter ISP
Forum Name: Spam Filter ISP Support
Forum Description: General support for Spam Filter ISP
Printed Date: 19 July 2018 at 11:14pm

Topic: Bayesian filter question
Posted By: corymckee
Subject: Bayesian filter question
Date Posted: 18 September 2007 at 2:21pm
I read the following on the website
A slider is used to control the accuracy of the statistical filter. Incoming emails are assigned a probability of being Spam, ranging from 0% (most likely a valid email) to 100% (most likely Spam). Any emails that have a probability of being spam above the value you set will be rejected. Typical threshold values are in the 99.9% range.
SPF - Sender Policy Framework
So here is my question. The info above states that typical threshold values are in the 99.9% range. Does this mean the slider goes to the right near 99.9 or to the left near 0.1. Also I can not seem to get close to 0 (which i am assuming i should be near for the higest level of spam catching). The closest I can slide (without going to 0 which states will disable it) is 7.226
Please advise

Posted By: LogSat
Date Posted: 18 September 2007 at 3:23pm
Setting a threshold for the Bayesian filter lower than 80%-90% would cause too many false positives, so we do not recommend using any lower settings, definetly not below 5%, which is which the GUI control will not alow it.

As a side-note, the Bayesian filter is usually very, very selective. Most emails will either score with values in the 99% range, or in the 0-1% range. There is very seldom an email that will have a score that appears in the middle (20-30-40-50-60%). 

Roberto Franceschetti" rel="nofollow - LogSat Software" rel="nofollow - Spam Filter ISP

Posted By: corymckee
Date Posted: 18 September 2007 at 4:34pm
I would rather have false positives than letting spam through. If this is the case, is it set as it should be?

Print Page | Close Window