The Bayesian spam-rejection filter gets "trained" in the normal use of Emmesmail, repardless of whether the Bayesian filter is used or not. That is, the valid-email database gets updated whenever you save a valid email and the spam database gets updated whenever you classify an email delivered to you as spam by clicking the "Spam" button at the bottom of the Read Action Console. By clicking the "Spam" button, the spam email also is automatically deleted from your mailbox.
Whenever Bayesian filtering is being used, emails classified as spam by the filter also are added to the spam database.
The spam mailbox periodically should be checked in order to check for valid emails mistakenly classified as spam. A listing of the emails in the spam mailbox may be viewed by clicking "View Spam" from the "Choose User" menu. You may select an email by subject and then click OK. This will provide the name of sender, additional information to help to decide whether the message is really spam or not. You then have five options. If you find an email you think may be a valid one, you may read it by clicking the "Read" button (first click OK) to be sure. If it is indeed valid, you should click OK again, and then click the "Switch corpi" button. This will delete the email from the spam mailbox and from the spam database and place it in the appropriate user's mailbox and the valid-email database. Emmesmail will exit after this operation, which can take a minute or so since the databases are large. If after checking the spam emails, you wish to do nothing, click the "Done" button.
After checking the spam emails, you may delete any one of them by selecting it and pressing the "Delete" button, or you may remove all the spam emails from the spam mailbox by hitting "Del. All".
We recommend against reading spam emails, no matter how curious you are. Many spam emails are configured so that information about which email address read it is relayed to the spammer. You can be certain that once the spammer knows you have shown even the slightest interest in one of their emails, it will be nearly impossible to get off their mailing list.
It is not necessary to manually delete any emails from the spam mailbox because emails greater than X days old are automatically culled from the mailbox daily. The value of X may be set by going to the configure menu, clicking on Options, and then selecting a value for spam-life.
Eventually the Bayesian filter, if it is trained by doing the above, will get more and more efficient at rejecting only spam, at a rejection rate > 98%, with less than 1% false positives.