About Bayesian filter

The Bayesian filter is a statistical process to indentify a spam mail message. For details about the operating princial of the Bayesian filter, please consult the article A Plan For Spam written by Paul Graham.

Configure action against messages marked as Spam by the Bayesian filter

You can select one of the following actions on spam message detected by your Bayesian filter:

Bayesian filter learning engine configurations

The Bayesian filter needs training before it can accurately detect a spam message. At the beginning, you may get false negative or false postive result from the Bayesian filter ( Note: false postive means a Spam message was incorrectly treated as a Good message; false negative means a Good message was incorrectly treated as a Spam message).

You can define two different folders, one for storing Good message collection and the other for storing Spam message collection. On the server, there will be a tranining program that runs periodically to update your good and spam message list. The more sample message you have in these folders, the more accurate the result can be obtained by the Bayesian filter.

You can click the "Update Database Now" button to train your Bayesain filter immediately. Depends on the number of messages you have in the Good or Spam message folder, this process may takes seconds to minutes to complete.