Spam Assassin's Bayesian filtering

[thedust2010]

I have a user whose Bayesian filtering eventually got way too specific and email was beginning to be marked spam that should not have. The crazy thing is this user NEVER logged in to Plesk and defined spam and ham emails. He just uses Outlook. So I'm wondering how spam and ham got defined over 800+ times without doing it from inside of Plesk. Does Spam Assassin still learn from your activities with your email once it's been downloaded via POP3 to your Outlook inbox?

 

[atinoco]

Wondering exactly the same thing

I enabled spamassassin for my email box a couple of weeks ago, and only defined whitelists and blacklist. There isn't any option for training SA in Plesk 7.5.6 for windows.

this week i noticed that bayesian filtering started working somehow, I guess autolearning its on, how does SA define spam and ham?

This its from one of the messages i got

<quote>X-Spam-Flag: YES
X-Spam-Checker-Version: SpamAssassin 2.64 (2004-01-11) on server242
X-Spam-Level: *****
X-Spam-Status: Yes, hits=5.5 required=5.0 tests=BAYES_99,HTML_MESSAGE
autolearn=no version=2.64
X-Spam-Report:
* 0.1 HTML_MESSAGE BODY: HTML included in message
* 5.4 BAYES_99 BODY: Bayesian spam probability is 99 to 100%
* [score: 1.0000]</quote>

I'd like to know how this works.

Thansk in Advance