Skip navigation, view page content (access key = C)

Begin OSU masthead and toolbar

The Ohio State University
www.osu.edu
  1. Help
  2. Campus map
  3. Find people
  4. Webmail

Ohio State University logo Office of Information Technology Technology Support Center (8help)

Obtain the Benefits of Low, Medium, and High in Anti-Spam while continuing to use Advanced Features



If an Anti-Spam user would like the benefits available in the Low, Medium, or High settings in the Simple Interface, but would also like to continuing using Whitelist and Blacklists special steps must be taken. For a general description of how to setup and use the system please refer to the article titled, "Anti-Spam Service: Getting Started - Turning on the Service"

Before attempting to adjust these Advanced Settings please be sure you are familiar with and have turned on the service. This article will focus just on the settings required to simulate the low, medium and high settings found on the simple interface, while continuing to use some advanced features.

To use the Advanced Settings to mimic the behavior of the Simplified Interface options, simply adjust your settings as follows for the level of spam protection you would like: Low, Medium, or High. The current default for students brought into the system is "Medium." You can reach these settings in antispam by clicking on "Preferences," then clicking on "Stream Settings."

    Stream Settings for HIGH
    • S-100 Automatically reject messages scoring more than this amount (5.0 to 2000) 6.5
    • S-200 Auto-reject messages scoring more than this amount without creating an incident (10.0 to 1000000) 10
    • S-300 Spam threshold (1.0 to 100) 5.0
    • S-500 Hold messages from hosts in administrator's real-time 'Hold' blacklists No
    • S-600 Reject messages from hosts in administrator's real-time 'Reject' blacklists Yes
    • S-1200 Only tag spam -- do not hold any messages Yes
    • S-1300 Convert any 'reject' settings to 'hold' in tag-only mode No
    • S-1600 Inherit rules from 'default' stream Yes
    • S-2300 Enable Bayesian analysis Yes
    • S-2310 Enable Bayesian training Yes
    • S-2500 Add links to messages to train Bayesian analyzer No
    • S-2700 Add training links in message headers No
    • S-3000 Score below which to auto-learn as non-spam (-1000.0 to 5) -0.5
    • S-3100 Score above which to auto-learn as spam (0 to 10000) 4.5 
    • Bayes Settings: Bayes Scoring Thresholds for HIGH
Percentage		Score
49			1
59			1
69			2
79			2
89			4
99			4
    Stream Settings for MEDIUM
    • S-100 Automatically reject messages scoring more than this amount (5.0 to 2000) 6.2
    • S-200 Auto-reject messages scoring more than this amount without creating an incident (10.0 to 1000000) 10.1
    • S-300 Spam threshold (1.0 to 100) 4.5
    • S-500 Hold messages from hosts in administrator's real-time 'Hold' blacklists No
    • S-600 Reject messages from hosts in administrator's real-time 'Reject' blacklists Yes
    • S-1200 Only tag spam -- do not hold any messages Yes
    • S-1300 Convert any 'reject' settings to 'hold' in tag-only mode No
    • S-1600 Inherit rules from 'default' stream Yes
    • S-2300 Enable Bayesian analysis Yes
    • S-2310 Enable Bayesian training Yes
    • S-2500 Add links to messages to train Bayesian analyzer No
    • S-2700 Add training links in message headers No
    • S-3000 Score below which to auto-learn as non-spam (-1000.0 to 5) -0.5
    • S-3100 Score above which to auto-learn as spam (0 to 10000) 5.5 
    • Bayes Settings: Bayes Scoring Thresholds for MEDIUM
Percentage		Score
0			0
59			0
69			0
79			0
89			0
99			0
    Stream Settings for LOW
    • S-100 Automatically reject messages scoring more than this amount (5.0 to 2000) 20.5
    • S-200 Auto-reject messages scoring more than this amount without creating an incident (10.0 to 1000000) 22
    • S-300 Spam threshold (1.0 to 100) 4.5
    • S-500 Hold messages from hosts in administrator's real-time 'Hold' blacklists No
    • S-600 Reject messages from hosts in administrator's real-time 'Reject' blacklists Yes
    • S-1200 Only tag spam -- do not hold any messages Yes
    • S-1300 Convert any 'reject' settings to 'hold' in tag-only mode No
    • S-1600 Inherit rules from 'default' stream Yes
    • S-2300 Enable Bayesian analysis Yes
    • S-2310 Enable Bayesian training Yes
    • S-2500 Add links to messages to train Bayesian analyzer No
    • S-2700 Add training links in message headers No
    • S-3000 Score below which to auto-learn as non-spam (-1000.0 to 5) -0.5
    • S-3100 Score above which to auto-learn as spam (0 to 10000) 23.9 
    • Bayes Settings: Bayes Scoring Thresholds for LOW
Percentage		Score
0			0
59			2
69			4
79			7
89			10
99			15
Setting the Bayes Scoring Threholds.
  1. From the expert interface, in the Blue Bar, click on "Rules"
  2. In the Orange bar click on "Bayes Settings"
  3. Wait while the system displays a table and a few settings, which shows the number of spam and non-spam votes that make up the Bayesian Database.
    - At some times during the day this takes a while to load
  4. Towards the bottom you will see two boxes, (Percentage and Score) and a "Submit Changes" button.
  5. To add a bayes threshold setting, type the corresponding percentage and score into those boxes and press the "Submit Changes" button.
    - For example Percentage = 99, Score = 15
    - Which means, if the Bayes scoring is 99% sure that a message is spam it should add 15 points to the basic score.
    - Between 89% and 98%, the score next to 89% will be used.



Current Record: 3191

Create Date: 09-22-2006
Last Reviewed: 09-26-2007
Please give us your feedback!
Was this document helpful?  




Home

 

return to top