On 9/21/2016 12:28 PM, Thomas Barth wrote:



Am 21.09.2016 um 18:00 schrieb li...@rhsoft.net:


the problem of the OP is that he starts things the other side round and
first reject without good evidence and don't have anything to make the
system bullet profe because it's rejected

I remembered that I read a book about Postfix with the topic "Training with SpamAssassin". And the author was against additional training. The more you train the worst the result. With the motto "I cook an egg for more than 15 minutes, but it is still hard." They re other arguments for not autolearning, but my english is not that good to translate a complete chapter. And if there are some mails breaking through the wall, than it is better to create rules against the header. Clear facts without side effects.

That is ridiculous. The more training bayes gets the better it works. And manual training is better than autolearning because autolearning can automatically learn false positives and false negatives and cause problems for the database.

If you are getting a bunch of similar spams coming through, you can definitely create custom rules, but you are going to be making new rules constantly. A good bayes database can automatically block some of those spams so you don't have to do as much work creating new rules for every new spam campaign.

He also wrote that Amavis/SpamAssassin is learning itself. Each mail classified as spam with a score of more than 12.0 is learned as spam and there should be a logfile entry with loglevel 2 if a mail has been learned as spam. I never increased the loglevel to check that.

SpamAssassin does autolearning by default, but it can make mistakes. It assumes that all high-scoring mail is spam and all low-scoring mail is ham. This is not necessarily true. You need to supplement this with at least enough manual learning to re-train the mistakes.

I followed his opinion because it is the best book I ve got (www.postfix.de, next SpamAssassin/Amavis training course in November, I m thinking of participation)

There are a lot of people on this list who have been using SA for a long time (about 12 years for me). You will sometimes get differing advice, but that is a good thing. I would much rather believe the consensus of a bunch of people with experience than one guy who decided to write a book.

--
Bowie

Reply via email to