Bayesian Self-Learning - (Option) bind autolearn value to block score
Currently the situation is as follows:
Bayesian self-learning is the answer to the Bayesian filtering problem.
The scoring system in SpamAssassin is used for this purpose. The higher the SpamAssassin score, the more sure we are that the message is a spam. The lower the score, the more sure we are that it is ham. The following criteria defines whether or not SpamAssassin will train the Bayes database about a message:
If the total SpamAssassin score is above 12, and both the header score and body score are above 3, then train the Bayes database about the spam.
If the total SpamAssassin score is below 0.1, then train the Bayes database that it is not a spam.
The problem however is is that this value of 12 is very high and arbitrary. Whether or not an email reaches 12 depends very much on how you've got your other settings in your server defined. For instance you could blacklist and add 5 points which would make it easy to reach12. But if you don't an email will rarely reach 12 unless its a piss poor spammer.
My suggestion would be to bind this autolearn value to whatever is set as 'block value'. This is the defined border of spam anyways, so anything above can safely be 'auto-learned' as spam.
Initial field tests show promising results! Please give me feedback for this idea / suggestion
Robin Ashton commented
I like this idea and think it should be implemented. Even just being able to alter this would be an improvement, or even JUST lowering the score from 12 to 9.5, the default block score.